INTRODUCTION

Who is this book for

Blender is an amazing program and there are many reasons why

someone would want to learn how to use it, but what are the
advantages of learning it from a book? When I was choosing the
content and deciding on the structure I had to make the following
assumptions about you, the reader:

 You are relatively new to Blender, at least in its current form, but
you can find your way round the internet and your computer, install
programs, and generally manage directories and do file house-
keeping.
 You wish to accumulate knowledge and skills in a systematic way
rather than wanting to jump straight in to find the answer to solve a
specific problem.
 You want to read detailed instructions at your own speed, and you
are occasionally frustrated by the, often excellent, online videos
(both the ‘read’ and ‘speed’ aspects being quite variable).
 You value the ‘random access’ nature of books that allows sections
in previous chapters to be quickly re-checked.

What this book tries to do

All practical skills can only be acquired by rolling up your sleeves

and “doing it”. The books and videos on Blender that you find online
(including mine) were not made by people who learned their
knowledge by reading a book. It was gained gradually, over a long
period, by using the software, getting stuck and discovering solutions.

i
 At the end of the day that is what you will have to do too but this book
should make your journey a little less frustrating.

My main aim has been to make learning fun and, in order to do

that, I haven’t stuck too rigidly to a learning hierarchy starting with
simple but rather boring exercises, only progressing to truly amazing
things right at the end of the book. Each chapter does have a main
concept or feature of Blender at its core, but it often includes extra
elements that are best learned in the context of a practical exercise, and
could have been legitimately included in many different locations.
Throughout the book the process of following the instructions and
finding the relevant menu or variable is a very valuable exercise in its
own right.

At the end of the book you will have a good concept of what
Blender can do. You probably won’t be able to remember many of the
specific details of what you did but you will know how to set about
finding out. Importantly you will be unafraid to try making things on
your own, and much less confused by the wealth of help online.

What are the non-aims of this book

This isn’t an exhaustive reference. Blender is a massive program

and mastering it entirely would fill a book much larger than this.

In this book I intentionally try to avoid using the “expert”

technique for doing things. When getting to grips with a new program
it is much easier to think “I want to Add a new Mesh object in the
form of a Cylinder” than, I need to press SHIFT A then M then Y. My
objective is to be explicit and easy to remember in preference to fast
and efficient. As the keyboard short-cuts are visible each time a menu
sequence is followed, it is very natural for you to switch to using them
when an operation has to be repeated many times.

I have also tried to explain how to do everything without resorting

to specific add-ons. This might seem a perverse stance as add-ons are a
really cool aspect of Blender that make so many things better. However
it will be hard for you to judge the value of an add-on if you have never
experienced the problem it sets out to solve. More importantly, when
you are looking for help online, it will be more difficult to follow
examples if you have not been using a “vanilla” Blender and have
always bypassed several steps using an add-on. The exceptions to this
rule are the use of rigify which has become a de facto standard for

ii
 posing and animating characters, and unavoidably where I explain how
to download, install and enable add-ons!

This book takes a subjective view of Blender. It represents my

personal experience and, although I have tried to find out as much as
possible about features I rarely use, there are some aspects that I have
decided to skip over or even neglect completely.

How to use this book

I have written the chapters on the assumption that they would be

read sequentially. Logically, the section on download, install and setup
should happen before any usage of Blender. However in other areas I
will describe something in detail the first time you encounter it, more
briefly the next time, and in one or two words thereafter. For instance:

In Chapter 1 “With the Cube still highlighted and the mouse

somewhere over the 3D Viewport press the S key on your keyboard.
Move the mouse backwards and forwards... then click anywhere in the
3D Viewport to fix the change…

Scroll back out and rotate view so you can see the whole of the
Cube then, using the object interaction mode drop-down (left, just
below the Header line), select Edit mode. Switching between Object
mode and Edit mode is such a common process that it is almost always
done by pressing the Tab key...

Now, with the top face selected subdivide it into 121 squares. You
can either use the menu option: Edge4Subdivide or the context
sensitive right click menu where Subdivide is the first option. Open the
Adjust Last Action panel and change subdivisions to 10 i.e. each
original edge has become composed of 11 edges”

Then in Chapter 2 “Add a Plane to the scene, scale it up about

40 times then switch to Edit mode and subdivide it about 20 times”

From time to time you will forget exactly what you did in a
previous chapter, but because this is a book it’s easy to flick back and
check!

Writing this book was an iterative process: making a screen-

capture video of each exercise, then writing up an explanation, then
revising the video, then modifying the description etc. etc. So my

iii
 recommended approach would be to read each exercise in the chapter
at the same time as watching the video then return to do the exercise
following the instructions, only resorting to watching the video when
you get stuck. Unless you have a powerful computer with two screens
it might be frustrating to try to watch videos at the same time as you
follow instructions in Blender.

Different versions

The examples in this book use Blender version 4 but the

appearance and functionality are very similar to version 2.8 onward, so
if you can’t upgrade to the latest and greatest, for whatever reason,
everything should still make sense.

The change from 2.79 to 2.8 was a more significant one and,
although most of the functionality covered by this book existed in
version 2.79, the layout and styling are quite different. However not all
graphics processors are capable of running versions after 2.79 so you
might be stuck with that! It will be significantly harder to find the UI
options, though you will quite quickly discern a pattern and, generally,
the short-cut key combinations have remained the same.

Big changes from version 2.79

Videos and Blend files

You can find links to the videos along with a sequences of blend
files accompanying each chapter on the web page get-into-blender.com.
I have also included full sized color versions of all the figures in the
book as the process of printing in black and white has made some of
them less clear.

iv
Where else you can find help

The online documents at docs.blender.org are comprehensive and

easy to use.

There is a large and active community of Blender users so,

obviously, an online search will often be your first resort for help. As
well as the official docs, blender.stackexchange.com,
reddit.com/r/blender/ and blenderartists.org all have copious questions
and answers, in the unlikely event that your question hasn’t already
been answered there are many active users just waiting to help you. I
would recommend creating accounts for all of these right from the
start.

There are a few youtubers that I have really enjoyed watching over
the years: Jan van den Hemel’s Blender Secrets, DECODED, Toni
Mortero and Sophie Jantak are all worth searching out. You will find
many more as you search for help or inspiration while you work your
way through this book.

v
 1
O R I E N TAT I O N

What the main elements of the User

Interface are called and where
they’re hidden

In this chapter you will take a whistle-stop tour

introducing you to the logic behind Blender’s structure
and readying you for the more intricate navigation later
in the book.
As you acclimatize yourself to the Blender user interface you'll be
building a realistic model of an emerald and diamond encrusted gold
ring displayed on a velvet cushion. Exploring the different regions,
editors, and menus will give you an inkling of how powerful Blender

1
 can be and how some relatively quick and simple steps can produce
impressive results.

Download and Install Bender

The instructions on https://docs.blender.org/manual for installing

Blender are comprehensive and clear; follow them carefully and
everything should work without problems.

If you are using Windows, the download link on the Blender

website will default to provide the standard Windows installer, which
works perfectly well and is the most straightforward route for that
operating system.

The alternative method, which is needed for macOS and Linux but
also works on Windows, is to download a compressed folder containing
everything you need to run Blender. For a long time Blender has
followed a policy of packaging all dependencies into one .zip, .dmg or
.tar file. The method of unzipping Blender into its own folder and
running the executable file when you want to use it is simple with
much to recommend it. It has allowed me to test the exercises used in
this book on versions 2.79, 2.83, 2.93, 3.6 and so on.

Blender is always being actively improved so you might want to

upgrade it in the medium term. For that reason, it’s not a good idea to
install from the Microsoft Store or use the Debian package manager, as
they sometimes lag several versions behind the current one. Blender
doesn’t have an automatic update system built in so you need to check
the blender.org website from time to time to see if there have been any
important changes.

2
 Blender Configuration and Setup

The configuration options within Blender are enormous but my

policy throughout this book has been to stick to the default setup with
zero modification. I would recommend that you also start that way so
that my descriptions, screenshots, and videos match your own
experience. The only exception is to enable emulation if you're using a
keyboard without a numpad, but don't make the change now. We'll get
to that later in this chapter, and the explanation of why and how to do it
will make more sense once you have a little experience of using
Blender.

Before you start any of the exercises in this book, make a folder
for storing all the files you create as you work your way through. The
name and location are up to you, but I recommend adding a subfolder
for each chapter to keep everything from getting muddled.

Start with a Cube

Once everything is installed, start Blender. You will be presented

with a splash screen with options to open previous work, but for now
press the ESC key or click to the side of the splash. You should see the
traditional starting cube (Figure 1-1) and many buttons and tabs, each
one leading to yet more buttons and controls. There is logic behind all
of this complexity.

Figure 1-1 The starting scene

3
 You’ll be making many simple alterations to this starting cube to
get a feel for the layout of the user interface, where to find the tools
you need and how to make basic changes to objects.

Blender has a lot of help built in, so once you get used to some of
the terminology and know how to look for information, your life will
become much easier. However, there are a few features that will benefit
from a little clarification before you start.

In Figure 1-1 there are actually four distinct areas, each containing
a different editor, the main window has been divided vertically and
each half has then been divided horizontally. The four arrows point
towards the top left of each area where there is a drop-down menu that
allows you to change the editor. Figure 1-2 shows the full list of
available editors..

Figure 1-2 Full list of Editors

There are a lot of options, and some of those editors would require
many pages to explain in detail, but the following are the main ones
you will encounter in the book, roughly in order of appearance.

• 3D Viewport is the place you will create and modify objects

in your scene. This editor can be switched between
different modes such as editing meshes, sculpting or
painting textures onto objects.
• Properties allows you control most aspects of your scene. It
is divided into many different tabs, some of which are listed
in Figure 1-3, other tabs are only visible when selecting
certain types of object or when using certain modes in the
3D Viewport. It isn’t too much of an exaggeration to say
that the rest of this book is essentially a description of what
all the tabs do.

4
Figure 1-3 Properties Editor Tabs

 Outliner shows a hierarchical view of the data used by the blend

file. This editor can be used to select, hide, organize components
and so on, a little like Windows Navigation pane.
 Timeline is the last of the four editors in the default layout and is
used to control animation. You will use it briefly in Chapter 2 and in
detail in Chapter 12.
 Shader Editor allows materials to be edited using nodes, in a
graphical, intuitive way. You will use this editor is every chapter of
the book.
 File Browser accesses the computer file system and is used by
Blender to save and load files when needed. However you can also
use it in one of the screen areas which allows you to drag and drop
media files.
 Image Editor lets you view or edit images and textures. You will
use it in Chapter 5, Chapter 8 and Chapter 9.
 UV Editor is used to edit how 3D objects are unwrapped to allow
images and textures to be applied to their surface. This editor will
feature in Chapter 8.
 Compositor combines images, usually produced by rendering
different parts of the 3D scene but also external images or movies.

5
 The methods of composition are extremely sophisticated,
comparable with fully fledged image editing software, but they use
the visually intuitive nodes interface. You will use this editor in
Chapter 13.
 Dope Sheet, Graph Editor, Drivers and Nonlinear Animation
editors are all used for controlling different aspects of animation
and you will encounter them in Chapter 12.

Mouse-over Tool Tips

Move your mouse over the tiny drop-down button just below the
Blender logo in the top-left corner of the screen (in Figure 1-1 the icon
looks like a ball resting on four crossed sticks). A tool tip should appear
telling you that the current editor type for this area is the 3D Viewport.
Now explore the screen using mouse-over, don’t worry too much about
remembering which elements are where but get used to how long you
need to hover and whether to hover over the label or entry field. Verify
that the four areas on this default layout workspace are 3D Viewport,
Outliner, Properties and Timeline Editors.

For the rest of the book, instructions will refer to the name visible
by hovering the mouse, for instance “Modifier Properties” (hover over
the blue wrench) or “Material Properties” (hover over the pink and
black sphere).

Find the Shortcut from the Menu

A significant past criticism of Blender was that everything required
shortcuts (CTRL, ALT, SHIFT) in combination with letters and
numbers. For power users this setup provided a very efficient work
flow, but for mere mortals, remembering all the key combinations was
almost impossible. Now, however, Blender provides a menu route to
each of these actions (as well as all the shortcuts) and in many cases, a
graphical manipulator as well. In addition, the right-mouse-button
context menu normally has options appropriate to whatever action you
intend to take.

As you moved the mouse over different buttons on the 3D

Viewport, you should have noticed the shortcut reference. The mouse-
over clue is missing for one or two shortcuts (for instance, selecting
edge loops). In those cases the right-click context menu on the tool or
menu option allows you to open the online documentation at the
specific reference.

6
 Figure 1-4 shows the right-click menu that appears for
Select4Select Loops4Edge Loops with relevant documentation
from the link to the online manual on the side.

Figure 1-4 Finding help and shortcut

As backup, if you can’t find a menu or button to give you a hint,

press F3 and search. New Blender users typically start with menus and
buttons, but after a while they start using shortcuts for more frequently
performed actions.

Throughout this book I'll generally include the full menu path,
which is more descriptive and provides a mechanism for finding the
shortcut. In cases where the shortcut is obscure for some reason or is
significantly easier to use than the menu option, I'll reference it when I
first introduce the functionality.

Changing the Cube

In this chapter you'll modify the default starting cube to become

the mounting cushion for an engagement ring. After that you will add
other objects and edit them to form different components of the scene,
and finally you will render your composition to an image file. Links to
the files and videos are on get-into-blender.com 1.ORIENTATION

Let's start with some of the basic methods available in Blender for
modifying objects and their properties. This scene has three objects:
Camera, Cube, and Light, and you can left-click each one to select
them in the 3D Viewport or Outliner Editor areas.

7
NOTE From this point on when I say “click” or “drag” without
mentioning which button, I mean left. The more specific mouse
button instructions may also be referred to as LMB, MMB, and
RMB, as shown in Blender's online documentation in Figure 1-4.

The selected object is highlighted in the Outliner as well as

outlined in orange in the 3D Viewport. Click the Camera and Light
objects before selecting the cube again, and notice that the options in
the Properties Editor change as you click from one object to the next.
Sometimes when you are following instructions and your screen has
different options from the ones you expect, check that you have the
correct object selected.

To change the size of the cube, select it then click the Scale button
in the Toolbar on the left (it’s the one highlighted in Figure 1-5). I've
indicated the four Regions in the 3D Viewport area that you will need
to use often and which don’t have a mouse-over giving their name.

Figure 1-5 The names of four regions on the 3D Viewport

The Toolbar menu will contain different options depending on the

type of editor (3D Viewport, Image Editor, UV Editor) and the mode
selected (Object, Edit, Sculpt, and so on). If a Toolbar exists, press
SHIFT-spacebar to access it as a pop-up, which also is a good way to
save a bit of screen space if you don’t have the luxury of a 4K monitor.

8
NOTE Before Blender 2.93 the default way to access the Toolbar was via
the spacebar, but that has been changed to toggle playing or
pausing the animation in line with most other applications. You can
change shortcuts such as this one in Blender with
Edit4Preferences4Keymap.

Scale the cube using the gizmo (yes, that’s the technical term for
the pink, yellow, and blue thing in a white circle, centered on the cube
in Figure 1-5). You can drag the colored points to scale in X, Y, Z, X +
Y, X + Z, or Y + Z directions. To scale uniformly, drag from anywhere
in the white circle that isn’t colored. Do several different types of
scaling operations, and then click to expand the Adjust Last Operation
where you will see the actual values of the last scaling. You can adjust
these values specifically, so try changing to X to 2.0, Y to 2.0, and Z to
1.0, and then in the Properties Editor, select the Object Properties tab
(orange square), and in the Transform section, look at the Scale field,
which is the open tab on the right of Figure 1-5. You should see
different values because those are the total of all the scaling operations
as opposed to just the very last one you did.

Now you will scale the cube using shortcut keys. With the cube
still highlighted and the mouse somewhere over the 3D Viewport, press
S. Move the mouse backward and forward, then press and release X,
then try Y, Z, SHIFT-X, SHIFT-Y, and SHIFT-Z, moving the mouse to
see the effect. Finally, type the number 3.14159 then click anywhere in
the 3D Viewport to fix the change. When you type a number after a
sequence of shortcut keys such as this, Blender will automatically
interpret the value as a scale, rotation or displacement. The formula that
will be executed when you finally click is visible at the top of the 3D
Viewport. You should have seen that the axis or plane for scaling was
constrained by following the shortcut with other keys to modify its
behavior.

This cube scaling exercise might seem like flogging a dead horse,
but there’s one last option for tweaking properties: in the middle of the
Scale X value of the Properties tab in the Properties area (where it
shows 6.395 on the right in Figure 1-5), while holding down the LMB,
move the mouse from left to right and back again. You should see the
scale increase and decrease as you move the mouse. Finally, in the
Scale value boxes, change X to 8.0, Y to 8.0, and Z to 1.0 which will
explicitly set the scale to exact values.

9
 Moving the Viewpoint

Another crucial aspect of navigating Blender is moving the point

of view around the 3D Viewport. With the mouse over the scene, press
the MMB and move the mouse around, then scroll the mouse wheel to
rotate and zoom. In order to “pan” the view, press and hold SHIFT on
the keyboard before pressing the MMB. The gizmos on the right of the
3D Viewport provide the same controls but also allow orthogonal
views along each axis or from the Camera position.

DIFFERENT VIEWS

Cameras in Blender are actual objects in the scene that are used to
“render” the final image or video. As well as being moved and animated
like a real camera, they have properties such as focal length and
aperture to control depth of field. The active camera’s view is different
from the working point of view you see on the screen but there is an
option to align the Camera to the current view and vice versa.

Occasionally, moving the viewpoint around becomes awkward. As

an example, try scrolling in with the mouse wheel or zoom gizmo until
you are inside the cube, then keep going even when the view doesn’t
move. Quite quickly you will find that you are “stuck.” You can rotate
the view, but panning and zooming don’t seem to do anything. If you
now needed to move so you were viewing the inside of one of the
corners from up close, you would have to scroll back out quite a long
way (eventually it would start to have an effect), then rotate 90 degrees
from the direction you wanted to go, pan sideways, rotate back toward
the target corner, and finally zoom back in again! It’s possible to do
once you get used to using the controls, but often it’s much easier to
change view with Walk Navigation which uses the “game” controls (W,
A, S, D, Q, and E) to move the viewpoint around. The shortcut for this
(View4Navigation4Walk Navigation) is SHIFT-` often called
backtick or AccentGrave in the Blender docs. Figure 1-6 shows a cross-
hairs in the center of the screen indicating the direction of travel.

10
Figure 1-6 Moving the viewpoint in Edit mode using Walk Navigation

To end Walk Navigation, press LMB. To instantly move close to

whatever is targeted by the cross-hairs, press MMB, and to jump back
to where you started, press RMB.

Finally, if you first switch to Camera view by clicking the icon on

the right, the location of the Camera will change as you “fly” around
the scene using Walk Navigation. You can also use the arrow keys on
the keyboard for Walk Navigation, but they lack the up and down
modes provided by Q and E.

Editing Meshes: Vertices, Edges, and Faces

The fundamental object used for 3D modeling is the mesh, which

is defined by a list of points called vertices, each with an X, Y and Z
location. In addition edges are defined connecting pairs of vertices and
faces are defined by three or more vertices. Editing a mesh enables
vertices to be manipulated individually or in groups, as well as
providing ways to add or delete the three mesh components.

Scroll back out and rotate the view so you can see the whole of the
cube. Then, using the object interaction mode drop-down (third line
down on the left in Figure 1-6), switch from Object mode to Edit mode.
Switching to and from Edit mode is such a common process, you'll
almost always press TAB to switch. To change modes without using a
mouse, a pie menu is also available via CTRL-TAB, which is useful if
you need to swap modes frequently.

In Edit mode, you can see the mesh that forms the object—in the
case of a cube, 8 vertices, 12 edges and 6 faces. All the vertices are

11
selected initially, but when you click one of them, all become
deselected apart from that one. Holding SHIFT while clicking on a
vertex toggles the selection of that vertex; it behaves like the CTRL
button in most other applications! It might take a bit of getting used to
so have a play around at selecting and deselecting vertices.

To see different parts of an object, it’s possible to swivel, zoom,

and pan the viewpoint; however, toggling the X-Ray view on and off is
a very efficient way of fine-tuning vertex selection. The X-Ray button
icon has two overlapping squares; in Figure 1-6 it is the furthest right
button visible. Finally, keyboard shortcuts A and ALT-A are frequently
used to select or deselect all the vertices.

Subdivide
To make the cube into a mount for a ring you will need to
subdivide the top face so that it can be made convex and also, so that
some of the subdivided faces in the center can be depressed to form a
slot. Subdivision is one of the simplest function in Blender to create
new vertices, edges and faces. If one face is selected then subdivided, a
number of smaller faces will be created along with the appropriate new
vertices and edges.

Being able to select edges or faces instead of vertices, and perform

appropriate actions specific to them, is another crucial aspect of using
Blender. Try switching between vertex, edge and face selection using
the three buttons in Figure 1-6 just to the right of the drop-down
showing Edit mode. Now, with the top face selected subdivide it into
121 squares using Edge4Subdivide. (There is no shortcut, but
Subdivide is the first option in the context-sensitive RMB menu). Open
the Adjust Last Action panel and change subdivisions to 10 so each
original edge becomes composed of 11 edges.

In general you move vertices in Blender by first selecting them,

then Mesh 4Transform4Move (or use the relevant gizmo in the
Toolbar). However, moving, scaling and rotating vertices, edges, faces
or whole objects are such common operations that you will almost
invariably do this using the G, S, and R shortcuts often in conjunction
with a subsequent X, Y, Z, SHIFT-X, SHIFT-Y, SHIFT-Z and
sometimes followed by a number representing a distance, scale factor
or rotation in degrees. The change becomes fixed when the LMB is
clicked. These shortcuts in Edit mode behave the same way as when
you were scaling the default cube in Object mode earlier. Select a few

12
 vertices and do some trial move, scale and rotate changes, undoing
each with CTRL-Z before continuing.

VARIATIONS OF THE MOVE, SCALE AND ROTATE SHORTCUTS

Several aspects of using the G, S, and R shortcuts are hard to find

using mouse-over help or search, but they can often be very useful:

 Use SHIFT while dragging with the mouse for fine control. This
also works when changing almost anything by moving the mouse such
as value sliders, color pickers, even snapping to grid increments.

 Rotate by pressing R twice to enable track-ball rotation and

simplify tasks such as rigging and posing. It's often a time-saver when
posing bones or positioning objects in a scene.

 Press CTRL while moving, rotating, or scaling to enable

snapping. You can press CTRL and SHIFT at the same time to snap at
a finer scale.

 In Edit mode press G twice to constrain vertex movement along

existing edges. You will probably find fewer circumstances when this
option is needed, but on occasion, it can prove very useful.

One button in the Header that is much used when editing meshes is
Proportional Editing (near the middle of the second line in Figure 1-6
under the word Shading). This allows graded movement of large
numbers of vertices. The size of the field that Proportional Editing
effects is shown as a circle, and you can adjust the diameter using the
mouse wheel. You can revise the proportional editing values in Adjust
Last Action as well as the actual amount of the change.

To make the top of the mount cushion convex you need displace
vertices downwards by increasing amounts, the nearer each is to a
corner, if Figure 1-9 you can see the curve that I achieved. Change
back to Vertex select, if it is still set to face select, and turn on
Proportional Editing. Now select all four corner vertices of the top face
and move them down using shortcut G then Z then LMB, in a similar
way to the scaling you applied to the cube earlier. The three
transformations; scale, move and rotate are so ubiquitous while
modeling that I will refer to them throughout the book by their
shortcuts rather than their menu path. You are aiming to form a nice

13
 convex surface, but you may get a better shape if you do this in several
steps with different sizes or profiles for the Proportional Editing. I
found that the Sharp profile was nearest to what I had in mind.

When the shortcuts are followed by X, Y, or Z to constrain

movement, scale, or rotation to one axis, a colored line will appear to
indicate the constraint.

Extrude
Once the top of the mount cushion is rounded, select five adjacent
faces in the middle to extrude downward to create a “slot” to hold the
ring., To do this, change back to face select, press the LMB to select
the first, and then press SHIFT-LMB to add four more to the selection.

Select Face4Extrude, move the mouse downward a small

distance, then press the LMB to fix it. Extruding is a common process,
so you'll very quickly learn to use the shortcut E. The default behavior
is extruding perpendicular to the surface (shown by a blue axis line),
and pressing Z repeatedly will toggle between local perpendicular, free
movement, and global Z axis. On the other hand, the Toolbar extrude
gizmo has a handle to drag for perpendicular extrusion, and you can
click and drag from anywhere else in the circle for “free” extrusion.

Creating the Ring

From this point on you’re going to add some more simple

geometrical shapes and edit their meshes and material properties to
make the ring, an emerald, a setting and some diamonds.

1. To create the band of the ring, switch back to Object mode and
select Add4Mesh4Torus. This torus will become the vertical
band in the cushion so select that object and rotate it through 90
degrees using R, Y, 90 then LMB and scale it to match the slot
in the cushion.

2. Add two polyhedrons to represent an emerald and its mount.

The one with fewer sides will become the mount and the higher-
sided one the emerald. Select Add4Mesh4Ico Sphere with
one division and do another with two divisions (expand the
Adjust Last Action to change the number of divisions). Scale
the two icospheres appropriate to the size of the ring,

14
 3. Copy the emerald to form a basis for four diamonds. With the
emerald selected copy it using CTRL-C and paste it with
CTRL-V. So long as the cursor is over the 3D Viewport the
copied object will appear in exactly the same location as the
object it was copied from.

4. Scale down the copied icosphere then copy it three more times.
These smaller icospheres will be diamonds set into the ring
(look ahead to Figure 1-10 to see the sizes I used, but design it
to your own tastes—after all, these carats cost nothing). Press
G, Y, then LMB to move the copy to one side along the Y axis.
Scale it down then repeat the copy, paste, and move to one side
step three times.
To make the torus look more like a ring, you'll change its cross
section from circular to oval. Your objective is to select a circle of
vertices on the inside of the ring and scale them up, then select a circle
of vertices on the outside and scale them down as shown in Figure 1-7.

Select the ring object then switch to Edit mode to allow you to
move vertices. However before you attempt to do this you need to learn
a little more about selecting vertices in Blender.

Figure 1-7 Scaling the inside and outside ring of vertices

More Ways to Select

There are several ways to select vertices, edges or faces beyond

simply clicking on them. When you first start Blender, selection
defaults to select box which allows you to draw a rectangle around the
vertices you want to select. There are actually quite a few other

15
 methods for selection but, initially, you will find that much can be done
using the default select box on the Toolbar and shortcut C to circle
select for fine-tuning.

With circle selection the size of the circle is controlled with the
mouse wheel and vertices are added by LMB and “painting”, removed
by MMB and the selection process ended by RMB or pressing the enter
key.

You can add to, or remove from a selection by pressing SHIFT or

CTRL at the same time as dragging the rectangle,

5. Your objective here is to select just the innermost vertices using

circle selection. Toggle X-Ray on and switch to orthographic
view so the ring can be seen from the side by clicking on the X
in a red circle on the view rotation gizmo. You will find that it’s
quite difficult to control, but making the circle size near to the
internal diameter of the ring does make it a little easier.

NOTE The shortcuts for changing to orthographic or camera view use the
numpad. If you are using a laptop or a keyboard without a numpad
it is still possible to emulate one by selecting from the main menu
Edit4Preferences4Input4Keyboard4Emulate Numpad. In
my opinion the numbers for changing the view are less intuitive to
use when they are laid out in a line instead of a grid and the view
rotation gizmo is quick and easy to use. However CTRL-PLUS
and CTRL-MINUS are very useful shortcuts for the unwieldy
Select4Select More/Less4More (or Less) so I strongly suggest
you enable the Emulate Numpad option. From this point on when I
refer to PLUS or MINUS I mean the keys on the numpad or
emulated numpad.

6. Now try to get the inside curve of the ring more oval. Scale the
innermost vertices up with proportional editing toggled on and
the diameter adjusted to an appropriate amount. This is a good
example of a time when it’s useful to open the Adjust Last
Operation pane and tweak the Proportional Size value while
watching how the mesh changes.

7. To select the outermost ring of vertices you will use by far the
most efficient method for selecting loops of vertices. Select a
representative edge, then Select4Select Loops4Edge Loops.

16
NOTE Edge Loop selection is such a useful feature that you will quickly
prefer the shortcut ALT-LMB rather than the menu route, which is
much slower. This shortcut also happens to be one that doesn’t get
a hint in the menu see (Figure1-4) because there is an alternative
shortcut to avoid conflict with 3 Button Mouse Emulation.

8. Scale the outermost vertices down with proportional editing on.

Remember to toggle proportional editing off again when you’ve
finished.

9. Blender allows objects to be “smooth shaded” or “flat shaded”

and for this exercise we want the ring to be smooth and the gem
stones and mount to look faceted. Switch back to Object mode,
select the ring then in the menu Object4Shade Smooth.

10. The icosphere with fewer sides will become the mount so you
will transform it into more of a conical shape. Select the
icosphere, switch to Edit mode and move the top vertex
downwards with G, Z then LMB, then select the five vertices
nearer to the bottom and scale them inwards. You could
constrain the scaling to the XY plane with S then SHIFT-Z, but
as all the vertices are in the same plane it will have no effect so
you can just use S.

11. In reality the ring and the mount will all be one piece of gold
and you can join them in a similar way in Blender. Switch back
to Object mode and using SHIFT-LMB, select the ring as well
as the mount object then, with the ring and mount both selected,
join them to form a single object using Object4Join.

NOTE This is a potentially confusing feature of Blender: When multiple

objects are selected there is a redder shade of orange used for
highlighting all objects prior to the last one and there is a similar
difference in highlighting in the Outliner Editor area on the top
right. The redder orange objects are termed “active” and only the
last selected object is termed “selected”. For this exercise the order
of selection (and shade of orange) doesn’t matter but for many
processes it is crucially important so this is something you need to
watch out for when you read the documentation or follow videos
online.

17
Adding a material to an object

For the rest of this exercise I’m going to guide you through aspects
of Blender that you will only cover in detail later in the book:
materials, shaders, nodes, textures, particles and rendering are all things
that you will definitely need to understand eventually, but not at this
stage. Your objective while following these instructions is just to
become more comfortable finding your way round the user interface.

1. To see what your gold and gem materials look like you need to
select Material Preview from the four buttons that control the
Viewport Shading at the top right of the 3D Viewport (to the
right of the X-Ray button).
The options are: Wireframe, Solid, Material Preview and
Rendered. They are all useful for different purposes but increase in
computational requirements from left to right. Rendered can be slow if
Blender has to simulate all the light paths being reflected, refracted and
scattered by many objects in the scene. Material Preview uses some
clever algorithms to give a good approximation to the final product.

2. First create a material to represent the gold. With the combined

ring and mount still as the selected object, click on the Material
Properties (pink & black sphere icon) in the lower right
Properties Editor area and click on + New to create a new
material. Note there is another + button on the right that will
add a new slot, you will learn a little more about material slots
in the next chapter.

3. The new material will have Surface set to the default Principled
BSDF Shader which is very flexible. To make it look like gold
you just need to scroll down and change the Base Color to
yellow, increase the Metallic component and decrease the
Roughness.

Using Nodes

You should already have a pretty good looking gold, however to

give an antique, hand-made look to your material I will quickly take
you through the process of adding some nodes. Again, don’t worry
about the details of this for now, using nodes is an important feature of
Blender so it will crop up again and again through the book.

18
4. To work with material nodes you need to set one of the areas to
use the Shader Editor. You can do this by clicking on the
Shading tab at the very top of the window which will change
the Active workspace. You can modify the type and size of the
different editor areas quite easily but the Active workspace tabs
are a convenient way to switch between tried and tested default
layouts (see Figure 1-8). The bottom of the screen will now
contain a Shader Editor area and you can zoom in on this with
the mouse wheel, pan the view with the MMB and select and
drag the nodes with the LMB.

5. The first node to add is one that will produce a random, beaten
pattern which can be used to change the reflection from the
surface of the gold. Using the menu at the top of the Shader
Editor area Add4Texture4Noise Texture, un-check the
Normalize option, then connect its output (Fac) to the Base
Color of the Principled Shader by dragging and dropping the
dot to create a connector line. This is just a temporary setup to
check what the texture looks like. Now increase the Scale to
about 18 so the pattern looks proportionate to the size of the
ring then drag the output connection from Base Color to
Roughness. Older versions of Blender had a dedicated node for
this called Musgrave texture.

6. To tone down the strength of the pattern you can feed the output
through a node to multiply it by a small value. Use the menu
Add4Converter4Math to create a new node and drop it onto
the connection between the Noise Texture and the Principled
Shader nodes, which should automatically insert the node into
the connecting “pipe”. If you miss then you can reconnect the
inputs and outputs so they match Figure 1-8.

7. To distort the surface of the metal further you can also use the
noise pattern as a displacement. Change the type to Multiply
and the value to 0.01 then drag a second connection from the
output Value to the Displacement input of the Material Output
node. You can adjust the multiplication value to vary the
unevenness in the final render.

19
Figure 1-8 Shading workspace

8. Next you need to create an emerald material in the relevant

object. In the 3D Viewport select the emerald object icosphere
and add a material as you did for the ring.

9. To make the material transparent, in the Material Properties tab,

click on the Surface shader and change it from Principled to
Glass BSDF, change the Color to green, Roughness to zero and
increase the IOR a bit (refractive index of emerald is about
1.58)

10. Select one of the diamonds and add a new material to that
object changing to the Glass BSDF shader as well, but with
Color white and IOR 2.4. For the other three diamonds select
the same diamond material from the drop-down list in the
Material Properties tab. You can see the drop-down half way up
Figure 1-8, just above the Preview section.

11. Finally arrange all the objects so they look to be part of one
ring. Move the ring down so it nestles in its slot then move the
emerald and diamonds to their positions on the ring, this will be
easier to do using the orthographic view.

Velvet with Quick Fur

Your next objective is to make a luxurious velvet material for the
cushion. As a first attempt it would make sense to use the Velvet BSDF
shader.

20
 1. Select the cushion object which started life as the default cube.
It will already have a white material called “Material”. In the
Material Properties tab change the Surface shader from
Principled BSDF to Velvet BSDF and the color to dark red.

2. With Viewport Shading as Material Preview it looks rather flat

so change to Rendered, the rightmost of the four buttons.
However it still isn’t very velvety.

3. “Quick Fur” sounds hopeful but you will find that it is applied
unevenly unless you apply the scaling. First click
Object4Apply4All Transforms then Object4Quick
Effects4Quick Fur then In the Adjust Last Operation panel
un-tick the three check-boxes; Apply Hair Guides, Noise and
Frizz. Reduce the View Percentage to 0.1 and change the
Density to High.

4. If you expand the Cube object in the Outliner Editor you will
see that a new object has been added called Curves, as a child of
Cube, as visible on the right of Figure 1-9. This has now
become the selected object. Click on the Modifier Properties tab
(blue wrench) and in the modifier called Generate Hair Curves,
reduce the Control Points to 2. In Set Hair Curve Profile,
increase Radius to 0.005 and in Interpolate Hair Curves,
increase the Density to 750.

QUICK EFFECTS

Adding Fur, Explode, Smoke or Liquid at the press of a button

sounds really amazing and I would certainly recommend playing around
with these, once you’ve finished the exercises for this chapter. However
the Quick Effects are just shortcuts to setting up Blender Physics,
Particles and Modifiers with appropriate settings and many of them will
be frustrating if you try to create a rendered image or animation before
you have covered the relevant topic in the later chapters.

5. You need to set the fur fibers to a similar red to the base cushion
but you can also add some some variety to that material using a
different texture node. With the Curves object still selected, in
the Material Properties tab a new Fur Material will have
appeared, change its color to dark red. Select the Cube object
(it should still have the original material called “Material”) and
in the Shader Editor area add three new Nodes
Add4Texture4Voronoi Texture then Add4Input4RGB
then Add4Color4Mix Color mix the Voronoi and RGB and

21
 feed the result into the Velvet shader, adjusting the controls until
you get an interesting result, the node arrangement is shown in
Figure 1-9.
In Blender version 3.5 the method for generating hair changed to
use geometry nodes. If you are using an older version then the original
version of these instructions can be found in Appendix A.

Figure 1-9 Velvet cushion material

Don’t spend too long fine-tuning the appearance of the velvet

material as this will undoubtedly change when you position the camera
and set the render engine in the next section.

Moving the camera

To compose the final image you need to position the camera and
run a few test renders to finalize all the settings.

1. For the exercise in this chapter you will use the cycles render
engine. In the Render Properties tab of the Properties editor (the
icon like the back of a camera) change the Render Engine drop-
down from Eevee to Cycles. You will learn more about the
differences later but, at a simplistic level, eevee is quick but
cycles is more realistic.

22
 2. Position the Camera by clicking on the camera button on the
right of the 3D Viewport. If you scroll out with the mouse wheel
you will see a rectangle surrounded by a dotted line to indicate
the image frame. Use SHIFT-` and WASD game style controls
to “fly” with Walk Navigation so the scene is well composed
within the camera’s field of view.

3. To start the render, select from the menu right at the top of the
window Render4Render Image. While fine-tuning the
material settings you can stop the render early by pressing ESC,
the first time will stop the render and pressing ESC again will
close the window. In subsequent chapters you will see how to
set a small render region so you can quickly check whether the
settings are correct.

A synthetic window

You will have probably noticed that the reflections from the gold
ring are much more interesting and produce a more realistic result
when the Viewport Shading is set to Material Preview compared with
the render. We will see later how to set this up properly but for the time
being you can improve the scene by adding a big square “light” behind
the camera with a vertical and horizontal bar in front of it. This will add
highlights and illumination as if from a window.

1. First of all create a rectangle that emits light. Create a plane

using Add4Mesh4Plane then scale it up, move and rotate it
vertically behind the camera. Add a new material to the plane
and change the Surface to use the Emission shader.

2. Now put a vertical and horizontal bar in front of the light. Add
an additional plane then scale, rotate and move it to form a bar
in front of the window. Finally copy, paste and rotate the bar to
make a cross.

The final render

The finished render (Figure 1-10) will probably take a few minutes
but when it’s finished you can save the image to your working folder
with Image4Save As

23
Figure 1-10 Final render

Conclusion

The main purpose of this chapter was for you to get used to
navigating the user interface. You saw that the default layout
workspace is divided into four editor areas each with its own special
features and functionality. As well as moving, rotating and scaling
objects and manipulating vertices you used two other methods for
creating new vertices in a mesh: subdivide and extrude. As the book
unfolds you will learn how to use the many other mesh editing tools
from the Toolbar and lots of ways to generate mesh real estate.

Selecting objects is straightforward at one level but it is also

something you will need to be careful of later when you learn about
parenting objects. Selecting vertices can also be simple, but being able
to do it efficiently when working with complex meshes is an important
skill and you will build an armory of different approaches as you work
your way through the rest of this book.

The next chapter introduces the subject of modifiers which will

provide you with a treasure chest full of methods to build complicated
geometry very quickly. As you complete the exercises you will need to
use the lessons you learned in this chapter and, at times, you may find
yourself racking your brains to remember what to do. But take your
time, refer back when you get stuck, watch the online videos and check
the documentation. It’s all part of the learning process and very soon
you will become familiar with all the intricacies and quirks of the
Blender user interface.

24
 2
MODIFIERS

A powerful and flexible mechanism

for making changes to objects

Having learned the basics of selecting objects and

vertices, rotating the view point, and switching between
regions and editors, it's time to get to grips with the next
core concept: modifiers. Once again, this particular
subject could fill several books; rather than try to
memorize everything, the goal is to explore and gain an
appreciation for the kinds of things you can do with

25
modifiers, bearing in mind that this chapter presents only
a tiny sample of that capability.
Modifiers (also called non-destructive modifiers) allow you to
apply a process or function to an object to produce the appearance of a
new object, but they leave the original object intact. In addition, you
can apply successive modifiers to the same object, each taking the
output of the previous one and modifying it further. In this chapter,
we'll start using modifiers to create three increasingly complex projects
so you can grasp their potential power. Files and videos for this chapter
are on get-into-blender.com 2.MODIFIERS

Booleans: Adding and Applying a Modifier

Boolean modifiers provide a way to construct complex shapes

from simpler ones by subtracting or adding one object to another. As an
example of working with Boolean modifiers, let's make a coffee cup:

1. Start Blender, delete the default starting cube, and select

Add4Mesh4Cone. In Adjust Last Operation, change Radius
2 to 0.6, and then rotate the object by 180 degrees about the X
axis.

2. From Object4Duplicate Objects, press Z, and move the

duplicate slightly upward then scale it down, but ensure the
smaller cone protrudes above the top surface of the larger cone,
as shown in Figure 2-1.

Figure 2-1 Boolean difference modifier using two truncated cones.

26
 3. Select the larger cone, click the Modifier Properties tab on the
Properties Editor area (blue spanner icon) and Add
Modifier4Boolean (Figure 2-1). Keep the combination type as
Difference, select the smaller cone (Cone.001) as the object to
operate on.

4. To see what the resultant shape looks like, click the eye icon in
the Outliner Editor to toggle the visibility of the smaller cone.

5. For the cup handle, go to Add4Mesh4Torus, rotate it 90

degrees about the X axis, scale it down, and move it to an
appropriate position on one side of the cup (easiest done in
orthographic view along the Y axis).

6. To make the handle more elegant, carve some of the ceramic

material from the outside of the torus using a cylinder:
Add4Mesh4Cylinder, rotate it 90 degrees about the X axis,
scale it down, and move it so it contains most of the torus as
shown in Figure 2-2. Scale it slightly more in the Z direction so
the handle is thinner at the top and bottom.

Figure 2-2 Handle made from a torus intersecting a squashed cylinder

7. With the torus selected, add a boolean modifier and change to

Intersect with the cylinder object. Toggle the visibility of the
cylinder in the Outliner Editor, and adjust the scale and position
values of the cylinder until the handle looks just right.

8. Select the larger cone object and add another boolean modifier,
but change the type to Union with the torus object as the target.
You should be able to see that something is wrong here: because
the torus is being added after the cone has been hollowed out, it

27
 protrudes inside the cup. To fix it, drag the boolean difference
modifier down below the union modifier using the area to the
right of the cross (it contains eight white dots; see Figure 2-3).
An option in the drop-down menu also allows you to change the
order of modifiers.

9. At this point, moving any of the objects would break the

assembly. You can fix this by “applying” the modifiers, which
effectively bakes them as a permanent change to the mesh (in
Chapter 3, you’ll learn a different method to solve this
problem). Select the torus, and in the modifier, choose Apply
from the little drop-down between the camera and cross on the
top line (Figure 2-3). Select the larger cone and apply first the
union, then the difference modifier. Delete the smaller cone, the
torus, and the cylinder objects.

Figure 2-3 Apply is hidden in a drop-down menu.

NOTE Watch out for the little downward pointing chevron indicating a
drop-down menu. In later chapters, when working with vertex
groups and shape keys, you will find crucial functions hidden
beneath a tiny button like this.

A significant difference between the boolean union modifier and

the join method we used in the last chapter to merge the emerald setting
to the ring (select multiple objects then Object4Join) is that the
modifier removes all the overlapping internal vertices (Figure 2-4).

28
Figure 2-4 Join on the left, union modifier on the right

Join also doesn’t integrate the two meshes. In Figure 2-4 the faces
of the two cubes on the left pass through each other without forming
edges and vertices at the intersection.

Subdivision, Bevel, Mirror, and More

It’s hard to quantify the relative importance of Blender’s modifiers,

but you are likely to use Boolean, Subdivision, Bevel, Mirror, and
Displace with high frequency. You will utilize all of these while
completing the main exercise for this chapter, but first let's look at
some details of these modifiers and review a few other interesting ones.
As an example, you will construct a simple ornament such as might
have been cast by a primitive Bronze-age craftsman.

Subdivision
Subdivision is the process of creating a smoother, higher polygon
mesh from a coarser one. When modeling with Blender, you should
always try to create simple meshes with as few vertices as possible that
generate smooth, detailed objects when a Subdivision modifier is
added. Easier said than done!

1. Clear the scene and start a new one: File4New4General

(don’t bother saving the coffee cup).

2. Add a second cube, scale it down slightly, then use the Boolean
modifier to add the smaller cuboid shape to the initial cube three
times, applying the modifier each time. Add the cubes on the
side facing the negative Y direction (so the description below is

29
 easier to follow when I refer to X and Y; see Figure 2-5). Delete
the smaller cube when you have finished adding it.

Figure 2-5 Don’t subdivide too much.

3. Select Add Modifier4Subdivision Surface and increase the

Levels Viewport a few times. Don’t increase it more than five or
six unless you have a powerful computer though, as each level
increases the number of faces by a factor of four. To witness the
geometric increase, right click at the bottom right of the Blender
window to toggle Scene Statistics.

Bevel
Bevel adds additional faces along edges to make them chamfered
or rounded. Obviously this is useful when an object's design requires
beveled edges, however, the modifier is often used to add definition to
edges that separate fairly flat surfaces, which turns out to be difficult to
do well in 3D modeling.

4. With the cube object selected, Add Modifier4Bevel, but

notice it has no visible effect because it is acting after the
subdivision and all the sharp edges were removed by that
modifier. The default setting is to bevel edges where the
difference from one surface to the next is greater than 30
degrees, but there are various ways to specify which edges to
bevel (we'll cover those different ways in Chapter 11).

5. Move the Bevel modifier up to above the Subdivision with the

Move to First option from the drop-down or drag it using the
top-right corner of the modifier pane. The subdivision surface
follows the cuboids more closely because of the extra edges that
beveling introduced.

30
 Mirror
The Mirror modifier is another way to create more vertices quickly
by duplicating a mesh across planes of symmetry. If you are making an
object that will be symmetrical, it is much better to make only a half or
a quarter of it and then mirror it.

6. Click the cross next to the little drop-down (top-right of Figure

2-5) to remove the Bevel modifier, then switch to Edit mode
and select the four vertices on the unaltered side of the original
cube. Select View4Sidebar, click the Item tab, and
Transform the values, setting the Y value of the vertices to 0.
Figure 2-6 shows the process.

Figure 2-6 Alter the position of a selection of vertices.

7. Select then delete the face with these four vertices at its corners.
When you press DELETE, you will get an additional menu to
clarify which combination of faces, edges, and vertices you
want to delete (in this case delete only the face, leaving the
vertices and edges intact). If you use the RMB menu delete
option, the choice will be more specific depending whether you
are working with vertex, edge, or face select.

8. Select Add Modifier4Mirror and set the axis to Y.

NOTE In general, the order of the modifiers is important and when

applying them, you need to do it from the top down. However, for
the Mirror and Subdivision surface, the result will be the same if
the mesh is mirrored then subdivided or subdivided then mirrored.

31
 Experiment with Other Modifiers
Try adding a few other modifiers to the stack. Most will have no
visible effect because they need other components for them to work;
however, the following exercise should yield interesting results. For
each modifier, move it up the stack to see how the final result changes.

9. Decimate is a way to remove vertices but preserve, as far as

possible, the original shape. Try Add Modifier4Decimate. A
clearer display of the scene statistics is also available using the
Viewport Overlays options (the drop-down next to the
overlapping circles icon as shown in Figure 2-7) as well as
other features to customize the 3D Viewport. With smooth
shading, you can often reduce the number of faces quite
significantly before the object becomes too misshapen, but
notice that the faces become rather randomly shaped, which
might be a problem for certain applications (as you'll see later).

Figure 2-7 Viewport Overlays Statistics

10. In the Decimate modifier settings, switch to Un-Subdivide,

change the Iterations, then switch to Planar and change the
Angle Limit.

11. Delete the Decimate modifier then Add Modifier4Remesh.

Remesh replaces the original mesh with one made from
quadrilateral faces, the size of which is controlled by the Voxel
Size. Be careful when reducing the Voxel Size, as you can
easily overload your computer! (I always type a value rather
than use the slider.)

12. Delete the Remesh modifier, then select Add Modifier4Build.

Build is an animation modifier, so the object will disappear after

32
 you add it! Use the buttons in the Timeline Editor at the bottom
of the window (Figure 2-8) to play the animation, which will
create the object one face at a time.

Figure 2-8 Build modifier showing the Timeline play button

13. Select Add Modifier4Solidify and increase the Thickness

value to 0.1m. When you run the animation, Solidify will build
the surface as solid blocks rather than an infinitely thin shell.

14. In the menu at the top of the 3D Viewport, select

Object4Shade Smooth to make the curved surfaces of the
object look better, but at the expense of the sharp edges created
by the Solidify modifier.

15. In the Object Data Properties tab (the icon is three tiny green
squares connected in a triangle, third from the bottom in Figure
2-5), open Normals and tick the Auto Smooth checkbox. For
many purposes this option will do the trick, but it leaves
perfectly sharp edges that never exist in the real world. In later
chapters, you'll use alternative methods that lead to greater
realism.

33
 Constructing a Complex Scene Using Modifiers

The rest of this chapter will round off your introduction to

modifiers with a more elaborate work - a Martian habitat. The
completed scene should look something like Figure 2-9.

Figure 2-9 Partly completed Martian habitat

You will make the terrain and some habitat domes, but your first
job will be to create simple manikins to populate them.

The Mirror, Skin, and Subdivision Modifiers

Starting with half of a stick figure, you'll use a mirror modifier to
make it symmetrical, then a skin modifier to make it into a solid, but
rather “boxy” figure. Finally you will add a subdivision modifier to
generate a smoother result; it should look something like Figure 2-10
when you're finished.

34
Figure 2-10 The completed manikin

1. Clear the scene and start a new one. Select the default cube, and
then in Edit mode, delete seven of the vertices.

2. Move the remaining vertex to a location above the origin (X=0,

Y=0, Z=1 for instance, but the exact height doesn’t matter).
Rather than use the Sidebar Transform to set the positions to
zero, as you did in the previous exercise, toggle the Snap
magnet at the top and set it to Absolute Grid as shown in
Figure 2-11. Use the X and Y orthographic views to position the
vertex above the origin.

Figure 2-11 Using Snap to position vertices

3. Select the single vertex, extrude it a few times in the Z direction

for the lower spine, upper spine, neck, and head. Then, using
the orthographic views, extrude sideways and downward from
the shoulders to create an arm and from the lower spine to form

35
 a hip and leg. You'll apply a mirror modifier, so you need to do
only one side. Although Blender can mirror in all three
directions, the default direction is –X to +X, so extruding
sideways in the X direction will make your life easier later on.

Switch back to Object mode and select Add Modifier4Mirror

(setting the correct axis to match the orientation of the stick
figure). Next, select Add Modifier4Skin and then Add
Modifier4Subdivision Surface.

4. Return to Edit mode and select all the vertices (toggle XRay to
be able to see them through the skin), then select
Mesh4Transform4Skin Resize to make it a reasonable
diameter. Selecting a few vertices at a time, adjust with Skin
Resize to form head, shoulders, knees, and toes (you will find it
much more convenient to use the shortcut CTRL-A at this
stage). When the manikin looks reasonably proportioned, move
it out of the way for later.

The Displacement Modifier

The surface of Mars where the research station is being built is a
small crater: a gentle hill with a depression at its top.

1. Add a plane to the scene, scale it up about 40 times, then switch

to Edit mode and subdivide it about 20 times. In Adjust Last
Action for subdivide, you can type in a value, but the slider is
limited to a maximum value of 10, so if you use the slider you
will have to subdivide the plane twice.

2. The Displace modifier can move vertices by a fixed amount or

an amount controlled by a Texture, so you can use that to make
your terrain uneven. Back in Object mode, select Add
Modifier4Displace, and in the modifier settings, click New to
add a texture.

3. Select the Texture Properties tab (pink checkerboard icon).

Select the texture just created in the modifier (probably named
“Texture”), and choose Clouds for the Type of Texture. This
will produce a very bumpy plane, so go back to the Modifiers
tab, reduce the Strength to around 0.1 and apply the modifier.

4. Switch back to Edit mode and toggle proportional editing to on,

then select one vertex and move it up so it creates a “hill” over
half the area of the plane. Reduce the radius of the proportional

36
 editing influence and move the pixel down to create the crater.
Remember to turn proportional editing off again.

Extruding a Cylinder
The habitat is made from geometric shapes “sunk” into the terrain
to form domes with an arched corridor linking them. The lower domes
have been glazed and pressurized with air but the top dome and
corridor are still under construction.

1. In Object mode, add an icosphere, then scale and position it in

the crater so it forms a dome. Setting the viewpoint along each
axis while positioning objects or vertices is the easiest way to
do this step. The advantages of this method are that the view is
orthographic (no perspective), so you can easily see what lines
up with what, and any movement or rotation (but not scale) are
constrained to a plane parallel with the screen.

2. Add a cylinder and scale, rotate, and position it so it's

“submerged” half into the terrain with one end inside the dome.
This will be the start of the corridor linking the top dome with
the lower domes. The cylinder's axis should be parallel with the
surface.

3. Switch to Edit mode with face select enabled, select the end of
the cylinder, and extrude it into a series of short sections
following the slope of the hill (Figure 2-12).

Figure 2-12 Extrude and steer the corridor.

4. “Steer” the corridor by rotating and moving the end face when
viewing from above and follow the contours of the hill by
viewing from the side.

37
 Combining a Collection of Objects
For the lower domes, you'll start by using a Boolean modifier to
join together four cubes, but rather than re-using a single extra cube,
you'll move the four cubes into a Collection and combine them in one
step. Sometimes assembling more than two objects at once is useful,
and this is a flexible way of doing that.

1. Add four cubes near the lower end of the corridor, scaling and
moving each one so they overlap each other corner to corner.
Select all four cubes then select Object4Collection4Move to
Collection4New Collection. Select just one of them (to
become the combined object), then Add Modifier4Boolean
(Union), and finally switch Operand Type to Collection and
select the one you just created (Figure 2-13).

Figure 2-13 Boolean modifier acting on a Collection

2. After applying the modifier, delete the new Collection to tidy

everything up in the Outliner Editor. That step won’t delete any
of its contents, in this instance the four cubes, but it will leave
them outside the original Collection. Delete each of the three
un-merged cubes, and drag and drop the combined object back
into the original Collection.

3. Select Add Modifier4Subdivision Surface then scale and

move the intersecting domes so they overlap the bottom of the
corridor and are half submerged in the terrain so they look like
domes.

38
 The Wireframe Modifier
The Wireframe modifier works in a similar way to the skin
modifier you used for the manikin, but it's more suitable for surface
meshes. It converts the edges of objects into 3D “wires” and removes
the faces. The resulting object is a cage resembling the framework of a
glass-house prior to any glass being fitted..

1. Use a Boolean union modifier to add the corridor to the top

dome apply the modifier then delete the corridor.

2. Select Add Modifier4Wireframe to the combined object, and

add the same modifier to the lower domes.

3. Select the upper domes, then the Material Properties tab, then +
New to add a material, increasing metallic value and reducing
roughness.

4. To check what this material will look like in the final render,
change the Render Engine from Eevee to Cycles in the Render
Properties tab (white back-of-camera icon).

5. Change the light from a point source to a “Sun.” Select the light
object either in the 3D Viewport or the Outliner Editor area. The
default light is named “Light” and its type is Point, so click the
Object Data Properties tab (the green lightbulb in Figure 2-14).
Under Light, select Sun and change the power to 4.0 W/m^2.
Rotate the direction of the sun to be more horizontal using R in
the 3D Viewport.

Figure 2-14 Sun settings

6. With the viewport shading set to render, select the upper domes
object, tweak the material settings, and see the result.

39
 7. Select the terrain, add a material, and change its Base Color to
a suitably Martian orange.

Copy and Paste or Duplicate

Pressing CTRL-C then CTRL-V is the general way to copy
and paste in most applications, so using that method means one less
thing to remember, but if you try right-clicking an object to access the
Copy context menu, you'll notice that Blender has an option to
“Duplicate selected objects and move them” with shortcut SHIFT-D.
To make several copies of an object, that method is slightly quicker and
is the standard way you will see objects copied on “how-to” videos
online. However, there are subtle differences between the two methods:
copy/paste will make duplicates of materials you have in any slots on
the copied object, but duplicate won’t.

1. Select then move the manikin you created earlier down to stand
on the terrain in the lower domes

2. Copy/paste and rotate it, so you have two characters in

conversation.

Viewing the Scene in Camera Walk Mode

As you did in Chapter 1, it’s really useful to “fly” around using the
camera walk mode (SHIFT-`) with viewport shading still set to render
to see what the scene will look like. To be honest, it’s not very
convincing! One way to make it more realistic would be to add some
atmosphere to the lower domes to differentiate them from the low-
density air outside.

40
 USE MEANINGFUL NAMES

Now is a good time to correct something I ought to have told you to

do right from the beginning: change the names of objects from Plane,
Plane.001, Cube.003, and Icosphere to more meaningful names like
Terrain, Person01, Lower_domes, and Upper_domes. To change them,
double-click a name in the Outliner Editor and enter the new name. A
menu option and shortcut are available for renaming objects, but the
double-click system applies to renaming materials, textures, vertex
groups, and so on. Giving objects sensible names as soon as they are
created is an important habit to get into.

Another way to improve the rendered scene is with texture and

scattered rocks on the terrain. You can do that using geometry nodes,
which was used by Quick Fur in Chapter 1, and will be described in
Chapter 14, or particle systems, which we'll cover in Chapters 4 and 5.

Volume and Surface Materials

One fortuitous feature of the Wireframe modifier is that you can
use it to create a frame but at the same time as keep the original mesh,
assigning a different material slot to each. In this case we'll use a
surface material for the frame so it appears to be metallic, but we'll use
a volume material with no surface material for the original object. This
will result in light passing through the domes but scattering from the
volume within, just as might happen with a glass house containing a
damp atmosphere.

1. Select the lower domes object, then the Modifiers tab, and look
down at the options on the Wireframe modifier you added
previously (Figure 2-15). One checkbox defaults to Replace
Original object when the wireframe is created, so un-check that.
Another option, Material Offset, allows one material to be used
for the wireframe and a different one for the original object.
This enables you to set a volume shader to give the effect of
light scattering from an atmosphere but still keep the frame as a
solid surface, so set this to 1.

41
Figure 2-15 Wireframe modifier

2. Switch to the Material properties tab (pink and black sphere,

shown in the bottom center of Figure 2-16), add a new material,
and rename it to atmosphere. Add a New Material Slot with the
+ button (top right of Figure 2-17), then with the drop-down
beneath the list of slots, Browse Material to be linked, select
the material you used for the top dome and corridor, which
should still be called Material.001. Rename it to alloy_frame.
With the atmosphere Material Slot selected, change the default
Surface shader from Principled BSDF to Remove and the
Volume shader to Volume Scatter as shown in the composite
Figure 2-16.

Figure 2-16 Change Material Surface shader to None and the Volume shader to Scatter.

Changes made in the Material Properties tab are reflected in the

Shader Editor area and visa versa. This is the reason there is no option
“None” available when you select a shader. If you view the nodes in
the Shader Editor, it is clear that there are two options: the node can be
deleted from the graph or else the link connecting the node to the
Material Output node can be disconnected.

42
 Adjust the color and density of the atmosphere material in the
Material properties tab, and if you need to further refine the frame's
Base Color, Metallic, Roughness, or other inputs, select the material in
the second slot to adjust those settings. To speed up the process of fine-
tuning the shader settings, you can define a small render region.

1. Select the Output Properties tab (white printer icon, second

from top in Figure 2-16) and tick the Render Region checkbox.

2. Toggle to camera view in the 3D Viewport, then set a small

rectangle to render: select View4View Regions4Render
Region.

3. If you’re not trying to get an absolutely photographic result,

switching on Denoising and reducing the number under
Sampling in the Render Properties tab is often worthwhile.

NOTE In "The Wireframe Modifier" on page 39, I suggested you switch to

use Cycles rather than Eevee rendering. Eevee is the default,
generally faster method, and it is possible to use it for volume
shading. However, as of Blender 3, Eevee can be applied only to a
bounding box, not the actual 3D shape of the object, so it would be
incapable of doing what we want here. Try rendering a small
region to see the difference between the two render engines.

4. Adjust the Surface Color for the background in the World

Properties tab (pink globe icon) to a dark blue.
I know the photos sent back by the various Mars rovers show a
surprisingly Earth-like sky, but I decided the contrast of the dark sky
with the volume shading of the habitat looked better. An alternative you
could experiment with would be to set the Surface Color to Sky Texture
type Nishita, and fix the sun elevation and strength as well as air, dust,
and ozone density to create a generally dark sky with a slight glow near
the horizon.

When working with materials, keep in mind that objects have a list
of “slots,” each capable of holding a material, but the materials
themselves are held centrally so multiple objects can use them. The
pool of available materials is accessed from the little drop-down below
the list of slots (Figure 2-16). In the next chapter, you will see how
different materials can be applied to different groups of vertices or
faces.

43
Conclusion

You now should have an inkling of how useful modifiers can be.
You should know where to find them and understand what at least a
few of them can do. In the remaining chapters of this book, you'll use
modifiers to do a host of jobs, and it’s no exaggeration to say that their
mastery is crucial to becoming a competent Blender user.

While getting to grips with modifiers in this chapter, you also

encountered some unrelated, but important, snippets of information
about materials. Not only is is possible to reuse the same material for
several objects, but each object also can utilize multiple materials. The
inner workings of the Wireframe modifier hid the details, but in the
next chapter, you'll explicitly assign different materials to different
parts of the mesh.

You also applied a fundamentally different type of shader to the

volume of the material output and disconnected the surface shader
completely. I mentioned in passing that the node view reflected the
relationships between the different components of the material in a
more logical way. In later chapters, you will work with more
complicated materials, and you will come to appreciate the benefits of
the node view of the Shader Editor.

In the next chapter, you will use some Blender objects that are not
meshes. Instead of being constructed from vertices, edges, and faces,
you will use curves and surfaces that have been defined by curves.

44
 3
NON-MESH OBJECTS

Exploring surfaces and curves

defined by functions

In the first two chapters, you learned how to switch from

Object view to Edit view, and when editing you could see
that each object was constructed from vertices, edges and
faces. Objects of this type are meshes, but when you
added them you will have noticed that Blender has many
other types of object beyond meshes, cameras and lights.

45
 In this chapter you will add curves and text to your repertoire and
find alternative, better, ways to create some of the objects you worked
on in the first two chapters

Curves

Firstly a word of reassurance: In the chapter subtitle I announced

that curves were defined by functions, but those functions are internal
to Blender and you don’t need to do anything mathematical in order to
use them! Think of the difference between meshes and curves as
similar to the difference between dot matrix fonts and TrueType fonts.
When you zoom in on a mesh the change in angle of edges at each
vertex becomes more and more noticeable but with a curve, as the
magnification is increased the resolution can also increase so that it
always looks smooth.

Convert Text to Curve

To get a bit more of a feeling for the relationship between meshes
and curves, and to witness the simplicity of converting one type of
object to another, you will generate a curve using text. Later in this
chapter you will learn how to manipulate Text objects but for the
moment you are just using this as a quick way to create complicated
curves.

1. Start Blender and delete the default cube. then select

Add4Text which will create a Text object with the word
“Text”, notice that the icon in the Outliner Editor is a letter “a”.
Convert the text with Object4Convert4Curve which will
appear to do nothing but the icon in the Outliner Editor will
have changed to a curve. Switch to Edit mode and select the
Object Data Properties tab which looks like a green curve
connecting two little squares, visible at the bottom of
Figure 3-1.

46
Figure 3-1 Editing Curves

2. If you zoom in, you will begin to see that the curve is
constructed from straight edges which would seem to contradict
my earlier explanation. However in Blender, rather than the
resolution changing automatically with the view, it can be
adjusted to suit the needs of the user. See how it changes as you
increase and decrease the Resolution in the Object Data
Properties tab. When you subsequently convert a curve into a
mesh this value will control how many vertices are generated
between the control points.

Control Points

3. To see how the curve is controlled by the different types of

control point, select a control point in the middle of a curve then
select a different one at a sharp corner, such as the ones selected
in Figure 3-1. You will see that handles (the lines controlling the
shape of the curve) are different colors: pink is used for Aligned
which is the default type, red is Free and yellow is Vector, you
can change the handles using Control Points4Set Handle
Type.

4. Experiment with moving, scaling and rotating the control points

and the ends of each handle to get a feel for how you can use
them to obtain the required curve. Because the curve is set to
2D all the transformations are confined automatically to the XY
plane. If you toggle 3D the fill will disappear but,
unsurprisingly, you will be able to form the curve into a 3D
shape.

47
 Beveling Curves

Your first real exercise of this chapter is to construct a ring and a

mask woven from gold wire each one made using curves. You will
learn how to create realistic reflections and lighting using an
Environment Texture as the background - much simpler and more
realistic than the synthetic window you constructed in Chapter 1. The
final render should look like Figure 3-2. Links to the blend files and
videos for the exercises in this chapter are at get-into-blender.com
3.NON-MESH OBJECTS

Figure 3-2 Golden rings and a wire mask

You will need two curves to define the ring. One for the axis of the
shape, which will be a simple circular curve, and one for the profile, or
section, to be extruded around the circle.

1. Start a new file and add two curves to define the ring. Select
File4New4General (there is no need to save your work),
delete the default cube then select Add4Curve4Circle for the
ring and Add4Curve4Bezier for the profile.

2. Modify the profile curve to suit your ring design. With the
Bézier curve selected switch to Edit mode and set the view to
orthographic from above, the Z direction. Now create an
interesting profile for the ring. I decided that it would be a token
of love given to a steam engine enthusiast so I made it in the
shape of train track but you can choose anything you like,
presumably the inside should be smooth against the wearer’s
finger. Use E to extrude the end control point then G, R and S to

48
 form the shape. My rail section is visible in the center of
Figure 3-3.

3. In order to make the ring solid you need to join the ends up
using Curve4Toggle Cyclic. When you are happy with the
profile, still in Edit mode, position and rotate the curve so the
outside half of the profile is to the left of the Y axis and the
inside half to the right, as in Figure 3-3.

Figure 3-3 Profile along the length of a Curve using an Object for the Bevel

4. To make one curve use the a different curve as a profile, you

need to specify it as a Bevel. In Object mode select the Bézier
circle and in the Object Properties tab open Geometry then
Bevel and select Object then choose the Bézier curve, shown
on the right of Figure 3-3.

Making 3D Text

You now have a 3D surface in the shape of a ring but, to make it

really special, add a message on the outside.

5. Create the text as a new object. Select Add4Text then switch

to Edit mode, delete “Text” and replace it with a word of your
choice.

6. You can make the text bend round a curve using a modifier.
Back in Object mode open the Modifier Properties tab and Add
Modifier4Curve choosing the Bézier circle as the Object.

7. Now give the text some depth. In the Object Data Properties tab
(green “a” icon) under Geometry, increase the extrude amount.

49
 8. In order to orientate the text correctly open the Object
Properties tab (yellow square, just above the wrench) and set X
rotation to 90 and Z rotation to 180 and tweak the Z and Y
Location and the Scale so that the text protrudes on the outside
but not on the inside.

9. When it all looks good select Object4Convert4Mesh

Correcting Normals on Curved Surfaces

If you look at some of the letters with curved edges you will see
that the combination with the curvature around the ring has produced
ugly edges running across the letter, even if you change to
Object4Shading Smooth the effect is still visible and, possibly, made
worse. You can improve this a lot by:

10. Switch to Edit mode, select all vertices then Face4Beautify

Faces. The rearrangement can be seen clearly in Figure 3-4

Figure 3-4 Beautifying an ugly arrangements of faces

The problem of ugly lines running across faces that should be flat,
or very gently curved, is a perennial issue with 3D modeling, especially
with hard surfaces such as needed for modeling machinery, furniture or
jewelry. The general solution is to add extra vertices next to sharp
edges to confine the effect of the curving surface normals to the part of
the mesh near the edge, often by adding a bevel.

50
 NORMSLS

The normal represents the direction that the surface is facing at any any given
point which effects shadows and reflections. When Shading Flat is selected then the
normal direction is the same for the whole of each face and there is a sharp change
at the edge. Shading Smooth, on the other hand, makes the normal change gradually
across each face so that there is no sudden change at the edge.

One thing to note here is that when you extrude a 2D curve then
convert it into a mesh, as you just did with the text, Blender creates
separate meshes for the front, back and sides. The benefit of this is that
it solves the problem of the normals softening the edges when using
smooth shading but, as I mentioned in the last chapter, it produces
unrealistic looking sharp edges.

11. In order to see the effect of the following changes, set shading
to smooth. In Object mode select Object4Shading Smooth.

12. In order to add a controllable bevel to your lettering you first

need to merge the meshes together. In Edit mode select all the
vertices then Mesh4Merge4By Distance which should
report, at the bottom of the screen, that a number of vertices
have been deleted.

13. Back in Object mode add a Bevel Modifier and increase the
Segments to 3. Figure 3-5 shows the three stages moving from
split edges to continuous mesh to a beveled edge.

Figure 3-5 Joining split edges and adding a bevel

As it stands the text and ring are not joined together which makes
it inconvenient to move and rotate them in the scene. But there is a
problem; in order to join them as you did with the ring and the mount
in Chapter 1, or to use a boolean modifier as you did in Chapter 2, you
would have to convert the Bézier circle into a mesh.

51
Parenting Objects

The method you will use here to effectively combine the two
objects is to make the Bézier circle into the parent of the Text object.

14. Select both objects, but in the right order. In Object mode select
the text then select the Bézier circle as well using SHIFT-LMB.
The order of selection is important here: the Text should be
highlighted in a slightly more orange color than the Bézier
circle object.

15. Now set the parent child relationship. Click

Object4Parent4Object and you will see in the Outliner
Editor that the text object has moved inside the hierarchy of the
Bézier circle object. When you move or rotate the ring, the Text
stays attached.
This is the alternative method I referred to in Chapter 2 when you
applied the Boolean modifiers to make the coffee cup. You could
instead have parented the smaller cone, the torus and the cylinder to the
larger cone, then hidden them in the Outliner Editor. This technique is
more flexible but at the expense of complexity and the potential to
inadvertently move a carefully positioned object.

16. Add a gold material following the same steps you used in
Chapter 1. Because the ring and the text are two different
objects you will have to specify the material for both of them
but as you can select the same material for multiple objects you
only need to set up the material properties once.

Drawing a Curve onto a Surface

To construct the mask you will make a mold in the rough shape of
a face then draw Bézier curves onto the surface.

1. Create the mold object. Move the ring to one side for the
moment and select Add4Mesh4Ico Sphere, scale it into a
flat oval shape then, in Edit mode, move vertices to form a ridge
for the nose, brow and lips and indents for the eyes. It only
needs to be very approximate.

2. Add a new curve object and set it up ready for drawing wires.
Create a new Bézier curve object with Add4Cuéérve4Bezier

52
 Curve and in the Object Data Properties tab under Geometry
and Bevel (the right hand part of Figure 3-6) select Round for
the Bevel geometry with Depth set to 0.4 and Fill Caps on.

3. Each time you draw will create a new length of wire so you
don’t need the original three control points and can delete them.
Switch to Edit mode, delete all the existing vertices then select
the Draw tool from the Toolbar. To make the lines draw onto the
surface of the icosphere you need to open the Sidebar and in the
Tool tab set the Depth to Surface. Draw by LMB and dragging
the mouse as in Figure 3-6.

Figure 3-6 Setting Draw Curve onto a Surface

4. Draw curves in alternating directions to give the impression of

weaving and, for added interest, alter the Radius of some of the
wires. If you just change the Depth of the Bevel you will see
that all the wires change, which is not what you want. However,
in the Toolbar you can select the Radius Tool and change the
size of specific control points as shown Figure 3-7.

Figure 3-7 Changing the Radius of control points

53
 5. To change the thickness of several wires you need to select one
control point on each wire you want to change then expand the
selection to all the control points on each wire by pressing
CTRL-PLUS several times, or better, Select4Select Linked .
This process is sufficiently repetitive that you will quickly start
using the shortcuts CTRL-L for selecting and ALT-S for scaling
the radius.

6. Hide the icosphere you used as a former, which you can do by

un-checking the eye and camera icons in the Outliner Editor.
The Outliner Editor has several optional columns that control
which objects are visible on the 3D Viewport or appear in the final
render, as well as more complicated features required for compositing
such as holdout, which you will learn about in Chapter 13. The drop-
down menu for setting these is called Filter and has an icon that looks
like a funnel in the Outliner Editor header.

NOTE On occasion there can be too many buttons and options to fit into
the header, for instance if you move the divider of the 3D Viewport
to the right, or when using Sculpting mode or Texture Paint mode.
When this happens hold the mouse pointer over the header and
scroll using the mouse wheel.

7. Set the Material to the same gold you used for the ring.

8. Add a Cube object as a display plinth with a suitable Material.

If you look at Figure 3-2 you will see that I adjusted the
Subsurface value and Color as well as the Roughness to give a
plastic effect.

Environment Lighting

The final touch, which you can also see in Figure 3-2 is to use an
Environment Texture for the background and set this to provide the
illumination and reflections for the scene. As a rule, using the
background texture as illumination gives a much more realistic effect
than even complicated lighting arrangements, and it’s easy to set up.

9. Open the World Properties tab of the Properties Editor (pink

globe icon) and click on the little yellow dot next to Color under
Surface then select Environment Texture

54
 10. Now you need to use the Open button to find a suitable image
file. If you are on Windows and used the installer then browse
to C:\Program Files\Blender Foundation\Blender 3.03\,
otherwise to the location where you unzipped the Blender
download. Inside that folder browse further to
3.03/datafiles/studiolights/world/interior.exr where the first
directory will the same as the version of Blender that you
downloaded, in this example 3.03. There are a few textures
included with Blender but you will be able to find many online
by searching for “HDRI environment”. Software is also
available that will allow you to create a bespoke background by
stitching together your own photos.
The simplest way to get the full effect of the the shadows and
reflections is to use Cycles rendering but you will probably find that,
even with denoising switched on, you need to increase the Sampling in
the Render Properties tab to get rid of the “fireflies”. When you are
happy with the overall effect, render the scene and save the blend file.

Surfaces

There are several ways to use Curves to generate graceful surfaces

very efficiently, and for the rest of this chapter you are going to try
three different approaches to model the shape of a sailing boat hull and
sail. But, before you launch into the exercise, search online for “lofting
sailing boat hull” and get an idea of the shape you are aiming for. It
doesn’t have to be precise but essentially the profile should have a very
narrow V shape at the prow, opening out to a fairly flat bottomed U
shape in the middle and stern.

Curve Tilt and Radius

For the first approach to forming a hull you will try using a
variation of the method you used for the ring, but adjusting the “Tilt” as
well as the radius of each control point to create a twisted shape.

1. First of all add a profile curve to use as the bevel object. Save
your last work and start a new blend, delete the default cube and
add a Bézier curve, renaming it “rib”.

2. Shape the rib so it looks like half of a boat section, one end
needs to be exactly on the origin. Switch to Edit mode and view
in orthographic mode looking from the Z direction. Move one

55
 control point to the origin with X and Y both zero using grid
snap absolute. Remember you can use CTRL while moving to
do this very easily when you need to position a single vertex to
a grid point.

3. Move the other point so it is above and left of the origin then
rotate and scale the points so the line forms a smooth curve. You
can see the rib curve at the top left of Figure 3-8.

4. Create a curve for the keel and orientate it so that a mirror

modifier in the X direction will create the full hull shape. In
Object mode add another Bézier curve, rename it “keel” then in
Edit mode, with both Control Points selected rotate it so it lies
along the Y axis using R, Z, 90 then LMB followed by R, Y, -90
then LMB If you rotated the rib or keel in Object mode then
you will find everything behaves rather strangely but you can
remedy it by Object4Apply4All Transforms.

5. With the keel selected, in the Object Data Properties tab, set the
Bevel to use the rib object.

6. Now you need to modify the size and rotation of the bevel to
make it look like a boat. In Edit mode select all the control
points and Segments4Subdivide so you have four or five in
total. Try to tweak each point to form a boat shape by adjusting
the Radius, Tilt, Z position, X rotation and Scale.
Figure 3-8 shows my attempt. You will probably find that this
method, though good for some shapes, doesn’t provide the flexibility
required for a boat’s hull.

Figure 3-8 Using Tilt and Radius to form a boat shape

56
 Don’t struggle on for long, the next method will give you more
flexibility to design surfaces of this type.

Nurbs Surface
For the second attempt at forming a boat hull you will use a nurbs
surface. This can be very powerful, but getting the shape you want
takes a bit of practice.

1. Delete the rib and keel objects then add a nurbs surface with
Add4Surface4Nurbs Surface. Switch to Edit mode and you
will see a relatively small surface suspended inside a frame of
sixteen points.

2. It will be easier to construct a shape for half the hull (you will
add a mirror modifier) if you make the surface reach right to the
outer control points. Open the Object Data Properties tab (green
surface icon near the bottom of Figure 3-9) and under Active
Spline tick the Endpoint U and V options.

3. Now move all the points to one side of the origin by selecting
all the points and moving them to position one edge of the
surface along the X = 0 plane. Move them approximately first,
then select just the points on the mirror plane and set their X
value using grid snap absolute then move along the Y axis while
pressing CTRL.

4. It’s a good idea to add the mirror modifier now before you start
moving the control points so can see the finished shape as you
work.

5. Move the control points to form the hull. The relationship

between point locations and surface shape takes a bit of getting
used to, so don’t alter the Y positions of the points much, apart
from at the top and bottom of the bow. To appreciate the 3D
shape as you create it I suggest staying in perspective view and
moving points with G then SHIFT-Y.
You will find it easier to make a streamlined shape using only the
sixteen control points of the starting nurbs surface but in figure 3-9 you
can see that, after forming the basic shape, I added some more points
by extruding the edge twice in the V direction (the pink lines running
across the boat) in order to flare the gunwale slightly and wrap the
surface over to form side-decks. Leave the stern open, you will fill it
after the surface has been converted to a mesh.

57
Figure 3-9 Boat shape using Nurbs Surface

6. When you’re happy with the shape, use

Object4Convert4Mesh which will also apply the Mirror
Modifier.

7. To form the cabin structure, in Edit mode select the edge loop at
the inside edge of the side-decks using SHIFT-ALT-LMB. In
orthographic view with X-Ray on deselect the rear half of the
vertices by pressing CTRL while box selecting them. Extrude
upwards, scale inwards and rotate slightly about the X axis. You
can see in Figure 3-12 the proportions I used.

8. Create a new face for the cabin roof by selecting all the vertices
that define the edges (which should already be selected after the
previous step) then select Vertex4New Edge/Face from
Vertices. This is such a common action that you will almost
always use the shortcut F. This will create a single face which is
often fine, you will look more closely at the different tools for
building mesh topology in Chapter 7 but for the moment keep it
simple – the boat will be too far away to see much detail in the
final render.

9. Add the transom by selecting the edge loop of vertices around

the open back of the hull and filling with a single face as you
did for the cabin roof. This might result in some messy edges
where the side-decks fold over but, for this exercise, you can
leave them as they won’t be visible.

58
 Bridging Edge Loops
For the mast and sail you will convert a series of Bézier curves to
meshes then join them using Bridge Edge Loops. At this point you can
probably see that it would be very efficient and easy to make a curved
and tapering mast from a Bézier curve with a round bevel and a
realistic sail from a nurbs surface. However the technique of bridging a
series of edges can be useful in many different places, in fact it would
have allowed a method of construction similar to the “lofting”
drawings of sailing boat hulls that you found online. You could have
constructed a series of ribs, each with a different shape but the same
number of vertices, with equal spacing between each rib, then built a
surface joining them all together.

10. In Object mode add a new Bézier curve and rename it “sail”.

11. One end of the sail curve will become the section of the mast.
Use the orthographic view from the Z direction and switch to
Edit mode then extrude the end control point twice so you can
form the mast.

12. Change the control point where the sail meets the mast so the
curve can bend sharply using Control Points4Set Handle
Type4Free as shown in the detail of Figure 3-10.

13. Duplicate the curve twice, and move the copies up to the mid-
point and the top of the sail. Scale, rotate and flatten the sail
profile appropriately: the top of the sail will be narrower, flatter,
and twisted more relative to the bottom. You will need to keep
swapping the view between the Z and Y direction as you do
this.

Figure 3-10 Sharp bend between sail and mast using Free Handle type

59
 14. The curves need to be converted into three one dimensional
meshes, essentially edge loops, before they can be bridged to
form a surface. In Object mode run Object4Convert4Mesh
then back in Edit mode select all the vertices then use
Edge4Bridge Edge Loops.

15. In the Adjust Last Operation pane you can increase the Number
of Cuts so the generated surface is a bit smoother.

16. Finally scale and rotate the sail then position it on the boat,
setting the boat as the parent of the sail as you did for the text
on the ring earlier in this chapter.

Multiple Materials per Object

The default white material might not be a bad choice for a

fiberglass hull or a sail but the overall impression would be much
improved by differentiating the mast and the side-decks.

17. With the boat selected switch to Edit mode and select just the
vertices for the side-decks - this is a good instance where
selecting Edge Loops would be very quick and efficient.

18. You now need to assign a material to a selection of vertices.

Still in Edit mode open the Material Properties tab then, if you
have a Material Slot with a Material defined double-click and
rename it “deck”. If there is no Material yet, add a new Slot
then add a new Material and rename that. If you compare
Figure 3-11 below with Figure 1-5 you will see that there is a
line of buttons for assigning and selecting vertices by Material
which is only visible in Edit mode. With the deck material
selected in the list of slots, click Assign to associate the
Material to the selected vertices.

Figure 3-11 Assigning Materials to Vertices

60
 19. Adjust the Base Color, Roughness and other shader parameters
to suit the scrubbed teak look of the deck Material.

20. Now create a different material to assign to the parts of the

boats that are not deck. Add a new Material Slot and a new
Material and rename it “fiberglass”. With your cursor back over
the 3D Viewport select the cabin roof and hull by using
Select4Invert, however if you are in vertex selection mode as
in Figure 3-11 this will leave a gap between the vertices
selected for the deck and the fiberglass. So change to face
selection mode, deselect all faces with ALT-A then in the
Material Properties tab select the deck material and click
Select. Now when you invert the selection there will be no gaps
and you can select the fiberglass material and click Assign with
the correct faces selected.

21. Reduce the roughness of the fiberglass shader to make it look

like plastic.

22. Go through the same process to allocate Materials for the sail
and mast.

Ocean Modifier

To make suitable surroundings for your boat you can quickly

generate some waves.

1. Create a plane to hold an ocean modifier. In Object mode use

Add4Mesh4Plane, rename it “ocean”, then add the Ocean
Modifier to it. In the settings increase Repeat to 5 in the X and
Y direction, change the Size to 2.0, the Spacial Size to 20 and,
under the Waves section increase the Choppiness to 3.0. Note
that you don’t need to scale or subdivide the plane itself so long
as the Geometry is set to Generate, the modifier does that for
you, indeed you could could have added any type of mesh.

2. Add a Material for the ocean, change the Base Color to a dark
blue-green, reduce the roughness and increase the Transmission
to 1.0

3. Check the view from the Camera and move the ocean so no
edges are visible. Pose the boat so it looks suitably dynamic and
rests on the surface. Alternating rotation using R and RR is a
useful technique for this.

61
 Rendering with Eevee

Because the scene lacks a lot of detail such as rigging on the boat
or surface foam on the sea, this is a good example to try using the
Eevee render engine, which is less photo-realistic but much faster. You
can also add some atmosphere using environment lighting, haze and
camera depth of field.

4. Use an environment texture again for background and lighting.

Delete the default point light then add an outdoor environment
texture, as you did for the ring and mask exercise. In
Figure 3-12 you can see that I used
3.03/datafiles/studiolights/world/sunrise.exr but turned the
Strength down.

5. To add a slight haze as the waves recede into the distance, while
you have the World Properties tab open, set a Volume Shader to
Volume Scatter with Density at 0.02.

6. When using Eevee, in order to enable the boat reflection on the

water, in the Render Properties tab tick Screen Space
Reflections.

7. Finally, to make the scene slightly more photographic, select the

Camera object and in the Object Data Properties tab (green cine
camera icon) enable Depth of Field, reduce the Aperture F-Stop
to around 0.1 and either manually adjust the Focus Distance or
use Focus on Object.

Figure 3-12 Final render: haze at dawn

62
 Using Eevee can often produce good results with only a slight
reduction of realism and taking a fraction of the time of a Cycles
render. For some projects, such as animations, that’s exactly what’s
wanted. There are several other optional features to improve render
quality: Screen Space Reflection, Ambient Occlusion and Bloom can
be turned on in the Render Properties tab.

Try opening the blend file for the ring and mask that you saved
earlier and change the render engine to Eevee, then experiment with the
optional features in Render Properties to see which ones make a
difference and how near you can get to the Cycles result.

Conclusion

Curves provide Blender with a quick method to generate graceful

shapes that would be hard to make using mesh tools alone. In this
chapter you very quickly dabbled with a few applications of curves and
saw how they could be created from Text objects as well as being
converted to meshes. As you cover different topics in this book you’ll
also find other uses for curves where they are used as “guides”: in
Sculpt and Paint mode they can define the stroke of a brush, the Curve
Modifier can deform a mesh to align with it, as you used with the text
on the ring in this chapter, and when animating, constraints can be set
so that objects or bones slide along or deform to match curves.

When you converted the lettering to a mesh you saw some of the
problems that can arise when trying to make large smooth surfaces with
realistic edges and you used some of the techniques available in
Blender to improve them.

You learned how different Materials could be allocated to specific

sets of faces within a mesh and, for the final render of both exercises,
you used an Environment Texture to provide the background, lighting
and reflections for your scene. Both these techniques can be very
effective for adding realism, as are the other two modification you
made to the standard setup: Camera Depth of Field and Volume Scatter.

Finally, when you added the Ocean Modifier to the plane you will
have noticed that there are still a lot more Modifiers I haven’t
mentioned yet. In the next chapter you will use two of these, the Array
and the Curve Modifiers, to great effect. You will also encounter
“empty” objects and, for the first time, use photographic images to
texture the surfaces of the models you construct.

63
 4
A R R AY S , M AT E R I A L S A N D
TEXTURES

Making patterns of an object from a

single instance

Blender has a surprising number of ways to generate

multiple copies from a single object: Array Modifier,
Instancing, the Spin Tool, Particle Systems and Geometry
Nodes all allow this to be done. In this chapter you will
use the first four.
In Chapter 2 you tried out several modifiers from the second of the
four columns in the drop-down list but you may have noticed that I

64
 skipped over the first one in the list; Array. The reason for delaying its
introduction till now isn’t because the array modifier is less useful than
the others, but there are some subtleties to its use that deserve more
explanation.

The exercise for this chapter will be to construct a rusty steam-

punk refrigerator and then to form an enormous sculpture by stacking
different circular arrangements of the refrigerators in a parody of a
Megalithic structure. The final result should look something along the
lines of Figure 4-1

Figure 4-1 Scrap Henge

Downloading a Texture Asset

Before launching into the modeling process on Blender you need

to get hold of an image file that you can use to texture the surface of
your refrigerator. I used the phrase “free download rusty metal
seamless texture” in an online search and found lots. The “free
download” part is to help find images that are not subject to tight
commercial conditions and restrictions, but read the terms and
conditions anyway when you find a good looking website. “seamless”
just means that the image can be tiled edge to edge without showing a

65
 discontinuity at the joins. Of course you could just take your own photo
and make it seamless with almost any image editing software. Save the
image file in the folder system where you intend to store the blend file
for this chapter’s exercise.

The ability to find resources online is definitely one worth

cultivating, but if you don’t find anything to you liking you can
download the image I used from get-into-blender.com 4.ARRAYS,
MATERIALS AND TEXTURES

A Material with Multiple Shaders

Making a convincing scrap refrigerator Material will be your first

task. Using the rust image for the base color gives the impression of
mild steel plates, but what we want is the original white or silver finish
with rust showing through around the corners and edges, which get
more wear. One way of doing this would be to assign different
materials to different vertices as you did for the boat in the last chapter,
but that gives a hard division between the two Materials instead of an
uneven pattern where the rust is gradually eating into the original
finish. A better approach is to create a single Material with two
Principled BSDF Shaders which are combined together using a Mix
Shader where the factor for mixing will take into account the
“pointiness” of each vertex as well as having some noise added to
make it look more natural. Links to the blend files and videos for this
chapter are at get-into-blender.com 4.ARRAYS, MATERIALS AND
TEXTURES

8. Start Blender, leave the default Cube selected and change the
workspace to Shading using the tab right at the top of Figure 4-
2.

9. You will have the default Material visible as nodes in the

Shader Editor area, rename it now to “rusty_fridge” either in the
Material Properties tab or at the top of the Shader Editor.

Bumps Using the Normal Input

Your material will have two shader outputs, one for the rust and
one for the smooth surface. The rust shader will need the rust image for
the basic color but you can also use the image to create a pseudo
bumpiness. In the Shader Editor you will add three nodes and connect

66
 them to the Base Color and Normal of the default Principled BSDF
Shader as I have in Figure 4-2

10. Add a node to supply the rusty texture. Select

Add4Texture4Image Texture then click on the little folder
icon and browse to the location of your rusty metal image.

11. The brightness of the of the texture will be a stand-in for the
height of the bumps, this can be extracted using a different
node. Use Add4Converter4Separate HSV and connect the
Color output of the image texture node to the input.

12. The brightness values need to be converted into a normal map

using a bump node so use Add4Vector4Bump. Connect the
Value output of the HSV node to the Height input, and the
output to the Normal input of the shader. This synthetic bump
image uses the value output but you could experiment with the
hue or saturation, alternatively there is a node to convert to
RGB and it might be interesting to try each of those outputs,
choose whichever gives the most interesting result.

Figure 4-2 The first part of the rusty_fridge Material

67
 The Mix Shader

13. Once you’re happy with the basic rust, add another Principled
BSDF Shader which will become the non-rusted part of the
material. Also add a Mix Shader to combine them.
If you add the mix shader first you can “drop” it onto the pipe
connecting your existing shader to the Material Output node, and this
will automatically rearrange the nodes for you. Make sure, when you
add the second Principled BSDF Shader that you land in a space not
overlapping any existing pipes, then connect its output to the first input
of the Mix Shader. You will see that the connection from the first
Shader automatically swaps over to the second input. The three shaders
are the nodes with green tops on the right of Figure 4-3

14. On the mix shader slide the Fac to 0.0 you should see a plain
white cube and when you slide to 1.0 it should be fully rusty. If
you see the opposite to this, drag the dots on the ends of the
input pipes to the Mix Shader to swap them as described above
Figure 4-3 shows the full layout of the nodes so you could use that
to position elements as you add them to save constantly moving things
round. But I advise you to add the components one by one, following
the steps below, so you see the effect of each - and the potential pitfalls.

Pointiness

15. You are going to control the mixing factor using the sharpness
of edges, which can be obtained from the geometry node. Select
Add4Input4Geometry then connect the Pointiness output to
the Base Color of the plain white Principled BSDF Shader and
set the mix Fac to 0.0.
You’re using the pointiness as the base color just to check what’s
going on - when you get the worn edge effect showing in black and
white you can connect the output to the Fac input of the Mix Shader. At
this stage there will be no visible edge effect for three reasons.

68
Figure 4-3 Full node tree for the rusty_fridge Material

The first is that the pointiness scale goes from 0.0 for a vertex that
is at the bottom of an acute hole to 1.0 for a vertex on the tip of a pin
point. For a vertex on a flat area the pointiness is 0.5 but for normal
edges and creases the pointiness will be very close to 0.5. So you need
to magnify the scale in the region around 0.5 and this is one of the uses
of the very versatile color ramp node.

The Ubiquitous Color Ramp Node

16. You need to add a color ramp node then move the stops so that
all of the variation in output take place over a narrow band of
input near to 0.5. You want the output to be maximum either
side and only drop to 0.0 where the vertices are flat. Select
Add4Converter4ColorRamp and drop the node onto the
pipe coming out of the Geometry node, use the + button to add
a third stop then slide the stops so one is on 0.5 and the others
are on either side, then change the colors using the color picker
at the bottom of the node so the sequence is white, black, white,
as you can just about make out in Figure 4-3.
This still won’t solve the problem because of the second reason; all
eight vertices on a cube have the same sharpness.

17. Subdividing will produce some vertices away from edges.

Switch to Edit mode and subdivide the cube with Number of
Cuts in Adjust Last Operation set to five.

69
 Frustratingly you still won’t see the desired effect for the third
reason: the Geometry node Pointiness is only available when using the
Cycles render engine.

18. In the Render Properties tab set the Render Engine to Cycles
and, at the top of the 3D Viewport, set the Viewport Shading to
Rendered. At last you should be able to see black sides and
white edges and corners.

19. At last you can connect the output of the color ramp node to the
Fac input of the Mix Shader.

20. To complete the Material you need to add some noise and
sharpen the cut-off between the original finish and the rusty
areas. Use Add4Converter4Math and drop the node onto
the pipe from the color ramp to the Mix Shader, it will default to
Add mode.

21. For the randomness here use Add4Texture4Noise Texture

and add the node in a space to the left. Connect its Fac output to
one of the Value inputs of the Add node.

22. A color ramp can again be used here to control the division
between rust and smooth surface. Select
Add4Converter4ColorRamp, drop the node onto the pipe
between the Add node and the Mix Shader then adjust the
sliders. The nearer together they are, the sharper the change
from white to rust and their position between black and white
will determine the amount of rust.

23. Change the parameters in the Noise Texture until you are happy
with the result, but you will probably need to make further
tweaks after the cube has been converted into a refrigerator.

Constructing the Prototype Object

With the material sorted out you can now move on to modeling the
refrigerator. The first modification will be lines of studs and
indentations on the refrigerator door, for which you will use an array
modifier.

70
 Array with Linear Offset

1. Change the Viewport Shading from Rendered back to Solid

2. You will use an array of cylinders to make the indents and studs
but, as they will be small, you should reduce the resolution from
the default 32. Select Add4Mesh4Cylinder then, in the
Adjust Last Operation pane, reduce the number of Vertices to
12, the Radius to 0.01m and the depth to 0.02m.

3. At this point the cylinder will be inside the cube so now would
be a good time to rename “Cube” to “fridge0” in the Outliner
Editor and, for the moment, click on the little eye icon to hide it.

4. To create an array of 8 cylinders, select the cylinder object, add

an array modifier, increase the Count to 8, un-tick the Relative
Offset checkbox and tick Constant Offset setting the Distance X
to 0.2m.

5. In the 3D Viewport rotate the object about the Y axis by 90

degrees. Note, this has done what we wanted it to do, you now
have a vertical column of horizontal cylinders, however this
may be a little surprising; the array offset specifies an X
distance so it would be reasonable to expect a row of cylinder
along the X axis, each one rotated by 90 degrees. You will see
later that controlling how the array is generated usually requires
more care than this.

6. Make the fridge0 more reasonably proportioned. It’s a 2m cube

at the moment so toggle the visibility back on and scale it down
along the different axes until it is the right size and move it up
so the bottom face is just on the origin. Making these
transformations in Object mode will cause issues later but do it
that way so you can see the problems and how to fix them.

7. To create a line of indents, using the orthogonal views and

move the cylinder object array so it overlaps the face of the
door in a line near the left side, then add and apply a boolean
modifier with Operation Difference to the fridge0 object using
the cylinder for the other object.

8. Repeat the process to form another line of indents towards the

right of the door.

9. To create a line of studs, reduce the Array Modifier Count to 4

and apply it near to the left hand line of indents but using a

71
 boolean modifier with type Union. Delete the cylinder object
after applying it for the last time.
Figure 4-4 shows how I arranged the indents and studs but
obviously you can use your artistic sensibilities to create whatever
pattern you like the look of: lines of rivets round all the panels or, if
you had reduced the number of vertices in the cylinder to 6 instead of
12, you could have made an even more steam-punk refrigerator with
lines of bolts.

Grooves and Edges Using Edge Loops

10. To make the groove running round the object to represent the
door seam use the Loop Cut tool in Edit mode. The way this
tool works is that it forms a highlighted edge loop at right
angles to whichever edge is nearest to the cursor. When you
hold down LMB you can slide the position of the edge loop
until you release LMB which fixes it. Position the new edge
loop a small distance from the first loop in from the front face
as shown in Figure 4-4.

11. You need three edge loops near to each other running around
the refrigerator so add one more using the Loop Cut tool just
next to the other two.

12. To form a notch running round the door, select the central edge
loop using SHIFT-ALT-LMB then scale it down a little bit.

Figure 4-4 The prototype refrigerator

The back of the refrigerator is going to sport a whimsical chrome

plated exhaust stack.

72
 13. To form the main straight section, in Object mode select
Add4Mesh4Cylinder then scale and move it so it is a small
distance from the back.

14. Connect the bottom of the pipe with a curved section. In

orthographic view from the side, switch to Edit mode and
extrude the bottom face of the cylinder and rotate and move it a
few times to form a bend, this is the same method you used for
the corridor in Chapter 2.

15. To make the silencer and final tube section extrude and scale the
top surface of the cylinder a number of times.

16. In Object mode create a new Material for the exhaust stack and
rename it “chrome”, setting Metallic to 1.0 and Roughness to
0.0.
You need to use smooth shading for the exhaust stack to make it
look cylindrical but when you do that you will see that the edges at
either end of the silencer become indistinct and unrealistic. This is the
same problem that happened with the Text object in the last chapter
which you solved by simply adding a Bevel Modifier. However,
because you will produce arrays in different ways using this
refrigerator prototype, you cannot Parent the exhaust stack to the
refrigerator, you must Join them together into a single object.

Bevel in Edit Mode

This raises the question of how to selectively bevel some edges of
an object but not others. For instance, if you add a bevel modifier to the
refrigerator object it will reduce the pointiness of the edges and modify
the rusty_fridge material. There are several solutions to this problem
such as assigning bevel weights to edges or using vertex groups.
However for this exercise you will simply modify the mesh.

17. In Edit mode, select the four edge loops at either end of the
silencer and bevel them using Edge4Bevel followed by LMB
then in the Adjust Last Operation pane set Offset to 0.05m and
Segments to 2.

Joining Objects with Different Materials

18. If you now Join the exhaust stack and refrigerator into a single
object, the materials assigned to different vertices will be

73
 preserved. In Object mode select first the exhaust stack, then the
fridge0 object and use Object4Join.

19. Save the prototype refrigerator as a blend file for a later chapter
when we will return to the topic of smooth shading, flat faces
and beveled edges.

Circular Array

For the first tier of the henge, you will use the third mechanism for
Offset positioning with the Array Modifier; Object Offset.

1. In Object mode select the fridge0 object and add an Array

Modifier, un-tick the checkbox against Relative Offset and tick
Object Offset.

Empty Objects
Although they sound slightly mysterious, empty objects are simply
a convenient way to hold information about location, rotation and
scale, for positioning other elements in the scene say, without the need
for an actual object.

2. You will see that this offset mode requires another object for
reference, so add an empty object to use to define the relative
offset. With the cursor over the 3D Viewport, select
Add4Empty4Arrows. As you can see there are a variety of
appearances for empty objects, they all do the same thing but
some shapes will make their use clearer later.

3. Now set the array offset object. Re-select the fridge0 object and
in the Object field select the empty object. You should see
something odd but to make it a bit clearer increase the Count to
8 and move the empty object a small distance in the Y direction
(see Figure 4-5)

74
Figure 4-5 Object Offset scaling issue

The reason this happened is because the Object Offset array uses
the difference between the Object properties of the fridge0 and the
empty. If you want the array to stay the same size you need to return
the scale of the fridge0 object to 1.0

4. To effectively convert all the changes made in Object mode as if

they had been applied to the vertices in Edit mode, select the
fridge0 object then select Object4Apply4All Transforms,.

5. In order that the refrigerators are set out in a ring, in

orthographic view from the Z direction, move and rotate the
empty object.

6. Now try selecting and rotating the refrigerator so the door faces
a different direction.
Because the Offset is based on the differences between the
rotations and positions of the two objects, the whole circle will change
as soon as you rotate the refrigerator!

7. The solution is to rotate the mesh in Edit mode. In the Object

Properties tab set the Z rotation back to 0.0 then, in Edit mode,
select all the vertices and rotate them about the Z axis.
Now you will see each of the refrigerators rotating nicely around
their local center point - in my experience this is the simplest way to
work with the Array Modifier; only use modification in Edit mode and
avoid any transformations in Object mode.

75
Arrays Along Curves

Now you will make a linear array of refrigerators then bend them
using a Bézier curve, but first, create three copies of the prototype
refrigerator to use for each of the higher tiers.

1. Duplicate the fridge0 object and move the copy up so it’s clear
of the first circular array.

2. Delete the Array Modifier in the copy so you have a single

refrigerator, then duplicate it twice.
If you check my render in Figure 4-1 you will see that the second
tier of the sculpture is constructed from an array of objects following a
spiral using the Curve Modifier, the third tier has five objects as
Instances, each at a vertex of a pentagon and the top tier has three
refrigerators constructed in Edit mode using the Spin Tool.

3. For the second tier add a Bézier curve object which the array of
refrigerators will follow.

4. To make the curve form about one and a half full turns, at a
similar diameter to the first tier circle, switch to Edit mode then
extrude one more control point. Move, scale and rotate each
point using different orthographic views. It won’t be possible to
make a very regular circle with only a few control points but it
doesn’t matter too much so long as the length of curve between
each point is about the same.

5. Arrange the circle while viewing from the Z direction then

adjust the heights and slopes of the control points from the X
and Y direction so there is a fairly smooth spiral.

6. First make a vertical array of refrigerators. Back in Object

mode, select the second tier refrigerator, add an array modifier
and increase the Count to 8. Use Relative Offset but change the
X Factor to 0.0 and and set the Z Factor so that each refrigerator
is separated from the next by a small gap.

7. Add a curve modifier, which needs to be below the array

modifier in the stack. Select the Bézier curve as the Curve
Object and set Z as the Deform Axis.

76
 8. To get the array to start and end on the curve adjust the Z
location of the refrigerator object in the Object Properties tab
and the Z spacing and Count in the array modifier.

9. You want to make the refrigerators look as if they are resting on

the first tier and on each other so adjust the positions of the
curve control points.
On a more general level, the technique of making arrays of objects
“line up” along curves is occasionally very useful; you can construct
cables, hoses and architectural features very quickly from a stack of
unit objects and a curve.

Instancing Objects

You will accomplish the third tier of the sculpture using instancing.
The method used here creates instances at each vertex of a parent
object.

1. First you need to add an object as the parent. Use

Add4Mesh4Circle then in the Adjust Last Operation pane
reduce the number of Vertices to 5.

2. Scale the circle up and move it above the second tier. Move the
third refrigerator approximately to its center.

3. Parent the refrigerator to the circle by first selecting the third

refrigerator then SHIFT selecting the circle. The order is
important here, the refrigerator should be highlighted in red-
orange and the circle in yellow-orange, then run
Object4Parent4Object. In the Outliner Editor you should
now have the refrigerator as a child of the circle.

4. Select the circle object then in the Object Properties tab (orange
square icon) under Instancing select Vertices and tick Align to
Vertices.

5. Move and rotate the circle object to rest the five refrigerators on
top of the second tier.

6. Rotate all the outside instances by selecting the central one and
rotating that about the Z axis. Note that, as with the array
modifiers, you will have issues if you rotate the object rather
than rotating the vertices while in Edit mode. The original,

77
 central refrigerator will be visible in the 3D Viewport but will
not show in the render.

THE BMW CAR DEMO

Instancing is a less frequently used feature of Blender. Though easy

to do for simple arrangements such as this, it does have limitations.
However, a close relative; making a collection into an instance to scene,
can be very powerful as it allows complex assemblies of components to
be quickly replicated in multiple locations. There isn’t space in this book
to use this technique in any of the exercises, however I recommend you
download the popular BMW Car Demo from
https://www.blender.org/demo/test, and check out the use of instancing
for the wheels (and the whole second car!) This demo also shows clever
use of subdivision and mirror modifiers, for instance the body shell
including details such as wing mirrors and door handles, only uses
1,765 vertices. Don’t worry at this stage if you can’t figure out how some
of the rather sophisticated modeling works.

Spinning Vertices

For the top tier of the sculpture you will utilize the Spin Tool in
Edit mode. On occasion this is a very convenient shortcut but it is less
flexible than the non-destructive array or instance methods you just
used. The result will be a single mesh so any tweaks you might need
later will have to be repeated on each of the refrigerators.

There are two different ways that the Spin Tool operates. If the
selected vertices form a mesh with no unconnected edges (you will see
this referred to as a manifold), it will create duplicate meshes at
intervals around a given angle. If there are unconnected edges it will
extrude the edges in a number of steps around a circumference.

For this exercise we want the first behavior but as an example of

the second: if you had converted the Bézier curve object from the first
exercise in Chapter 3 into a Mesh, you could have use the Spin Tool to
make it into a ring as shown in Figure 4-6.

78
Figure 4-6 Spinning a ring from a mesh

1. Select the top refrigerator object, switch to Edit mode then

select all the vertices and choose the Spin Tool from the
Toolbar.

2. First create some duplicates by dragging on one of the blue plus

buttons at either end of the blue arc (visible in the center of
Figure 4-6). It doesn’t matter how far you drag it because you
will set specific values afterwards.

3. Set the number of duplicates and the angle between them by

opening the Adjust Last Operation pane and change the Steps to
2 and the Angle to 240°. If there happen to be any unconnected
edges on your refrigerator (for instance if you used an open-
ended cylinder for your exhaust stack) then you will need to tick
the Use Duplicates checkbox.

4. To get the refrigerators orientated the way you want change the
Center X and Y values.

5. Position the top tier so it’s resting on the one beneath by

changing back to Object mode and moving and rotating it.

Scattering Objects over the Surface of Another Object

The setting for your sculpture will be a grassy mound with early
morning mist against a suitably rural background using an Environment
Texture. The grass will be scattered over the mound using a particle
system.

79
NOTE From Blender version 3, the standard way to scatter instances of
one object over another surface is Geometry Nodes which will be
covered in Chapter 14. However old versions of Blender can get
the same effect using particles.

Collections
Before you start adding the components, this would be a good
opportunity to introduce collections as an excellent way to tidy up your
project and help your workflow.

You used a collection with a boolean modifier to create the

Lower_domes object in Chapter 2 and they can also be used for
instancing as I mentioned above, however their normal use is as a kind
of folder system to help organize complicated scenes. A very real
benefit of using collections is the ability to toggle them off in the
Outliner Editor which allows you to keep alternative outfits for a
character, for instance, in the same blend file but switch between them
easily.

You will encounter another important function of collections in

Chapter 13 when you use the compositor. Blender can render different
parts of a scene as different layers for later reassembly with enormous
flexibility, and collections are the mechanism for controlling this.

For the exercise of this chapter you will simply use collections to
tidy up your workflow.

1. Make a new collection for the fridge arrays. In the Outliner

Editor RMB over Scene Collection then New Collection and
rename it “fridges”.

2. Create another new collection and rename it “ground”.

3. Drag and drop each of the refrigerator related objects into the
fridges collection. The five sided Circle, the empty and the
BezierCurve all belong inside, leaving only the Light and the
Camera where they are.

An Object to Hold the Particle System

The mound will be a simpler version of the crater you constructed
in Chapter 2, but before you add it, move your Camera so you frame

80
the sculpture in a pleasing way. You can see in Figure 4-1 that I
changed Resolution X in the Output Properties tab (printer icon) so that
I rendered a square image. That resulted in only a small amount of hill
being visible in the final render – there’s no point adding polygons to
the scene if they will never be seen.

4. Select the ground collection in the Outliner Editor then add a

plane to the scene. It will default to being inside that collection.

5. Rename the plane to “hill” then scale and subdivide it and form
a mound in the middle by moving a vertex upward with
Proportional Editing on.
The mound will be covered with grass so it doesn’t need to be very
high resolution or carefully crafted.

An Object to Distribute as Particles

The last method you will use for generating multiple copies of an
object is the Particle System. This is a very versatile feature but that
versatility comes at the cost of quite serious complexity, and in many
cases the settings that need to be adjusted are quite hard to find.

Before you add the Particle system to the hill object you need to
make a tuft to be scattered over the surface.

6. Add a plane inside the ground collection and rename it “tuft”.

7. Modify the plane to form a single blade of grass. In Edit mode

and in orthographic view from the Z direction move the
vertices, I subdivided one of the edges twice so I had a total of
six vertices to allow better shaping, see Figure 4-7.

8. When it looks OK select all the vertices and duplicate the blade.
When the tuft is duplicated by the particle instance the Y
direction will be vertically upwards so you need to rotate it
around the Y axis so the tuft is later visible from all directions.
Scale and move the duplicate to form a second blade.

9. Repeat six or seven times until you have a little tuft, as shown
in Figure 4-7. You will need to move some blades up and down
in the Z direction too to spread them out in two dimensions.

81
Figure 4-7 The Tuft object

10. To add a particle system to the hill switch back to Object mode,
select the hill object and open the Particle Properties tab (the
icon is a blue dot connected to three other blue dots). Choose
Add Particle System by clicking on the + button.

11. Particle Systems are really an animation effect so they are

designed to produce particles over a period of time, but you
want all of your grass to exist from the start without needing to
run the animation so under Emission set Frame End to 1.

12. The default particle is a glowing dot, so, to change that to tufts
of grass, you need to open the Render section then change
Render As from Halo to Object, make the Scale 1.0 and the
Scale Randomness 0.5. In the Object section below select the
tuft as the Instance Object.

13. You can generate more tufts by adding children. Open the
Children section and select Simple and increase the Radius
value to around 0.8 or until the Children objects spread
naturally into the spaces on the hill.

14. Create some random rotation as well by ticking the Rotation

checkbox then open that section and increase the Randomize
Phase value to 0.5 or more.
To increase the variety of the grass it would be good to add a
second, slightly different, object as well.

15. Duplicate the tufts object and rename it “stalks”.

82
 16. In Edit mode extend one of the blades so it is significantly taller
than the others.

17. Back in Object mode select the hill object again and add an
additional Particle System, renaming the two System slots and
associated Settings with an appropriate “tufts” or “stalks” name.
Note that the Particle System works a little like the Materials
setup, there are slots each of which can hold a Setting item, and
different objects can share the same settings.

18. With the stalks System selected go through the settings and
change the Render, Children and Rotation as you did for the
tufts but using the stalks as the Instance Object.

19. Reduce the Number under Emission to 100.

20. To increase the effect of variability, change the Children Render

Amount to 10, Size to 1.25 and Random Size to 1.0.

21. Add three Materials, one to the hill and one to each of the grass
objects, with different shades of green and brown. If the hill
becomes invisible once you add the particle systems to it then
you need to tick the checkbox Show Emitter in the Viewport
Display section of the Particle Properties tab.

22. Use the Render Region option in the Output Properties tab to
check the overall effect when rendered using Cycles. Adjust the
colors and Particle System settings until you get good looking
grass.

Background
In the World Properties tab choose a suitable outdoor Environment
Texture to use as the Color in the Surface section. I used the sunrise
image again and turned the Strength down to 0.2 but you can try
different options until you find something that shows off your sculpture
and grass to good effect.

Adding a Volume Shader

Back at the start of this chapter I mentioned that the Geometry
node Pointiness only worked with the Cycles Render Engine, so that
should already be set in the Render Properties tab. However one
consequence of using Cycles is that if you use the Volume Scatter

83
 shader in the World Properties, as you did in the last chapter, it is too
opaque, even on the lowest possible setting. The shader attenuates the
light coming from the environment texture into the scene rather than
the light being scattered from objects back to the camera, but the result
is the same; a black cat in a coal cellar!

A technique often used with Cycles is to add a mist pass in the

View Layer Properties tab. However that requires setting up the
compositor to do post-rendering image processing, which is a big topic
you will cover in Chapter 13.

An alternative though, is to add a sphere with a Volume Scatter

shader so it surrounds everything in the Camera view. This is the
technique you used to fill the Lower_domes with atmosphere in
Chapter 2 so check back and follow the instructions for that, but
obviously don’t add a wireframe modifier or any shader on the
material’s Surface. Figure 4-8 shows the position and size of the sphere
with the Material settings I used.

Figure 4-8 The mist sphere

Conclusion

In this chapter you delved into four quite different ways to

generate multiple clones from a prototype object. Each technique has
strengths and weaknesses and it’s hard to be prescriptive over which
one to use in a given set of circumstances. My suggestion would be to

84
experiment with all of them and gain a feeling for each one’s
idiosyncrasies.

Using the Particle System to generate instances of another object is

a way to scatter things over a surface. This is how the Quick Fur works,
which you applied to the velvet block in Chapter 1, and it would be
ideal for spreading rocks randomly over the Martian landscape you
created in Chapter 2. In the chapter introduction I mentioned a fifth
method of creating multiple instances of objects, geometry nodes,
which you will use in Chapter 14. The Blender Foundation has stated
that geometry nodes will become the standard method for processes
such as scattering grass on hills, as well as many other techniques, and
in my opinion they are significantly easier to use than particles.
However, as at version 3.0, functionality is being added and removed
with each release so it’s still very much worthwhile learning about the
established particle system.

In the next chapter you will learn to use brushes in Blender, for
sculpting 3D models but also for Texture Painting. This latter, along
with Vertex Painting, allows extra flexibility to be added to the
emission of objects by particle systems. You can paint areas to control
the Density and Size (as well as many other properties) of generated
particles. You could also have used the same method to paint the
corners and areas of the refrigerator where you wanted more wear and
rust to show through, which would have avoided some of the
restrictions caused by using the Geometry node Pointiness.

There is a lot to cover in the next chapter, but at the end I will
quickly revisit the Martian terrain as well as the rusty refrigerator
material to show you how they could be improved.

85
 5
S C U L P T I N G A N D PA I N T I N G

Using brushes to create 3D models

Sculpting in Blender opens up a whole world of

creativity. A quick search online will find a host of
fantastic work, each taking significant amounts of
imagination, hundreds of hours of work and non-trivial
computer power.
For this exercise we will not be so ambitious - taking one of the
simple manikins from the Martian scene of Chapter 2 you will develop
it slightly towards a detailed, pose-able model. Realism will not be an
objective but it should be fun! Links to the blend files and videos for

86
 the exercises in this chapter are at get-into-blender.com 5.SCULPTING
AND PAINTING.

Getting Assets From Other Blend Files

Reusing assets from earlier work makes lots of sense and it’s very
easy to do in Blender.

1. Start a new blend file and delete the start Cube

2. Select File4Append then navigate to the blend file you saved

back in Chapter 2 and under Object select the name Person01
(or whatever you named the manikin).
Append does seem a slightly obscure menu option for importing
assets from other blend files but, presumably, alternatives such as
“import” were already taken. Note that you can append other
components from previous projects such as materials and textures so
this is potentially a real time-saver.

Object Transformations and Vertex Positions

Once imported you need to do a little bit of tidying up to avoid

problems later.

3. To make the object stand at the origin, select the object (It will
probably be highlighted in red-orange showing that it is active,
rather than yellow-orange showing that it is selected) and, in the
Object Properties tab, change Location X and Y to 0.0 and
adjust Z so the bottom of the feet are on the ground.
In Chapter 2 I suggested using the X axis for the symmetry of your
figure, which is the default and makes life easier, but if the mirror
modifier on your figure uses the Y axis then you can fix that now by
rotating the object about the Z axis by -90°. If you did swap the axis of
symmetry from Y to X you will need to go into the Modifier Properties
tab and, under the mirror modifier change the Axis from Y to X.

87
Figure 5-1 Manikin Appended Prior to Applying Transforms

4. In the 3D Viewport menu select Object4Apply4All

Transforms which will make the vertex locations match the
positions as they appear on the screen.

NOTE The difference between object position, rotation and scale, and the
resulting vertex locations is often a source of problems for people
getting to grips with Blender. Part of the trouble might be that the
origin of each object is an inconspicuous orange dot, similar in
appearance to a selected vertex. Also, most of the time it isn’t
important: The figure would look to be standing on the ground if
its origin was in its head which was positioned a distance above
the ground or under its feet and positioned on the ground.
Sometimes, such as when using the array modifier in the last
chapter, the positions of vertices relative to the origin of the object
is critical. When things don’t work the way you expect this area is
always a good place to check, and applying all the transformations
to the object often fixes it.

If you had applied a scale factor to the object in the blend file from
Chapter 2 then the proportions might be wrong but even if that’s not the
case, now is a good time to adjust the mesh size.

5. Switch to Edit mode and with the relevant vertices selected run
Mesh4Transform4Skin Resize

Make a Basic Armature from the Skin Modifier

You will cover animation and rigging in Chapter 12 but there is a

convenient way to convert the stick-man mesh into a simple skeleton
(the technical name is Armature) using the skin modifier. Before doing
that you need to apply the mirror modifier.

88
 6. Move the mirror modifier to the top of the stack if it isn’t there
already and fix any gaps or overlaps in the skin. In my version I
could see that there were some overlapping skins caused by the
vertices on the spine and head not all having X value zero. You
can fix that by selecting the vertices and setting the X value to
0.0 or by increasing the tolerance of the Merge Distance in the
mirror modifier.

7. Once everything look OK apply the mirror modifier.

8. To create and select a new Armature object, in the skin modifier

click on the Create Armature button.
Unless you were lucky where you started extruding your skeleton
you will find an extra bone sticking out from the back of the neck. This
extra bone is the “root” and conventionally should be at the hips of a
humanoid Armature in order to make the posing process more intuitive.

9. Revert back to the skin modifier without an armature by

undoing the last action using CTRL-Z, as with normal
applications, then switch to Edit mode. You should be able to
see a dotted line around the vertex that had the root bone.

10. Select the hips vertex then click the Mark Root button in the
skin modifier.

11. Click Create Armature again and the extra bone will be
sticking out of the hips.

12. Switch to Edit mode (the Armature should already be the

selected object) and select the end of the root bone and move it
in using G then Y so it’s a more discrete size. When editing
Armatures you can select either the whole bone or the “ball” on
the end of the bone so you need to be careful.

13. Check that the Armature is working correctly by switching to

Pose mode, selecting a bone and rotating it. Experiment with
track-ball rotation as well by pressing R twice.

14. Return to the non-posed position by CTRL-Z or, if you’re not

sure how many tweaks you made to the pose, select all the
bones then from the menu Pose4Clear Transform4All. That
final All refers to transforms not bones, in case you don’t get the
result you expect because you didn’t first select all the bones.
As I mentioned above this Armature isn’t as sophisticated as the
one you will use to rig a model in Chapter 12 and it would be quite

89
 Sculpting the Main Features
It’s time to start creating a face. Begin by forming the eye line.

11. Mark a horizontal indent across the middle of the face using
CTRL Draw or Crease with a small brush size.

12. Enlarge two eye sockets just below the horizontal indent by
increasing the brush size and using CTRL Blob.

13. Shape the nose and brow using Clay Strips to build up or
scrape. Smooth from time to time as you go along with SHIFT
and whatever brush you are using at the time.

14. Rotate the view point as you work and use the Grab brush to
move the surface in and out where needed. Sometimes you can
use a large brush to move whole areas of the face such as the
chin but sometimes it is better to use a smaller brush size and
lower strength to move a series of points in small increments.
There are three different Grab brushes which behave in slightly
different ways, all are useful at times so try each of them.

15. Shape the outsides of the nostrils possibly using Draw Sharp or
Crease for where it meets the cheek.

16. Indent the nostril itself using CTRL Draw or Blob.

You may notice an idiosyncrasy that, even with dyntopo on, Draw
Sharp doesn’t create new faces although Crease and Draw do. Also, the
default behavior of this brush is negative or indentation though the little
picture shows a ridge (Crease also defaults to negative but the picture is
clearly a crease).

At this point it is a good idea to add some eye balls. Although you
might be tempted to try sculpting them, it’s better to just use two UV
Spheres. This will allow you to rotate the eyes and move the eyelids
independently.

17. Change back to Object mode, Add4Mesh4UV Sphere then

scale and rotate it 90° about the X axis.

18. Move the sphere so it’s in the correct position in the right
socket.

19. Add a mirror modifier to the sphere.

94
 The reason why you can’t now see two eyes is because the sphere
is being mirrored about its center point then the combined double
sphere is being moved to the right eye socket (re-read the NOTE earlier
in this chapter about object positions and vertex positions).

20. Using Object4Apply4All Transforms should fix this. To

finely adjust the position of the eyes you need to switch to Edit
mode and select all vertices so you move them all relative to
their origin.

21. When the eyes look to be in the correct position select the
Person01 object again.

22. Switch back to Sculpt mode and start to build up the eyelids
using Clay Strips.
As you try to create a square edge to the lid (where the lashes
would grow, see Figure 5-3) you will find that it is hard to make the
corner clean, even using Stroke4Stroke Method4Curve, as the
faces are arranged in a jumble rather than neatly along the contours of
the shape you are sculpting.

Figure 5-3 Jagged eyelid edge from using Dyntopo

This same problem will crop up with the lips and other areas and
the solution is the subject of Chapter 7, for the time being don’t worry
about these few rough edges.

Masking when Forming Larger Details

Ears have a sharp crease and the “clay” doubles back on itself
around the top and back. Using a mask will help here.

95
 23. In order to allow a smooth curve to form for the crease around
the ear, use a brush in smooth mode first to subdivide the large
faces.

24. Build up the area of the ear slightly using Clay Strips while
zoomed in to get a reasonable number of small faces.

25. To prevent the head around the ear being distorted, draw using
the Mask brush then, with the Grab brush, stretch the edge of
the ear out into, well, an ear shape, shown on the left of Figure
5-4.

Figure 5-4 Mask around the ears or hide vertices

The Mask can be fine tuned using options from the menu. As well
as clearing it when no longer needed you can invert, grow, shrink,
smooth, sharpen and alter contrast.

96
 MASKING

Masking is an important part of sculpting and painting on 3D objects, with

several methods available. One approach, also shown in Figure 5-4 is to switch to
Edit mode, select the vertices you want to work on, then Mesh4Show/Hide4Hide
Unselected. The hidden vertices will also be hidden when you switch back to Sculpt
mode.

An even better method is to use the Draw Face Sets brush to define areas you
want to keep separate. Each time you draw with the brush it produces a different
color and a new face set, if you want to spread an existing set press CTRL as you
start to draw. Once face sets are defined you can hide everything apart from the color
under the cursor by pressing H which allows very fast and efficient workflow.
However you cannot use face sets at the same time as dyntopo so, at th e stage of
the sculpting process where you need to subdivide faces, it is better to stick to Mask,
or hide vertices in Edit mode.

26. Sculpt the ear further with Clay Strips, adding and scraping
away and finally making an indentation for the ear hole with the
Draw or Blob brush. Have a look at a photograph of an actual
ear before you go too far with this.

27. To define the mouth, build up a flat dome using Clay Strips mid
way between the nose and chin, smooth it then add the upper
and lower lips, fattening then smoothing to achieve the correct
shape. For the sharp crease between the upper and lower lips
you will again improve your work flow using masking as you
did for the ears. Also don’t forget that switching to Edit mode
often allows you to make selections or move vertices in a more
precise way than in Sculpt mode.

28. Finally add the shape of the jaw, cheek and chin along with the
typical creases running from the edges of eyes, nose and mouth.

29. Make adjustments to the overall shape of the face using the
Grab brush with a large diameter.

Fixing the Bone Weights

When you’re happy with your sculpt save the blend file. As I
mentioned at the start, don’t expect to create a work of art, or even
anything very realistic. The sculpting process will have scrambled the
way the armature controls the mesh so this will need to corrected.

97
 1. To check the damage, switch to Object mode, select the
Armature then enter Pose mode.

2. Select the head bone and rotate it and you will see a horrible
mess. Most of the vertices don’t move, including the eyes,
which we know are still a completely different object. That will
be the first thing to fix by joining the two objects.

3. Select all the bones and clear all the transforms for the pose.

4. Back to Object mode select the eyes and apply the mirror
modifier. Then with the eyes still selected SHIFT-LMB on the
Person01 object so both are selected, then Object4Join.
The Sphere object should have disappeared from the Outliner
Editor and if you go to Edit mode you will see that you can select
vertices of the eyes and skin at the same time, it’s all one object now.

5. You can use Weight Paint mode as a visual check of bone

influence. With Person01 selected, choose Weight Paint from
the object interaction mode drop-down.

6. Open the Object Properties tab on the Properties pane (three

green dots connected in a triangle).

7. In the Vertex Groups list select a few bones and watch how the
coloring changes in the 3D Viewport.
Blue means no effect and red means full effect. You will see that
limb bones are affecting both sides of the manikin, and most of the face
remains blue when the head bone is selected. There are menu options
available in Weight Paint mode to help fix this but it is actually much
easier to just re-parent the armature and skin objects. For all practical
purposes armature objects should be the parent of the mesh objects they
deform, that way you can move the armature around the scene and the
model will follow it. The skin modifier doesn’t automatically set the
generated armature as the parent of the skin mesh, so this is an
operation that has to be done anyway.

8. Switch to Object mode and select the Person01 object, then

SHIFT select the armature then Object4Parent4With
Automatic Weights.
Your model may be fine after sculpting but it is possible to have
problems because of the complexity of the mesh, essentially vertices
that were at the same location as other vertices. This can frequently be

98
 cured by merging vertices by distance, however you need to make sure
you only select vertices in the head mesh excluding the eyes, so they
don’t get accidentally fused together.

9. In Edit mode select one vertex on the body then expanded the
selection to the whole mesh but excluding the eyes
Select4Select Linked4Linked

10. Use Mesh4Merge4By Distance and, in Adjust Last

Operation, increased the Merge Distance until the message at
the bottom reports some vertices being removed, but without
any dramatic simplification of the mesh.
Merging by distance will normally be sufficient but if the
automatic weight assigning still doesn’t work there are other options in
the mesh clean up menu, such as decimate, that should eventually fix it.

11. Check that the armature now poses the Person01 object as
expected. Clear all the transforms from the pose when you’ve
finished.
One thing you might find when you use automatic weight
assigning, especially for complicated or unconnected bits of mesh such
as the eyes, is that large movements of the bones produce unwanted
distortions: vertices get “left behind”. Sometimes you can fix this in
Weight Paint but often it is hard to find which bone is having an
unwanted influence. A useful way to solve this problem is to use Edit
mode as shown in Figure 5-5

Figure 5-5 Fixing bone influence in Edit Mode

99
 12. In Edit mode open the Sidebar on the right of the 3D Viewport
and select the Item tab

13. Select a vertex on the mesh that is showing a problem, for

instance the eye, then select a vertex that is behaving correctly,
the eyelid. Compare the Vertex Weights for each section in the
Sidebar
In Figure 5-5 you can see that the eyes were being influenced by
the neck bone Bone.01 with a very small weight factor.

14. Select all the problem vertices then, in the Object Data
Properties tab under Vertex Groups (on the right of Figure 5-5)
select each bone and either Remove it using the button below or
change the Weight value and Assign it so that the Vertex
Weights match the correct vertex.
As you can see, using an armature to pose a sculpted mesh has
many problems, and a much better approach is to create a new, simpler,
mesh over the top of the sculpted mesh and parent the armature to that.
You will use this technique in Chapter 7 after reworking this Person01
mesh and in Chapter 12 when you learn more sophisticated rigging
techniques.

Texture Painting

Texture Painting allows you to draw on an object using the mouse

or graphics tablet and to save the result as a texture for use in materials.
Painting is not just limited to solid colors, images can be cloned onto
surfaces, and textures can be used to generate patterns on the surface.
As well as the base color of the material, the other inputs to shaders
such as roughness, displacement and metallic can be painted.

There are two prerequisites for Texture Painting: The first job is to
generate the map that controls how the colors on the texture are
positioned on the 3D surface. This is referred to as a UV map and you
will look in more detail at this in Chapter 8. For now you will keep it
as simple as possible so these are the basic steps needed to set
everything up ready for painting..

1. To create the UV map select the Person01 object, switch to Edit

mode, select all the vertices then UV4Smart UV Project,
increase the Island Margin to 0.002 then click OK

100
 2. The second prerequisite for Texture Painting is to have an image
texture available for painting onto. Select the Material
Properties tab in the Properties Editor, create a new material and
rename it to “person_base”. Change the Timeline Editor panel
at the bottom to Shader Editor, make it larger, then add a texture
node to the material with Add4Texture4Image Texture and
connect the Color output to the Base Color input of the shader.

3. In the Image Texture node, create a new image by clicking

+NEW. For the moment leave the resolution at 1024x1024
(though you will see that produces a very pixilated material).
Give the image an appropriate name such as “person_base”, un-
tick the Alpha check-box, set the Color to an approximate skin
tone then click OK. Switch the Viewport Shading to Material
Preview to see the result.

4. Select the Texture Paint preset from the Active workspace

option in the menu at the top and the Active Tool and
Workspace Settings tab in the Properties Editor on the right. On
the left, in the Image Editor pane (the icon is a little paint can)
you will be able to see the 2D texture being filled in, as you
paint on the 3D object. You can also paint directly onto the
texture in this window which is occasionally useful for tidying
up areas.
The Image Editor pane also shows the UV unwrapping: notice that
the projection is not smart enough to know which areas needed to be
large because they might contain a lot of detail, and which areas could
be shrunk down relatively small.

5. Back in the 3D Viewport try painting on the figure, use F and

SHIFT-F to change the brush size and strength as you did when
sculpting.

6. Change the color of the paint either using the tool controls in
the header at the top of the editor, or in the Active Tool and
Workspace settings tab. Most of the options are duplicated in
the two places and many are the same as for sculpt brushes.
However, before Blender 4 it was possible to specify Material
or Single Image in the properties tab, but now this must be done
in the header, see the top of Figure 5-7.

7. Change the Falloff to the square edged profile and draw two
contrasting colors next to each other.

101
 Sometimes it is still hard to paint different colors next to each with
a sharp edge between, though a higher resolution image would
obviously help. In those cases it can be useful to mask the mesh as you
did in sculpting.

8. To mask an area you need to switch to Edit mode, set face

selection then select a group of faces. Switch back to Texture
Paint mode and toggle Paint Mask using the button just to the
right of the object interaction mode drop-down (the drop-down
you used to switch between Object, Edit and Texture Paint
mode). The Paint Mask icon is a solid square in front of an
outline square.

Vertex Groups
In order to swap from one set of faces to another set, then back
again, for instance to paint eyes, hair, skin, shirt, pants and shoes it
would good to be able to save the selections for re-use. This is exactly
what vertex groups do, although their use extends far beyond; almost
all modifiers can use vertex groups and some, for instance the armature
modifier, depend on vertex groups entirely.

Just as when assigning vertices to materials, vertex groups used as

masks really define the faces between the vertices, so the vertices at the
edge of the hair group, say, should also be in the group defining the
adjacent skin.

9. To create a vertex group for masking as you paint a shirt, in Edit

mode use face select mode and select the the faces to define a
shirt for the manikin. Using X-Ray and orthographic view will
help this.

10. To add a new group, in the Object Data Properties tab, under
Vertex Groups click the + button on the right. Double click to
rename it from Group to shirt. With the shirt group selected,
Assign the selected vertices using the button below.

11. Repeat the above process to create groups for hair, eyes, pants,
shoes. Because the eyes are unconnected meshes you can select
a single face on each eye then click Select4Select
Linked4Linked

12. Finally use the Select button beneath the list of vertex groups to
cumulatively select the faces for all the groups, Select4Invert

102
 then create a group for the skin. This is why you needed to
change to face select mode, if you had defined the vertex groups
in vertex select mode there would be gap when you inverted the
selection.

13. Switch between Edit and Texture Paint mode using the different
groups as masks, painting with a suitable color.
The result is probably rather uninspiring - everything looks like it’s
made from plastic! Adding a Texture to the brush might improve that:

14. In Texture Paint mode open the Active Tool and Workspace
setting tab and scroll down to Texture and add New. The texture
will be called “Texture” and will show as solid black which will
prevent the brush from adding any color.

15. A brush texture is subtly different, and actually much more

powerful, than a simple texture such as you have used
previously as part of materials. It requires a second stage to
specify how the texture is to be generated; select the Texture
Properties tab (pink and black checkerboard right at the bottom)
and in Type select Voronoi and change the Coloring to
Position.
When you draw with this brush it will combine the pattern and
colors of the texture with the color selected for the brush. The size of
the pattern depends on the size of the brush so as you zoom in and out
the pattern size will change relative to the object you are painting
unless you change the size of the brush at the same time.

16. Paint the shirt, pants and shoes using different textures and
colors.

17. Change the Texture to Type Wood with Pattern set to Band
Noise and paint the hair.
One thing you might find, when over-painting with a different
texture brush, is that the black color “blots out” the design that you
have already done, even if you turn the brush strength down with
SHIFT-F. Often you want the black part of the texture to be transparent,
in which case you need to:

18. In the Texture Properties tab, under Color, tick the check-box
against Color Ramp. The positions of the sliders can also be
used to make the pattern sharper or softer, as well as the color
and alpha value at each stop.

103
Displacement texture

Once you have some interesting base color textures drawn onto
your model it would be fun to add a displacement texture as well.

19. Switch the Active workspace preset back to Layout using the
menu at the top. In the Shader Editor at the bottom add a second
Image Texture node and create a new texture renaming it to
“person_displacement” with a black default color (back is 0.0
displacement, white is 1.0 displacement and, if the scale of the
object is small, even low values can look extreme). Connect the
Color output to the Displacement input of Material Output
node. You might get a better result, with finer control, if you
add a math multiplication node between the displacement
texture and the material output.

20. Change back to the Texture Paint preset using the menu at the
top then click on the texture selector in the header, visible at the
top of Figure 5-7. This should display the name of the new
image texture you just created but sometimes Blender doesn’t
refresh this straight away. Change the Mode from Material to
Single Image and make sure person_displacement is selected.
Draw with a fairly dark gray in the hair area.

21. Draw displacement onto the jacket area as well using the same
wood texture but with reduced turbulence. In Active Tool,
Texture settings you can adjust the angle and scale of the pattern
to create the effect of horizontal bands on a quilt.
If there are multiple textures involved with a .blend file, then you
need to tell Blender which one to paint over by either selecting it in the
Image Editor drop-down, clicking on the node in the material Shader
Editor or, much more reliably, specifying it in the Texture Paint header

When you draw the iris and pupil on the eyes you will find that
there is a shortage of pixels to give anything like a sharp edge, even if
you draw directly onto the 2D image texture in the area on the left. In
Chapter 8 you will see how to solve this problem by using different UV
unwrapping and a different material for highly detailed areas such as
eyes.

104
 Pack your textures or save them

After you have finished creating the base color and displacement
textures for your character you would be forgiven for assuming that
they would be included in the blend file when you save it. However,
this is not the default behavior of Blender. You must either save
textures as image files or pack them into the blend, though, if you try to
exit the blend file, you will be warned and given an opportunity to save
any unsaved images.

22. The Save All Images button is now hidden inside the material
or texture selector in the header, visible at the top of Figure 5-7.
Click on that button now.
This is the most straightforward way to avoid losing painted
textures and the data is packed into the blend file. Frequently you need
to save a texture as a separate file, for instance if you want to use it in
different blend files later or edit it with an external application. To save
an image as a separate file:

23. In the Image Editor area select each image with the drop-down
then from the menu Image4Save
To pack textures, that were previously saved separately, into a
blend file:

24. In the main Blender menu File4External Data4Pack All

into .blend or better, set this behavior as default so you don’t
forget later to tick the check-box File4External
Data4Automatically Pack into .blend

NOTE At some stage you will find that an object in one of your scenes has
turned to a strange pink or purple color. This is a sign that there is
a missing image file referenced from a material in the blend file.

There are options under File4External Data to help find and link
missing image files, as well as options for packing and unpacking
images in different ways. In general, problems with external images
only become apparent with larger projects, blends imported from other
application and online resources or where images need to be processed
externally.

105
 Improving the end result

There are many deficiencies with attempting to paint skin this way,
not least is the fact that photographic images are readily available and
can be incorporated into skin materials quite easily. However there are
two things you can do to improve things a little. Firstly, subsurface
scattering better represents the fact that skin is actually a rather
yellowish gray color but the blood underneath makes the overall effect
pink. However the default BSDF shader produces a rather “plastic”
look that needs to be toned down in many areas. Secondly, eyes are
highly reflective and need to have a much lower roughness than the rest
of the skin:

1. Change back to the Layout workspace using the tab at the top of
the screen. Set the Viewport Shading to Material Preview.

2. Select the Person01 object and open the Material Properties tab.

3. You can copy and paste one material to another using the tiny
drop-down menu button to the right of the material slots list,
just visible in Figure 5-6. However there is a quicker way, using
the button, just to the right of the current material name, with an
icon of two overlapping squares. First rename the material from
Material.001 (or whatever default name was created) to skin.

Figure 5-6 Copy the skin material settings to the eyes material

4. Add a new material slot using the + button to the right then
select the skin material from the drop-down below.

106
 5. Click on the the little New Material button to the right of the
drop-down then rename the new material from skin.001 to eyes.
Both the skin and the eyes material now use the same texture
file you just painted, so you can add a subsurface color to the
skin one and increase the reflection for the eyes.

6. First you need to assign the vertices appropriately. Switch to

Edit mode, de-select all vertices, select a vertex in each eye then
Select4Select Linked4Linked. In the Material Properties tab
Assign the selected vertices to the eyes material.

7. With the eyes material still selected decrease the roughness to

zero and increase the Specular value to above 0.75

8. All the vertices not in the eyes already use the skin material so
you don’t need to assign those.
In reality the amount of subsurface scattering varies from lips to
cheeks to chin so this would be a good instance where it would be
useful to be able to paint a texture which can then used in a more
general way. This is the technique I referred to in the previous chapter
for controlling the worn areas of the refrigerator rather than relying on
the Geometry node pointiness. You can see the screen layout for the
following instructions in Figure 5-7.

9. Change back to the Texture Paint workspace using the tab at the
top of the window then swap the panel on the left from Image
Editor to Shader Editor.

10. Create a new image texture node and rename it to

“person_subsurface” with initial color white. The default
method of subsurface scattering, Christiensen-Burley, gives a
good result for skin, and you will mainly need to tone it down
by texture paining darker colors.

11. Connect the Color output to the Subsurface Scale input of the
shader as shown in Figure 5-7.

107
Figure 5-7 Painting a subsurface map texture

12. With Viewport Shading set to Material or Rendered, draw areas

of decreased subsurface scattering onto the Person01 object.
The whiter the paint color the stronger the effect.

Using an Image Texture as Blend Factor

Try using this technique on the rusty refrigerator material from the
last chapter. You can replace the Geometry and first ColorRamp nodes
with a hand painted texture representing areas of wear as in Figure 5-8.
As this method doesn’t rely on the cycles render engine the texture
painting can be done with Viewport Shading set to Material Preview
which is normally much more responsive.

108
Figure 5-8 Painting areas of wear

Although texture painting will probably work after a fashion with

the default UV map from the original cube, it will be much better if you
generate a new one following the first step under the Texture Painting
heading in this chapter..

Painting Particle Density

The other promise I made earlier was in relation to the scattering
of rocks on the Martian landscape from Chapter 2. To do this you need
to follow the steps you used in Chapter 4 to create grass on the hill but,
obviously, using simple rock shapes. In Figure 5-9 you can see that I
have used two different sizes for rocks (they are the same shape,
basically a cube subdivided once then slightly randomized with
Mesh4Transform). I have also appended the grass tuft object from
Chapter 4 and scattered that as a third particle system inside the dome.

109
Figure 5-9 Improved Martial landscape

Painting particle density requires a slightly non-obvious step

involving a type of texture that is different again from images, brushes
or procedural textures. It also involves quite a few steps that need to be
set up in different places, so follow the instructions carefully. Links to
the blend files and videos for this modified landscape are at get-into-
blender.com. 5. Sculpting and Painting

1. Open the blend file from Chapter 2. Create and rename the
boulder object then duplicate and scale it down as a pebble
object. Append the tuft object from the blend file for Chapter 4.

2. With the Terrain object selected switch to Edit mode, select all
the vertices then UV4Smart UV Project then switch back to
Object mode

3. Set up the three particle systems following the instructions from

the last chapter. Name them “pebbles”, “boulders” and “grass”.
For the moment leave them at the default number of particles
which will be spread over the whole area.

4. In the Shader Editor panel create three solid black images.

Name the images “pebbles_texture”, “boulders_texture” and
“grass_texture”. You don’t need to connect these textures to
anything in a material but you need to associate them with the
terrain object and select each one here to paint onto it.

110
 5. Back in the Particle Properties tab, for each of the three system,
scroll down to the Texture section and create a texture for each
renaming them “pebbles_tex”, “boulders_tex” and “grass_tex”.
This is the non-obvious step as these textures are rather
different from textures you have encountered so far.

6. With each particle system selected in turn click on the Texture

Properties tab (pink and black checkered icon at the bottom).
This should have defaulted to ParticleSystem with the correct
particle texture selected. Ensure that the texture Type is Image
or Movie.

7. Under the Influence section un-tick General Time and tick

Density and also Size.

8. Under the Mapping section select Coordinates UV and Map

UVMap

9. Under Image, Settings select the relevant image you created in

the Texture Paint workspace. The Alpha check-box should
match the alpha setting you used to create the textures for
painting onto.

10. Draw onto the terrain where you want each of you particle
systems to appear. In the Texture Paint header you need to
change which texture is selected to ensure that you are painting;
pebbles, boulders and grass. Occasionally I find that I have to
switch to and from Object mode to refresh the relevant particle
system.

11. In the particle settings adjust the number of particles emitted,

their rotation, randomness and children to get the desired effect.

Conclusion

You have covered a lot of ground in this chapter. Using brushes on

3D models can go far beyond sculpting and creating surface textures.
However you also saw that there were potential deficiencies with these
techniques.

There is a large element of skill required for sculpting and the

more you practice the better you will get, however, large, detailed
objects will inevitably need a powerful computer and benefit from the
use of graphics tablets. You also found that the jumble of faces
produced by sculpting makes it difficult to get clean lines for both the

111
edges of different materials and in order to allow the mesh to be
deformed when it is rigged.

When texture painting, the default UV unwrapping is often

inadequate for the detail required in areas such as the face. Also the
layout of the different skin components has discontinuities in visible
locations. As with plastic surgery, the seams need to be hidden at the
back of the head or above the hair line, certainly not in the middle of
the face.

In Chapter 7 and Chapter 8 you will learn how to overcome these

two problems of topology but the next chapter will concentrate on
some of the methods in Blender for importing and exporting models.

112
 6
I M P O R T, E X P O R T A N D A D D - O N S

Increase productivity by borrowing

and sharing work

One of the appeals of Blender is the massive community

of enthusiasts as well as professional artists. This means
that any problem you run into is likely to have cropped
up before and you will either find an answer on a forum
or wiki, or somebody will have made an add-on to get
round the issue: There are thousands of add-ons and the
majority are free.

113
 Blender is used by many game creators to make 3D assets as well
as by artists as a render engine for models made with other 3D
modeling software. As a result there are many built-in import and
export functions, and when they're missing there's a good chance
someone has written an add-on that can help.

In this chapter you will enable some add-ons that are already built
into Blender waiting to be used. In addition you will download and
install some add-ons as well as some 3D models in various formats.

Lots of Different File Standards

A quick search online will reveal that there is a bewildering

number of 3D model file specifications. This situation is related to the
evolution of computer hardware and software allowing more
complicated information to be stored, but it has been compounded by
companies making propitiatory specifications for commercial reasons.
You will most frequently encounter Wavefront .obj, Autodesk .fbx and
Collada .dae but there are increasingly lots of .blend files available.

NOTE Blender allows scripts to be included inside files and executed

on load. Although automatic execution is disabled by default,
there is a real security risk if you do allow them to run. For this
reason you must ensure that any file you download comes
from a reputable, trusted source. If you have the slightest
misgivings about this then certainly do not allow a script to
run.

Download and Import an Asset

OBJ Files
The first format to try is Wavefront .obj which is an older
specification and holds only information about the mesh and associated
materials, so no armatures for rigging characters and no animations.
For this exercise you can search for “download free 3d obj angel” and
you should get many results. I will use the model in Chapter 12 on
rigging and animation so it would be ideal if you download the same
one as shown in Figure 6-2. It seems to be available on free3d.com and

114
 open3dmodel.com, probably more, but any character model with arms
and fingers, described as low or medium poly and in .obj format would
do. The majority of websites offering free downloads require you to
register but many do not. Links to the files and videos for the exercises
in this chapter are at get-into-blender.com 6.IMPORT, EXPORT AND
ADD-ONS.

1. Download the .ZIP file from wherever you find it and unzip into
the folder you created for the exercises of Chapter 6.
You will see that it contains an .obj file, a .mtl and several .jpg
files. The .obj and .mtl files are plain text and very simply formatted so
you can open them with a text editor and see the values of vertices,
normals and texture coordinates.

2. Open Blender and delete the default cube then select

File4Import4Wavefront (.obj) and navigate to the location
of the .obj file and select it.
The model that appears will display some of the issues of
exporting from one application and importing into another. Not all
software uses Z for the vertical axis, sometimes Y is up and Z is
towards the camera. Also the scale will often vary from one unit equals
one meter used by Blender so the result you see will be a giant lying
face down as in Figure 6-1.

Figure 6-1 Imported .obj objects can be too large to see the distant parts

3. Rotate it -90 degrees about the X axis and scale it down so it is

just under 2m tall. Use the orthographic view to see the grid

115
 lines, if you’re not sure if the lines are for 1, 10 or 100m you
can see the overall dimension in the Sidebar.

4. As should now be second nature: Object4Apply4All

Transforms
As everything in the scene is relative to other objects it might seem
unnecessary to scale it down, however, viewing the object with the
camera and illuminating it with lights will be much simpler if you scale
it down. The default camera and point light are five or six meters above
the origin and can be used as reference when scaling by eye.

The other problem you will frequently encounter when importing

files created on different applications is that the materials will not be
converted correctly on import. Blender will try to allocate the
information from the original materials to the inputs of a Principled
BSDF Shader but sometimes the process is too difficult to do
automatically. If the surface has patches of partial transparency, as
visible on Figure 6-2, in the Material Properties tab under Settings, try
switching the BlendMode from Alpha Blend to Opaque or Alpha Clip.

Figure 6-2 Partial transparency due to Alpha Blend

.fbx is another ubiquitous format. It has the advantage that

armature and animation information can be incorporated in the file.

FBX Files
To follow the steps in this chapter and Chapter 12 you will need to
sign up with mixamo.com - it’s owned by Adobe so it’s reasonably
reputable and I certainly haven’t been bombarded with spam since

116
 joining. Mixamo has a selection of different characters and animations
that you can select in combination. As an example of a low resolution
mesh character select the Peasant Man and Salsa Dancing as the
animation, download it with the default settings (Binary FBX, with
skin). Then, as an example of a slightly higher polygon count mesh
download character Lola with animation Climbing Up Wall.

1. In Blender start a new scene with File4New4General, delete

the cube then use File4Import4FBX and select the salsa
dancing peasant file from where you saved it

NOTE If you use an .fbx file from Mixamo downloaded prior to 2022 you
need to do a couple of steps before you click on the Import FBX
button. In the import options on the right, under Transform un-tick
the Use Pre/Post Rotation check-box. Then, Under Armature tick
the Automatic Bone Orientation check-box. Without these settings
the bones would all be at ninety degrees and, although the
animation would work, it would be hard to pose the armature later

The character should be about the right size and the right way up!
The armature is obviously present as the bones protrude in various
places..

2. Switch Viewport Shading to Rendered and zoom in to look at

the detail of the face.

3. Open the Overlays drop-down at the top of the 3D Viewport and

under Geometry switch on wireframe as shown in Figure 6-3.
(The icon is a solid circle overlapping an outline circle.) This
draws wireframes over all objects in the scene, to apply it only
to a specific object use the option in the Object Properties tab
then the Viewport Display section.
Notice how few polygons there are (only 2,400 for the whole
model) but there appears to be quite a lot of detail in the rendered
material, such as the garments and beard. This has been accomplished
by storing surface detail on a normal map image. If you look at the
Material in a Shader Editor area you will see there is a specular texture
to control shine as well a normal map.

117
Figure 6-3 Surface detail using a normal map

In Chapter 9 you will learn different ways to “bake” detail from a

high resolution mesh such as one created by detailed sculpting onto a
normal map texture. For the time being try running the animation. If
you changed the Timeline Editor to a Shader Editor in order to see the
details of the Material, change it back and reduce the End frame from
the default 250 so it matches the last frame of the dance and it loops
smoothly.

4. In order to see the polygon count, open the Overlays again tick
the Statistics check-box.

5. Add a Subdivision modifier and increase the levels in the

viewport to 1 then 2 then 3, which will generate a mesh of
9,000, 38,000 or 150,000 faces.

6. Run the animation again for each level of subdivision and you
should see a noticeable slow-down and, hopefully, appreciate
the need for lower resolution meshes for animations, especially
in games.

7. Start a new scene and repeat the process for Lola Climbing Up
Wall.
This model has more than twice as many polygons as the Peasant
Man and if you add a Subdivision surface you will slow the animation
even more dramatically. Although the subdivision makes little

118
 improvement from a distance, when viewed close up it is quite
noticeable, especially where edges are viewed in profile. This is the
reason there are two settings for the level of subdivision: rendering a
scene can take quite a while so the marginal cost of calculating two
levels of subdivision is negligible, but when modeling in the viewport
it can be very frustrating.

Compare the arrangement of the mesh on Lola’s face with the

mesh on the Peasant Man. As well as just dividing the surface into
more polygons, they have been arranged to follow contours forming
concentric patterns around the lips, eyes, nostrils and ears.

Figure 6-4 Relatively clean topology for Lola’s face

By doing this, and by keeping all the polygons as quads, the mesh
can be subdivided predictably and can be deformed without causing
glitches. In the next chapter you will learn methods for creating clean
topology like this.

Export an Asset

Being able to create assets with Blender, then use them in other
applications is very useful, but there are obviously limits to the
functionality that can be directly transferred. Much of the mesh,
armature and animation information is common to many different
systems but physics, particles and much of the material nodes is not. A
good common denominator format is the .fbx file so try exporting the

119
 sculpted manikin you made in the last chapter, then re-import it to see
what fell through the cracks.

8. Start a new scene and delete the default cube then append the
Person01 object from the blend file you saved at the end of the
sculpting and texture painting exercise in Chapter 5.

9. Select both the mesh and the armature then click

File4Export4FBX. Browse to a location to save the file but,
as with importing, do not yet press ENTER or click the Export
FBX button yet. There are two non-default options that need to
be selected first.

10. In order for the textures that are packed into the blend file to be
embedded in the .fbx file you need to set Path Mode to Copy
and click the inconspicuous button beside it Embed Textures

11. Under the Include section, Limit to, tick the Selected Objects
check-box.

12. Now verify the information preserved in the export by re-

importing it. Click Export FBX then start a new scene, delete
the default cube and import the .fbx file you just saved. You
don’t need to make the changes needed to sort out the bone
orientations that you had to do for the Mixamo imports.

13. Check over the imported object. You should find that the
armature, mesh, material base color, roughness, specular and so
on are OK but the displacement and specular textures have not
survived the transfer.
Often the information in the .fbx file is sufficient for your
requirements, where you are constructing 3D assets for a video game,
for instance. If you need to have more than this, you will have to export
all the texture files manually by unpacking them, then reconstruct the
material or physics in the destination application by hand.

Add-ons

It is probably fair to say that there is an add-on for almost anything

you could possible need to do in Blender. Many add-ons are already
included and simply need to be enabled, other add-ons are developed
by enthusiasts and are available to download for free, and yet more
add-ons are made by professionals and distributed commercially.

120
 The choice of which add-ons you should use depends entirely on
the problems you encounter so I will try to avoid coloring your
judgment by giving you advice in this book. My approach would
always be to get as far as I could without resorting to an add-on in
order to understand the issues as thoroughly as possible. Then I would
research forums and videos online and use that information to make my
choice.

The add-ons you will enable or install as part of this exercise are
intended to be “a bit of fun” rather than useful but they should give you
a good idea of the general process.

Enable a Pre-installed Add-on

A key part of Blender’s design philosophy has been to keep the
core application as general purpose as possible but allow extra
functionality to be supplied relatively easily using add-ons. The import
and export menu options you used earlier in this chapter are add-ons
that are enabled by default and, rather surprisingly, even the cycles
render engine is an add-on.

However this flexibility means that add-ons are allowed to

introduce menu options, control panels and buttons, almost anywhere
in the Blender UI without restriction. The majority of community
contributed add-ons, once installed and enabled, can be found in the
Sidebar, but not all of them. This means that, in some cases, the only
way to discover how to use the add-on is by carefully reading the
manual and watching a few videos online. Fortunately, the process of
enabling a pre-installed add-on is very straightforward and happens all
in one place.

1. Using the menu at the top of the window hit Edit4Preferences

then the Add-ons button on the left.

2. You can use search to find add-on. Leave the default filter using
the Official and Community options at the top then in the search
box type “extra” which should find two add-ons, Add Curve:
Extra Objects and Add Mesh: Extra Objects. Tick the check-box
on both in order to enable them, then close the Preferences
window.

3. In the 3D Viewport if you select Add4Mesh or Add4Curve

you should now see a host of new options that weren’t there
before.

121
 4. As an example try using the Celtic Links curve. Start a new
scene, delete the default cube and add an icosphere, switch to
Edit mode then squash it to about 20% of its original height.

5. Switch back to Object mode then

Add4Curve4Knots4Celtic Links, if the last menu option is
grayed out it is probably because you didn’t have an object
selected beforehand.

6. In the Adjust Last Operation panel change Weave Up to 0.05,

Weave Down to -0.05 and Bevel Depth to 0.07. The result is
shown in Figure 6-5. Select the icosphere and delete it.

Figure 6-5 Celtic Knot wrapped around an icosphere

As you can see this would have been a very efficient way to
generate the gold wire mask you constructed in Chapter 3. I am going
to make a quick detour here to show you a slick way to add some dirt
in all those cracks formed where the wires cross each other. You will
use an input node to the material called Ambient Occlusion which is a
way to add realism to computer generated scenes without having to
crunch through the the full ray tracing work that is undertaken by the
cycles render engine. Ambient Occlusion basically finds, from the
perspective of a given point on a surface, how much of the sky is being
obscured.

7. To give the knot some context add a plane, scale it up, move it
just below the Celtic knot and create a shiny dark gray material
for it.

8. There’s no point re-inventing the wheel so append the gold

material you created for the first exercise of Chapter 3 and
select this material for the Celtic knot object.

122
 9. As you did in Chapter 3, change the background lighting to use
an environment texture and check what this looks like by setting
Viewport Shading to Rendered.

10. In the Render Properties tab on the right leave the render engine
set to Eevee but tick the check-box for Ambient Occlusion and
Screen Space Reflections.

11. Now you can start to modify the material to use ambient
occlusion. Switch the Active workspace using the Shading tab
at the top of the window, then add an Ambient Occlusion node
using Add4Input 4Ambient Occlusion.
Now see if you can figure out a way to use the AO output of this
node to change the color and roughness of the wires where they cross
over each other. Try to think through the logic before looking at Figure
6-6. As a hint the AO output will go from 0 where the point is at the
bottom of a deep crack to 1 when there is nothing around the surface,
you can use two color ramp nodes to map this value to a range for the
Base Color from dirty to clean gold and for Roughness from 1 to 0.
Remember white is 1 and black is 0.

Figure 6-6 Using Ambient Occlusion for dirty corners

By this stage I’m sure you appreciate that you could have used the
AO output to combine two different shaders just like you did for the
rusty fridge material to get a much better grungy look in the nooks and
corners. Also, see what it looks like using the cycles render engine, it is
a little bit more realistic but eevee is pretty good. In Figure 6-7 the left

123
 half was rendered using cycles, taking five minutes, and the right half
was rendered using eevee in five seconds. It’s quite hard to see the
difference.

Figure 6-7 Cycles on the left, Eevee on the right

Spend a little time experimenting with all the extra meshes and
curves. There is so much and I couldn’t hope to do the add-ons justice
here, however you will use the wall factory in Chapter 12 to provide
Lola with something to climb up.

Download and Install an Add-on

Downloading and installing add-ons carries the same warnings as
when downloading a .blend file but even more so, in order to use an
add-on you have to execute someone else’s code. My advice would be
to stick to add-ons that are popular in terms of reviews on websites and
discussion in Blender forums.

Blender add-ons are written using the python scripting language

and, although some simple add-ons might be contained in a single file
with a .py extension, they are usually distributed as a .zip file. Blender
can install add-ons in either form but the normal route uses the zip file
without expanding it. The example I suggest you install is Oliver
Weissbarth’s book generator from
https://oweissbarth.de/software/book-gen-blender-addon/.

124
 1. Download from the link and save the .zip file in your
downloads directory without unpacking it.

2. In Blender select the top menu Edit4Preferences then Add-

ons as before.

3. Click Install at the top and navigate to the .zip file you just
saved then click Install Add-on

4. Tick the check-box next to the add-on Add Mesh: BookGen in

order to enable it.

5. You might be forgiven for expecting to find all the controls for
BookGen under the menu option to add other mesh objects,
after all that’s where the mesh extras appeared. However this
add-on follows the normal convention of putting all its controls
in a tab in the Sidebar. The functionality is fairly simple so
using it should be intuitive but there is a good video on Oliver’s
website if you get stuck. Figure 6-8 was very quick to set up
using the wall factory and this book generator. Each book is a
separate object so it is quite simple to create random colors for
each cover, using the object info node random output and a
color ramp, as you can see in the material nodes at the bottom.

Figure 6-8 BookGen with random cover colors

125
Conclusion

Although there is enormous scope and variability of resources, the

process of importing and exporting to and from Blender is quite
straightforward. Using other peoples’ models, components and add-ons
can, potentially, transform your productivity and push your creativity to
a whole new level.

Seeing how other artists have gone about making things is also a
good way to learn. You saw the difference between the low poly model
Peasant Man and the medium poly model Lola and how they both
added surface detail using a normal map. In Chapter 9 you will learn
how to transfer the detail from a jumbled, high poly mesh, such as the
one you created by sculpting a face in the last chapter, onto a normal
map matching a clean, low poly mesh. However, before that, you will
need to make that low poly mesh, and that is the subject of the next
chapter.

126
 7
M O R E T H A N O N E WAY T O S K I N A
C AT

Building one mesh over the top of

anther one

Making a mesh conform to the shape of another one is a

surprisingly common job in certain areas of 3D
modeling. The process of creating a regular low poly
mesh after sculpting is called re-topology but the mesh
editing techniques used for it are important for all areas
of 3D modeling in Blender.

127
 Over the course of this chapter I will demonstrate the value of
making objects comprised from quadrilateral faces as opposed to
triangular or higher sided polygons. I will then show different methods
for making and fitting clothes to your manikin from Chapter 2 before
starting to apply the re-topology process to the sculpted face. As I
mentioned when you did the sculpting, there are many excellent, ready-
made face meshes available free on the internet, so your objective here
is not to reproduce the work of expert modelers and anatomists. Doing
re-topology on a complex mesh can be rather slow, and there are
several add-ons that could make your life much easier, so you will
simply be getting an appreciation of what’s involved and learning skills
you can transfer to other modeling you do in the future.

After the re-topology exercise you will learn two useful features of
Blender that allow you to change the shape of an object by moving a
slider, Shape Keys and Drivers. In some ways both of these
complement the use of armatures to deform meshes and they would
have fitted logically with the work on animation in Chapter 12. Links
to files and videos used in this chapter are at get-into-blender.com
7.MORE THAN ONE WAY TO SKIN A CAT

The Value of Quads

After a while watching Blender videos online, or reading helpful

posts in forums, you might become aware of a general sentiment that
“clean” topology is good, and clean topology always consists entirely
of quads. This isn’t absolutely true, like all good rules there are
exceptions, but it is a good rule!

As a quick exercise make an object from quads then see what it

looks like when you convert it to triangles.

1. Start a new scene and, with the default Cube, switch to Edit
mode and extrude one of the faces five or six times so each
extrusion is about the same size as the original Cube.

2. With the extruded face still selected turn Proportional editing on

with a radius of influence a little less than the length of the
object. Rotate the end face and move it so the line of cubes
forms a curve.

3. Switch back to Object mode and add a subdivision modifier.

The result should look like the “banana” in Figure 7-1

128
Figure 7-1 Subdivision of quads

4. In Edit mode select all the faces then Face4Triangulate

Faces.
Notice that the whole banana surface has become subtly eccentric
and distorted as shown in Figure 7-2

Figure 7-2 Subdivision of triangles

There are further complications associated with meshes that are

not constructed from quads and, as you complete the exercises in this
chapter, I will point them out.

Create a New Object using Part of Another Mesh

For the first approach to making a mesh following the shape of

another mesh, you are going to make a shirt for the Person01object
from Chapter 5. You will copy a part of the body mesh, modify it as
required, and then shrinkwrap it back onto the original mesh.

Converting Triangles to Quads

Before making a copy of part of the mesh you can improve it by
running the reverse process of the triangulate faces function you used
above.

129
 1. Start a new scene and delete the default cube, you don’t need to
save the banana! Append the Person01 object from the blend
file you saved at the end of Chapter 5.

2. To ensure that the armature is unposed select the armature

object, switch to Pose mode, select all the bones and hit
Pose4Clear Transform4All.

3. Switch back to Object mode, select the Person01 object then

change to Edit mode. You will see that the body and arms are
built from triangles rather than quads. This is a by-product of
using the dyntopo option while sculpting. Deselect all the
vertices with ALT-A then try to select an edge loop running
round an arm with SHIFT-ALT-LMB on a circumferential edge.
See Figure 7-3 You will find that you only select a single edge.

4. Blender can try to convert triangles to quads for you. Select all
the mesh with A then click Face4Tris to Quads, deselect all
the vertices then retry selecting an edge loop round an arm. This
time you will find it works, but note that tris to quads didn’t
convert all of the quads as you can see on the right of Figure 7-
3.

Figure 7-3 Loop Select and Loop Cut only work with quads

The reason that the loop selection didn’t work with triangles is
because the algorithm works from vertex to vertex by finding the
opposite edge from the one joining the previous vertex. This only
works reliably for vertices with four, or at the very least, an even
number of edges.

Possibly more significant than loop selection is the ability to make

loop cuts, a technique often used when building meshes “by hand”,
both for re-topology and for hard surfaces.

130
 Separating a Selection
You will now duplicate part of the body mesh and use it to create a
new shirt object that you can further refine.

5. Switch to orthographic view in the Y direction, toggle X-Ray on

then using circle selection select the vertices for a short sleeved
shirt. You will add a mirror modifier so you only need half a
shirt. Your selection should include vertices on the center line
but none from the opposite half. Figure 7-4 shows where I cut
off the sleeves and neck, but use your own sartorial judgment.

6. Duplicate the selected mesh using SHIFT-D then LMB without

moving the mouse at all. If you move the mouse before
clicking, it will move the duplicated mesh, which you don’t
really want.

7. You can create a new object from the mesh. The duplicated
mesh is still part of the Person01 object but the vertices will still
be selected, so use Mesh4Separate4Selection

8. Switch to Object mode, select the new object and rename it

Shirt. You will find that the object is a child of the Armature
and, if you check the Modifier Properties tab, it still has an
armature modifier though if you pose the character will only
pose half the shirt.

9. Add a mirror modifier to create a full shirt. If you try posing the
armature now you will find that the mirrored vertices move too,
which is not what’s required. You will need to regenerate the
armature modifier later so, in the Modifier Properties tab, delete
the existing one. At this point your shirt should just have a
mirror modifier, if the object has inherited any others you added
to Person01 experimentally then delete those as well.

10. In the Material Properties tab delete the two existing materials
skin and eyes then create a new material, rename it shirt and set
the Base Color to a strong color different from the existing skin
material. Change the Viewport Shading to Material Preview
and, hopefully, you won’t see your new shirt in all its glory,
though you might see some of it with strange banding. The
faces of both objects are in the same location so you need to
make sure that the shirt is outside the body.

131
 The Shrinkwrap Modifier

11. To ensure the shirt is visible above the surface of the Person01
mesh as in Figure 7-4, add a shrinkwrap modifier to the Shirt
object. Set the Snap Mode to Outside, the Target to the
Person01 object and the Offset to 0.001m.
The shrinkwrap modifier can be a very useful aid when making
clothes like this, but it can also be used in many other situations where
you need one mesh to line up with the surface of another one.

The two settings, Wrap Method and Snap Mode are a little hard to
understand at first.

Wrap Method controls the way that wrap locations on the

underlying mesh are found. For simple meshes like this, the default,
Nearest Surface Point is mainly fine, and it’s quick to calculate so the
modifier can sometimes be left in place on a garment that is being
animated and subdivided, to prevent skin poking through. In some
locations the calculation may not work perfectly, for instance under the
arms of my manikin where the surface folds back on itself. In these
cases the Target Normal Project will often give an improved result and
for even more difficult situations the Project option can sometimes be
needed as it has more fine-tuning available.

Snap Mode controls the way the vertices stick to the surface they
are wrapping. The name of each option gives a good description of how
each behaves but Inside and Outside need a little clarification. If
Outside is chosen then, where a vertex is already outside the target
object by more than the Offset amount, it won’t be moved, only
vertices that are inside the target object (plus Offset) will be moved
outwards. This is useful where garments have pleats, pockets or collars
that you don’t want to get “sucked” in against the target body.

Basic Alterations to the Mesh

You will now add some features to the shirt creating a collar and a
breast pocket.

132
 Extrude and Tweak

12. Switch to Edit mode and using front and side orthographic
views move the vertex at the neck away from the center line to
form an open collar as in Figure 7-4 on the left.

13. Select the vertices around the collar, extrude them in the Z
direction then, by rotating the view often and moving each
vertex a little at a time, create a folded over collar. For this kind
of adjustment it’s often handy to switch the select mode to
Tweak by long pressing the top entry in the Toolbar. In tweak
mode you can select and drag a vertex with one mouse click.
The result should look something like Figure 7-4

Figure 7-4 Extrude then form the collar

Sometimes it’s tricky to see what’s happening in Edit mode if there

are modifiers, such as shrinkwrap or subdivision acting on the object.
You can control which modes show or don’t show the modifier by
toggling the four buttons next to the name in the modifier controls, in
Figure 7-1 and 7-2 you can see that three button are on but the left-
most, On Cage, is off. Use the X-Ray view in order to be able to see
hidden vertices and, occasionally, for modifiers such as shrinkwrap, it
helps to to switch On Cage on. For the subdivision modifier it is
usually better to turn Realtime off, so you don’t see the effect of the
modifier in the 3D Viewport. Juggle these different settings as you
work to keep track of what you are changing and what the final result
will be.

Usually, when creating a mesh like this, your objective will be to

make something as simple as possible, but which can use a subdivision
modifier to create the desired detail. The target mesh here, the
Person01 object, is rather angular and coarse so the results will be
peculiar, but it’s still worth checking what the problems will be. The
next thing you are going to do is add a breast pocket on one side of the
shirt which will break the symmetry.

133
 14. First of all apply the mirror modifier, then add a subdivision
modifier and move it to the top of the stack. If the shrinkwrap is
applied after the subdivision it will cover the bumps better.

15. Several issues will now be apparent, the most obvious of which
is the skin poking through the shirt. Although you will solve
that problem with a mask, which you will use later in this
chapter, you should see what the effect is of changing the
shrinkwrap modifier Wrap Method to Target Normal Project
and increasing the Offset to 0.003m.

Bevel
The second issue is that the collar has smoothed out the sharp bend
and reduced its width. This will also be a problem when you add the
shirt pocket, as you will see later. The solution is either to add some
edges near to the sharp corner to help define the subdivision process or
alternatively, add a bevel modifier before the subdivision. There is
actually a third options that applies to the subdivision modifier and that
is to increase the edges’ crease value. For this exercise all three
approaches will produce similar results and have similar drawbacks,
but you will use the one that creates new edges in the mesh. Being able
to see the edges will help you understand and solve the problems
caused by non-quad faces later in this exercise.

16. Bevel the mesh by switching to Edit mode and selecting the
vertices along the inside edge of the collar. Then select
Edge4Bevel Edges, move the mouse to create the bevel then
LMB to fix the it. Don’t worry about getting the right amount
for the bevel as you will now open the Adjust Last Operation
pane and alter the Width to 0.01 and Segments to 3. You will
have to adjust these values to suit your model, in my case I
found that the shrinkwrap modifier was pushing the the
underlying mesh through the layer above. Even then I needed to
tweak the vertices in a similar way to moving the control points
to define the nurbs surface in Chapter 3. Figure 7-5 shows the
before and after screenshots.

134
Figure 7-5 Tweaking vertices to fix the bevel for subdividing

The Knife Tool

Your next job will be to create a breast pocket for your shirt using
the knife tool to create the extra edges and vertices in the desired
location. In the description below I refer to using the tool from the
Toolbar menu but when you are making lots of changes to a mesh using
different tools you will find it more convenient to use the shortcut. In
this case it’s easy to remember; K.

17. Still in Edit mode select the Knife tool from the Toolbar and
draw a rectangle where the breast pocket will go. See Figure 7-6
stage 1. You will find this easier if you toggle the shrinkwrap
and subdivision modifier Realtime option off so they don’t
distort or hide the mesh in the 3D Viewport. Vertices will be
created where you LMB or where the red line crosses another
edge. There is automatic snapping to edges or vertices so this
tool is very easy to use. When the shape has been defined, press
ENTER to fix it.

18. To make the patch shape into a pocket, extrude the rectangle
you have created by a small amount as in Figure 7-6 stage 2, but
note that when you enable the subdivision modifier it distorts
the pocket drastically, as seen in stage 3.

19. The simplest way to stop the edges being smoothed like this is
to add a bevel. Select the pocket edges, but excluding the ones
crossing the center of the rectangle and bevel them, in Adjust
Last Operation set Segments to 1 and an Offset of 0.001. Now
the subdivision and shrinkwrap should look much better, see
Figure 7-6 stages 4 and 5. However the faces around the pocket

135
 look distorted because they are no-longer quads. That’s what
you should tackle next, it’s the hard part of re-topology, but it’s
also the most satisfying.

Figure 7-6 Cut, extrude and bevel a pocket

There are two ways to revise the arrangement of edges in an

existing mesh. The first is to select edges that are in the wrong place
then press DELETE and use the Dissolve Edges option, which leaves a
larger face, then add back in edges in the correct location using knife,
loop cut, bevel or subdivide. The second method is to delete faces and
edges leaving a hole, then fill them back in manually using F.

Both the above methods are fine and useful in different

circumstances, but they cannot be easily be mixed without the danger
of creating faces on top of each other. Figure 7-7 shows the sequence
using the first technique. You will essentially use the other technique
for the whole second half of this chapter.

Figure 7-7 Using the knife tool to create quads

136
 20. Select each edge that needs to be removed then
DELETE4Dissolve Edges you can probably find some edges
that will immediately turn two triangles into a single quad.

21. With the knife tool divide up polygons with more than four
sides. Your objective is to make quads with corner angles
neither to acute or too obtuse. If Figure 7-7 stage 2 you can see
that the quad above and to the left of the pocket has three
vertices nearly in a line which will limit the ways it can bend,
and should be avoided.

22. In Figure 7-7 stage 3 I have used Edge4Loop Cut and Slide
to reduce the gap between the bottom of the pocket and the
horizontal edge below it. By Figure 7-7 stage 5 the strange
distortions around the pocket have become much improved.
Before you start constructing the mesh for the face it’s worth
spending a little time reviewing methods to avoid triangles and also
how to cope with bits of skin poking through clothing, which can
happen even with the optimum shrinkwrap settings.

Avoiding Triangles

If you look at my mesh in Figure 7-8 stage 1 you can see that,
although there are lines of edges looping around the shoulder and under
the arm, there are areas where it seems impossible to get rid of all the
triangles. The problem is that if a polygon has an odd number of sides
it can’t be divided up into quads, even if extra vertices are added in the
middle. And, if an extra vertex is added on one side to make the
number of sides even, then the polygon sharing that side gains an extra
side. Figure 7-8 stage 1 shows my shirt mesh to start with, then stage 2
shows it with all the obvious triangles removed, leaving a problem
polygon with seven edges.

There are different ways to solve this problem but they all boil
down to converting polygons to even sided ones. Sometimes there are
two polygons near to each other that allow edges to be moved so one
gains an edge and the other loses one. In this case there is no obvious
way to do that, so the simplest solution is to add a loop cut as in Figure
7-8 stage 3 or, better, stage 4.

137
Figure 7-8 Converting polygons to quads

When making the better shaped quads in Figure 7-8 stage 3 and 4 I
moved the vertices along edges using Vertex4SlideVertices to keep
the mesh shape as close to the body shape as possible. Though the
shortcut in the menu for sliding vertices or edges is SHIFT-V you can
also press G twice, which is more like the normal move shortcut so
might be easier to remember.

The Mask Modifier

You may have noticed, while working on the shirt, that bits of the
Person01 mesh poke through, however much you adjust the shrinkwrap
settings. The classic solution to this problem is to add a mask.

1. Unfortunately the mask modifier doesn’t work well with the

shrinkwrap modifer still on the clothes, so, once you are happy
with the shirt topology, apply the subdivision and the
shrinkwrap modifiers.

2. Remember you also need to regenerate the armature modifier

now the mirror modifier is no longer there, so select the shirt
then SHIFT select the armature and click
Object4Parent4With Automatic Weights

3. Create a vertex group on the parts of the underlying mesh you

want to hide. Select the Person01 object then in Edit mode
select all the faces that will be hidden by the shirt and add them
to a new vertex group “under_shirt”.

4. Add a Mask modifier to the Person01 object and specify the

under_shirt vertex group. Click on the Invert button as the
default behavior is to show rather than hide the vertex group.

138
 For such a coarse mesh as the body of Person01 you will find it
difficult to get the balance of vertices to include in the under_shirt
vertex group. You will probably have to add and remove them in a
process of trial and error.

Re-Topology

For the second half of this chapter you will create a new, clean
mesh for the face that you sculpted in Chapter 5. Take a quick look
back at Lola’s face in Figure 6-4. You can see that the quads form
loops around the eyes, mouth, nostrils and ears. Also the edges of quads
run along natural ridges or grooves on the face. Something to bear in
mind when creating a mesh like this is that it is very easy to subdivide
later by adding loop cuts or bevels, so keep everything as coarse as
possible to start with. There is a bit of setting up required:

1. You will start your new head topology from the simplest
possible mesh, so add a new plane to the scene and immediately
switch to Edit mode. All the vertices will already be selected so
scale the whole thing down to about the size of an eye and move
the plane up and to one side so it is half in and half out of a
cheek.

2. Rename the plane in the Outliner Editor to new_head. Add a

mirror modifier to the new_head object, which you can do
while still in Edit mode.

3. Change the snap mode to Face and switch it on.

These first three steps are essential, the next four just enable you to
see what you’re actually doing and they make the process much easier.

4. In the Object Properties tab, Viewport Display section, toggle In

Front to on.

5. In the Active Tool and Workspace settings tab, under Options,

toggle Auto Merge on. As you work you will probably have to
adjust the Threshold value, as well as the merge threshold in the
mirror modifier. Coarse settings make it easy to place vertices
so they snap together but sometimes you will find that a vertex
jumps from where you placed it because it was too near to
another vertex.

139
 6. Set the select tool to Tweak mode and the selection mode to
vertex and edge. It is possible to use more than one selection
mode at once by pressing SHIFT when you click on the buttons.

7. Use Viewport Shading set to Solid and in the options drop-

down set Backface culling on.

Extruding and Filling

Once you have set up snapping to a target surface, and you can see
the mesh as you work on it, you are good to go. For the beginning of
this exercise you will use the traditional extrusion and tweaking
system. There is much to recommend this method but we will cover
some additional tools that are more efficient in some circumstances.

8. Observe how the snapping onto a surface works by moving

each of the vertices of the new plane a small amount. Select one
of the side edges and extrude it then rotate it and repeat a few
times to form part of a ring of quads going around the eye.
Figure 7-9 stage 1 shows the ring complete but needing the last
vertex to be dragged so it snaps onto the corner of the starting
plane.

Bridging Edge Loops

After a bit of practice extruding, rotating, scaling and positioning
the edge of the new plane, this method becomes rather efficient. You
can probably see how you could tile the whole head using this
technique. However for rings of quads such as these around the eye it is
possible to use the edge loop bridging method you used to make the
sail in Chapter 3. Although you could use just two loops, one at the
inside of the existing faces and one on the inside of the eyelid, and set
the number of cuts as required, for good topology you should aim to
position the edges more precisely.

There are natural ridges and creases around the eye and you should
aim to make the edges of the new topology coincide with these.

140
Figure 7-9 Extruding edges and bridging edge loops

9. First make the edge on the inside of the eyelid by selecting the
edge loop around the inside of the ring of quads you just
created, then duplicating it and scaling it down a little. For this
stage of your work it is worth toggling the Auto Merge off in
the Active Tool and workspace tab, as you want vertices to stay
where you put them.

10. Move each vertex of the new ring to lie at the junction of
eyeball and the inside of the eyelids as in Figure 7-9 stage 2.

11. Repeat the process for the ridge on the edge of the eyelid and
for the crease above and below the eyelid as in Figure 7-9 stage
3

12. Select all four loops and Edge4Bridge Edge Loops.

Sometimes the arrangement of vertices on sharply curved
surfaces can cause confusion for Blender which might generate
crossed edges and overlapping faces. If this happens you have
to select a smaller number of loops and fill the area in a number
of stages. For instance you should be able to bridge the outside
three loops unambiguously then bridge the inner two edge
loops, where there is a sharp fold, with a single ring of faces.

NOTE There are no simple keyboard shortcut for Edge4Bridge Edge

Loops and Face4Grid Fill so it might be convenient to add them
to Quick Favorites by RMB on the menu entry. This can then be
accessed later by pressing Q

13. To work on other areas of rings, extrude a vertex out towards

the ear, nostril, mouth and neck and build up rings of quads in a
similar way to the eye. You can see in Figure 7-10, where the
faces have been built for the nostrils, that Blender sometimes
points the normals inwards instead of outwards. This happens if

141
 you bridge loops without any references faces around the edge.
You can fix this by selecting all the vertices facing the wrong
way, then select Mesh4Normals4Flip. However to avoided
this problem in the first place you should always extrude out
from an edge so that at least one part of the edge loop can be
correctly orientated when you start filling in quads.

Figure 7-10 Flipped normals around nostrils

Filling with One Face at a Time

In Figure 7-ll stage 1 you can see that filling in the ear with edges
along the ridges and grooves runs into a problem where there is a
relatively featureless area in the center but there needs to be more
detail, and a hole, at one side. If concentric loops continue to be added
they will become far to congested below and to the right of the ear
hole. Rather than using bridging edge loops, you can add quads and
edges one at a time using Vertex4New Edge/Face from Vertices, or
in practice, the shortcut F.

142
Figure 7-11 Filling in the ear

14. Figure 7-11 stage 1 shows the stage where two loops have been
duplicated and scaled down but not yet bridged because they
have started to overlap the ear hole.

15. To create a single quad select four vertices or, more easily, two
unconnected edges opposite each other, and press F. You can
see the new quad formed in Figure 7-11 stage 2.

16. Select one of the sides of the newly created quad which is
facing down a “track” of unconnected edges. In Figure 7-11
stage 2 you can see the edge I have selected. Now press F and,
not only will a new quad appear, but the next edge down the
track is automatically selected so you can press F repeatedly to
fill in a strip of quads. This is actually a very fast and
controllable way to build the mesh.

17. In Figure 7-11 stage 3 the circumferential quads are being filled
as far as the awkward area in the bottom right. In stage 4 the flat
area on the left was chosen to have an even number of edges
around it, which are now filled in. The hole on the right, also
has an even number of edges.

18. Figure 7-11 stage 5 shows what sometimes happens with

automatic merging in confined spaces. One of the vertices from
the duplicated edge loops forming the bottom of the ear hole has
snapped to merge with a vertex on the loop outside it. Don’t

143
 worry if this happens. You can detach the miscreant by selecting
the vertex and hitting Vertex4Rip Vertices.

Joining Mismatched Quads

A perennial problem in re-topology is joining a neat arrangement
of quads in one area with the quads emanating from another area.
Figure 7-12 stage 1 shows an example where the smaller quads needed
around the nostrils and tip of the nose meet up with the quads from the
eye. You can see that an elongated quad between the nose and the
cheek has effectively turned the corner from around the eye to across
the nose, but the next rows of quads going up the nose will all have to
be merged into a single quad. Ideally the next few rows of quads would
be fairly square, like the ones in Figure 7-12 stage 2.

The first thing to notice is that there have to be an odd number of

rows of quads merging into one row as that creates an even number of
sides to the polygon. You can see, in the middle of Figure 7-12 stage 2
that there will be a purple rectangle with six sides once the top vertices
are joined. Stage 3 shows one reasonable way to fill it, there are hosts
of other possible arrangements.

Figure 7-12 Merging one set of concentric faces into another

You can see in Figure 7-12 that the mesh at the top of the mouth
could be joined to the nose and eye relatively easily continuing with
this extrusion technique but, instead, I will introduce you to a useful
alternative.

144
 Poly Build
When you were using the method of bridging edge loops you
probably noticed that, after you duplicated a loop, the new vertices
didn’t snap to the surface until you had tweaked each of them in turn.
Even if they were in approximately the correct position when viewed
face-on, you had to move them around a bit to force the snap to work.
So, in some ways it might have been just as efficient to create each
quad individually if that could be done as succinctly as extrusion. Well,
in many ways, it can, using the Poly Build tool.

1. Select the Poly Build tool from the Toolbar, then move the
mouse around noting when vertices or edges turn blue. Press
and hold CTRL and move the mouse around noting when
triangles or quads appear. Now hold SHIFT and see when
vertices or faces become highlighted red for deletion.
Hopefully you found some difficulty in a couple of locations. This
is the reason, in my opinion, that the Poly Build tool isn’t the perfect
way to do all re-topology. It’s often quite difficult to make it select the
right edge, vertex or face for extrusion or deletion. However sometimes
it’s very efficient so it’s worth knowing how to use it.

2. Move the mouse until a corner vertex is highlighted blue as in

Figure 7-13 stage 1. Now press CTRL to see the new quad, if
you press CTRL when an edge is highlighted you will get a new
triangle. I find that it is better to press and hold LMB very
quickly after pressing CTRL, before poly build changes its
mind. The alternative approach is to press CTRL all the time
and move the mouse around to get the quad forming where you
want it. The latter method is probably quicker in simple areas
but the former method seems less frustrating where the
topology is already congested. Drag the new vertex to create the
correct shape for the quad, as in Figure 7-13 stage 2.

3. To extrude a new quad from one edge of an existing quad, press

and hold LMB while an edge is highlighted as in Figure 7-13
stage 3. Move the remaining vertex after extrusion by dragging
a single blue dot shown in Figure 7-13 stage 4, exactly the same
as when using tweak select.

145
Figure 7-13 Using the Poly Build T ool

Use the poly build system to fill the areas joining together the
mouth, nose, eye, ear and neck. Try to figure out how to divide up the
shrinking spaces as you expand the areas towards each other.

As an example, Figure 7-14 stage 1 shows the underneath of the

chin once I had expanded the rings around the neck up and joined them
to the columns that had developed running vertically down between the
ear ring and the eye ring.

Figure 7-14 Filling under the chin

To start with, the purple area was surrounded by eight edges which
could have been divided up into quads, but that would have left a very
long edge along the center line, which actually needs a slight curve. To
keep an even number of edges around the area to be filled, I had to
divide the long edge into either three, as in Figure 7-14 stage 2, or five
as in Figure 7-14 stage 3.

146
 This section on using poly build would not be complete without a
mention of two other add-ons that are included in Blender, but not
enabled by default. The first is F2 which changes the behavior of the
shortcut F such that if a corner vertex is selected, in a similar position
to the the blue highlighted one in Figure 7-13 stage 1, you can create a
new quad in a very similar way to poly building. The second add-on is
Bsurfaces which allows you to draw lines onto a surface using the
annotation tool, and then create additional mesh surface by converting
the lines to edge loops and filling between them, in a similar way to the
edge bridging you used earlier. Bsurfaces can give a massive increase
in productivity, especially if you have facilities to draw with a graphics
tablet rather than the mouse, however you should learn the basic,
manual methods first, in order to derive the most benefit from any add-
ons that you decide to use.

Grid Fill
Grid Fill is a method for instantly filling areas with mesh, very
similar to bridging edge loops. Try using it to complete the mesh for
the cranium. In Figure 7-15 you can see the approach I took to create a
rectangular area to fill.

Figure 7-15 Grid Fill the cranium

147
 1. Start by building a line of quads over the top of the head then
mark out a rectangle so each side has the same number of
vertices as the opposite side as shown in Figure 7-15 stage 3.

2. Fill the triangular areas at the front and back of the head with
quads. You might need to add or remove edge cuts in order to
make the number of edges around the perimeter of each triangle
an even number. Don’t worry if the sizes of quads are uneven,
that will make the next section, where you use the relax tool,
more impressive.

3. Select the edge loop around the outside of the rectangle

ensuring that one of the corner vertices is “selected” in the
Blender sense, showing yellow rather than orange. In Figure 7-
15 stage 4 you can see that the top left corner of the rectangle is
the selected one. You might have to deselect a corner vertex
then reselect it.

4. Select Face4Grid Fill and the area should be tiled with quads
correctly. If the result is arranged diagonally or has triangles in
places check that the edge numbers are correct and that a corner
vertex is showing yellow. It is also possible to tweak the
arrangement in the Adjust Last Operation pane.

Sculpt Slide Relax

Sometimes you will end up with an area of quads that you can see
are imperfectly laid out, for instance in Figure 7-16 stage 1 there are
elongated rectangles and quads that are almost triangular. Rather than
move each vertex individually, you can use a relax brush to quickly
tidy up the geometry.

Figure 7-16 Using Sculpt Slide Relax

148
 5. Switch to Sculpt mode, select the Slide Relax brush and adjust
the size and strength so it doesn’t accidentally move edges that
you have carefully placed on ridges or groves.

6. To relax the edge orientation, press SHIFT and hold LMB as

you paint with the brush as shown in Figure 7-16 stage 2. You
need to paint carefully and not go too near the ear or eye.

7. You can also move edges towards areas where you want a
higher concentration using the same brush without SHIFT.

Shrinkwrapping
Both grid filling and edge bridging, as I mentioned earlier, do not
snap the vertices they create to the underlying mesh. If you
subsequently relax the mesh in Sculpt mode, rather than moving each
vertex individually, the vertices still won’t be snapped to the mesh.
However shrinkwrapping gives you a quick way to fix this.

8. Switch to Object mode, add a shrinkwrap modifier and, if you

have any other modifiers such as mirror or subdivision already
there, drag shrinkwrap to the top of the stack.

9. Set the Target of the shrinkwrap modifier to the Person01 object

and check that the vertices are being fitted nicely to the target
mesh. If there are problems try changing Wrap Method but
leave Snap Mode as On Surface. If everything looks OK then
apply the modifier.

Fine Adjustments with the Subdivision

Modifier
Towards the end of your mesh building exercise, and certainly
before you apply the mirror modifier, it’s a good idea to add a
subdivision modifier to see where you need to add extra topology to
help define ridges and creases. The eyebrows, eyelids, ears and lips
will probably all need extra loop cuts adding. When doing this final
tuning it will be much less frustrating if you turn off auto merging and,
for some fine adjustments, snapping.

A final tool that will help a lot as you make adjustments is

Edge4Slide. Pressing G twice works as a convenient shortcut for this,
and as I mentioned earlier in the chapter, this is also the way to move

149
 vertices without changing the overall shape of the mesh when you have
toggle snapping off or don’t have a model to snap to.

Join Two Meshes

To wrap up this exercise you should stich the new head onto the
old body and re-apply the armature modifier.

10. Make sure you have saved your blend file.

11. Switch back to object mode, select the Person01 object and
select all the vertices apart from the eye meshes by selecting a
single vertex then hitting Select4Select Linked4Linked then
deselect everything below the neck with CTRL and box select.

12. Delete the old head!

13. Back in Object mode select first the new_head object then the
Person01 object and Object4Join

14. The new_head and old body are all one mesh now but there is a
gap between them that needs to be filled. Switch back again to
Edit mode and join up the vertices from the top and bottom of
the neck. You will have an even number on both sides of the
gap so it will be possible to do it entirely with quads.
Figure 7-17 shows my attempt.

Figure 7-17 Stitching the head onto the body

15. You can merge vertices by selecting two then selecting

Mesh4Merge and choosing First or Last, so be consistent
keeping either the head or body location and moving the other
to match. Change the shape of the new quads to make them
more regular by moving vertices using G G.

150
 16. Finally, to recreate the armature modifier and bone weights for
the new vertices, in Object mode select the Person01 object first
then the armature and select Object4Parent4With
Automatic Weights.

Shape Keys, an Alternative to Bones

Shape Keys are a way for Blender to store multiple different

positions for the vertices of an object. They then allow the user to slide
a controller to smoothly transition the vertices from one position to
another. In some ways they do overlap with the functionality of an
armature but they can be used where vertices need to move in different
directions which would be hard to do with an armature, requiring
complicated linkages of bones. Shape keys are frequently used to add
expressions to the faces of characters such as the Person01 object and
you will now make a very simple “smile” shape key.

1. To add a shape key, select the Person01 object then in the

Object Data Properties tab under Shape Keys click +. However
there always has to be Basis key and this will be the name of the
newly created shape key. Click + again then change the name
from “Key 1” to “smile”.

2. In the list select the smile shape key then switch to Edit mode.
Now, when you move the vertices they will represent the shape
of the mesh when the smile shape key is set to 1.0. This means
that at a later stage, in order to see, or edit, the basis mesh you
must re-select Basis from the list of shape keys. For the moment
leave it on smile.

3. You might want to make a crooked smile but it will probably be

easiest to set the X symmetry on. Also switch on proportional
editing and change the diameter of the influence to something
reasonable, say, about the diameter of an eye. In the
proportional editing drop-down options you should tick the
Connected Only check-box, otherwise you will distort the eyes
as you move the eyelids to form the crowsfeet creases when
smiling.

4. Now tweak the edges of the mouth, indent the bottoms of the
cheeks and crinkle the eyes. When you’re happy switch back to
Object mode.

151
 5. Try the slider on the smile shape key as shown in Figure 7-18.
You can add other shape keys, and when you edit the mesh with
each shape key selected you will create the positions of vertices
equivalent to that specific shape key being set to 1.0. However
in Object mode you can position several shape keys to different
values and the results will all be combined.

6. As an experiment try setting the value of Range Min to -2.0 and

Max to 3.0. You will find that you can “overclock” the sliders a
little, but too far will mangle the mesh!

7. Don’t forget to switch proportional editing and X symmetry off

when you have finished editing. Along with forgetting to switch
off snapping these are often the cause of Blender “stopping
working” and they are inconspicuous settings that will take a
long time to spot!

Figure 7-18 The Smile Shape Key

Conclusion

This chapter has packed in a lot of material. I hope you found the
process of re-topology to be an enjoyable challenge rather than a
frustrating chore, but either way you need to understand the problems,
and how to solve them.

If you ever find yourself needing to do much re-topology work

then you should invest in one of the various excellent add-ons
available, especially as some of them are free! However, knowing how
to achieve the results manually, will enable you to judge which tool to
use when and allow you to appraise how good they are at doing what
they claim.

152
 You used many mesh editing tools, some new ones and some
familiar ones being used in a new way or with some additional subtlety.
Many of the techniques you used here to make a “soft” mesh that could
be deformed cleanly, are actually identical to those needed to make
good “hard surface” meshes. Dividing areas into quads, and placing
edges in strategic positions is a requirement running through the whole
of 3D modeling. Your appreciation of the problems involved, and their
solutions has moved incrementally forward and it will help you in
many ways over the rest of this book and thereafter.

At the end of the chapter you glimpsed the potential of Shape

Keys. They can be used in a multitude of ways beyond making facial
expressions. In many ways they belong in Blender’s arsenal of
mechanisms for controlling animations, so in Chapter 12 you will use
Shape Keys again and see how they complement constraints and
drivers.

In the next chapter you will unwrap UV maps in a much more

controllable way than you used in Chapter 5. And in the following
chapter you will learn how to transfer the detail that you sculpted on
your original head to a normal map texture for including in the material
definition.

153
 8
F L AT PA C K F U R N I T U R E

Gaining more control when

unwrapping a 3D object to form a
flat surface

Blender does a good job of automating the use of UV

maps and you can get quite a long way without having to
worry about them at all. At some point, though, you will
need to apply more control than is available with the
default shapes and the smart project option.
In Chapter 5, when you started texture painting, you used smart
UV project to create the UV map for the subsequent painting. As the

154
 chapter progressed I pointed out problems caused by this imperfect
method of unwrapping. In this chapter we will revisit the same model
but use the mesh that you created in Chapter 7 and apply some more
sophistication. You will see that it is not only possible to make the scale
of the UV mapping correspond with the level of detail you want to
paint, but it is possible to use different UV maps for different parts of
the model and use them to apply different resolution textures to
different materials.

Different Ways of Unwrapping

For making a realistic model of human character it would be quite

normal to divide the mesh into separate materials for the head, eyes,
torso, arms and legs. However your Person01 object is far from
realistic so it will be sufficient to just use three materials, head, eyes
then everything else. Links to the files and videos for this exercise are
at get-into-blender.com 8. FLAT PACK FURNITURE.

1. Open the blend file you saved at the end of Chapter 7 with your
completed, re-skinned, Person01 object. Save it now with a
different name in the folder for this chapter so you don’t
accidentally overwrite it.

2. Select the Material Properties tab. You might still have two
materials left from Chapter 5, if so rename them and add a third
one. You want a head, eyes and body material.

3. For this exercise you will need materials set up as in Figure 8-1
for the head and body, and one without the subsurface texture,
bump texture and normal input for the eyes. You will probably
have a similar setup from before called skin and eyes so modify
the skin one and save it as head then use the New Material
button, with the overlapping squares icon to the right of the
Name field, to duplicate that and rename the duplicate as body
and use eyes as it is.

155
Figure 8-1 Material Nodes for head

4. There are probably existing images from the old file which you
can rename. Add some new ones so, in total, you have;
head_base_color, head_subsurface, head_bump,
body_base_color, body_subsurface, body_bump and
eyes_base_color. To get the detail for a realistic face you should
make the textures at least 2048 by 2048 pixels, they won’t need
the Alpha option. You can create images in the node in the
Shader Editor panel, using the menu options to the right of the
image name. You can increase the resolution of existing images
if you reuse some by opening an Image Editor in a work area
and using Image 4Resize.
Once the materials are set up ready you need to create the UV
maps for the textures to use. The first job is to draw seams where you
want the skin to be “peeled” back. The standard locations are round the
neck, up the back of the head, down the back of the body, down the
back of each arm and down the inside of each leg. The eyes can be
sliced around their circumference. See Figure 8-2 for the seams and
unwrapping on the head.

5. Switch to Edit mode and select the edges that will be seams.
You can do them one seam at a time rather than trying to select
them all in one go.

156
 6. Select Edge4Mark Seam and the selected edges will be
marked red.

7. Select the head vertices down to the neck seam, but excluding
the eyes. Use Select4Select Linked4Linked then deselect
with CTRL box select from the neck seam down, with X-Ray
on.

8. Change the bottom screen area to UV Editor then in the 3D

Viewport select UV4Unwrap. You should get something like
Figure 8-2 stage 1

Figure 8-2 Unwrapped Head

It is now possible to edit the UV map in a similar way to editing

meshes in the 3D Viewport. However before you do that it’s a good
idea to temporarily swap the head_base_color image for a UV grid.
Then, if you set the viewport shading to Material Preview, the UV grid
will cover the head and you will be able to tell where the UV map is
becoming too badly distorted. Your objective is to scale up the central
part of the UV mesh so it can contain more detail, at the expense of
scaling down the bits of the mesh that are under the hair or behind the
head.

9. Change the bottom area back to Shader Editor and make sure
you have the head material selected. In the node providing the
texture input head_base_color click the button to create a new
image then select from the Generated Type drop-down UV

157
 Grid and click OK. Your head should now look like Figure 8-2
stage 2.

10. Change back to the UV Editor. Select all the vertices by

pressing A while the mouse is over the UV Editor, then rotate
them by -90 as in the lower half of Figure 8-2 stage 3.

11. You need to keep the squares as regular as possible but make
them smaller in the middle of the face. Toggle proportional
editing on with linear falloff, then select a vertex on the tip of
the nose in the UV Editor. Scale the vertices up, adjusting the
radius of the proportional effect so no overlaps or serious
distortion occurs. You can see my attempt in Figure 8-2 stage 3.

12. Repeat the unwrapping process for the body and the eyes as
shown in Figure 8-3. There is no need to swap the texture to use
the UV grid image for these two.

Figure 8-3 Unwrapped body and eyes

13. Change back to the Shader Editor in the bottom area and in the
drop-down for the base color texture for the head, select the
head_base_color again. Ensure that this node is selected and is
outlined with a white line.
You should be set up correctly now for texture painting. In the past
I have saved the blend file at this point and, when I opened it later, all

158
 the image files I had created vanished. So I always use the Save All
Images button in the texture paint settings before I save, just to be on
the safe side.

14. In the 3D Viewport switch to Texture Paint mode and in the

Active Tool and Workspace setting tab set the brush, color and
other settings as you did in Chapter 5. You should see the image
texture being over-painted in the Image Editor panel on the left.
You should be able to get much better detail now, and the seams
are hidden away, although you will need care to match the different
textures across boundaries.

Don’t spend too much time painting, as I mentioned in Chapter 5,

unless you are an exceptional artist it will be hard to make realistic 3D
objects by texture painting. So, for the rest of this chapter you will
learn how to use photographs as textures.

Taking Photos for Textures

There are several problems when taking photos to use as textures

in Blender materials. The image should ideally contain accurate
information about the base color of the object that you are trying to
represent. However photos always capture the light being reflected
from the surface of the subject, which depends very strongly on the
lighting conditions at the time. If you take a photo of something
outdoors in strong sunlight, at twilight or under strong artificial
lighting, the results will all be very different. In addition, there are
almost always shadows caused by the direction of the incident light
which will duplicate the shadows calculated by Blender when it does
its rendering. Most cameras will also modify the light being recorded to
match the sensor, often moving the profile of colors up or down the
spectrum to maintain a “white balance”. And finally, photos are
necessarily a perspective view of just one side of whatever object is
being recorded, so even though Blender can unwrap a 3D object to 2D,
it’s not possible to take that “all around” photo with any normal kind of
camera.

Solving all these problems well takes quite a lot of expensive

equipment, but it’s possible to do a surprisingly good job using just a
phone, some reasonable lights and the clever functionality built into
Blender. Often flash can help to help fill in shadows, and if you have

159
 access to some basic photo-editing software such as GIMP or
Photoshop you will be able to improve your results even more.

Applying Images to an Existing Mesh

You will need some photos of a willing subject to use for the
head_base_color texture. Take a profile and a frontal image as a
minimum but ideally use other pictures as well. You can see the set of
mugshots I used in Figure 8-4. Give them distinct, meaningful names
such as “full_face”, “face_slight_left”, “right_side” and so on.

Figure 8-4 Mugshots. That’s me looking very serious.

Cloning from Perspective Images

When you used the texture painting in Chapter 5, you may have
noticed other tools apart from Draw. Soften, Smear, Clone and Fill all
work in a similar way to other painting or image editing software.
However there is a useful variation for cloning which allows you to
paint onto a texture using, not only a different texture, but a different
UV map. This means that it is possible to paint from several different
perspective views of a real head onto the head_base_color texture using
the clone tool.

1. You now need to set up each of the mugshots with a perspective

UV that you can use to clone from while painting. Change the
workspace layout to UV Editing using the tab at the top of the
window, then from the menu in the UV Editor area choose
Image4Open and navigate to the first of your images,
full_face.jpg, for instance.

2. Create a new UV map for this image by opening the UV Maps

section of the Object Data Properties tab and selecting + New.
Rename the UV map in line with the image, UVfull_face, say. If
you look at the top of the UV Editor area you can see that the
currently selected UV map is displayed in the header and you

160
 can change the current selection here as well as in the Object
Data Properties tab.

3. You can now make a UV map to match the view of the head in
the image you just loaded. To do this you must adjust the
camera position in the 3D Viewport until the model looks as
close as possible to the image showing in the UV Editor. It is
much better to do this using the camera view, positioned using
walk navigation, because it allows you to change the Focal
Length value in the camera Object Data Properties tab. Mobile
phone cameras have very short focal lengths so you probably
need to reduce the default 50mm setting. Using the camera view
also allows you to return to the exact same position you used to
generate the unwrapping if you change the view part way
through for some reason. Figure 8-9 stage 1 shows my attempt.
When it’s as good as you can get it, in Edit mode with select
mode set to faces and all the head faces selected, run
UV4Project from View.

4. The UV map is unlikely to line up with the image in the UV

Editor area, so you will need to ease it into position. To start
with it will be rather a jumble as all the faces from the back are
overlapping the ones from the front, as shown in Figure 8-5
stage 1. Start by scaling in the X and Y direction then use
proportional editing with a linear fall off, large radius and
Connected Only to nudge the UV map to line up reasonably
well with the image.

5. You should aim to use only the part of each image and
associated UV map, to paint the areas that are clear and
undistorted on that image, so concentrate on getting those bits
more accurate. Now you can gradually reduce the radius of the
proportional editing and line up edges precisely, as in Figure 8-5
stage 2.

161
Figure 8-5 UV Projected from View

6. Work your way through each image you intend to use. A critical
thing you need to make sure of before generating each new UV
projection is that you have selected a new map. If you
accidentally leave the UV map set to a previous one, when you
hit project from view, it will overwrite all your careful work.

7. You can reuse images taken from the left, say, for painting the
right by selecting Image4Flip4Horizontally in the UV
Editor and using the UV projected from the opposite view.

8. To clone the perspective images onto the head_base_color

texture, first specify the destination texture and UV map which
you should do in the header of the Texture Paint panel, visible at
the top of Figure 5-7. For Mode choose Single Image,
specifying head_base_color and the original UVMap. At this
point, in the Shader Editor, make sure you’ve changed the
texture for the head from the UV grid back to head_base_color.

9. To set the source to clone from, select the Texture Painting

preset from the workspace options tabs at the top. On the left of
the 3D Viewport select the clone tool from the Toolbar then, in
the Active Tool Properties tab, Brush Settings tick the checkbox
Clone from Paint Slot. Expand the details for this and specify
the Source Clone Image texture and Source Clone UV Map.
These options will not be available unless you have chosen

162
 Single Image Mode in the panel header. Each time you start
cloning double check these settings are correct as well as the
ones listed in the previous paragraph.

10. The first image you clone can be done using full strength for the
brush and you don’t need to be too careful where you paint but
subsequent cloning must be done carefully. Use a low strength
and build the texture with multiple brush strokes so you get
smooth transitions between the different areas.

11. You will probably find some small areas that are inadequately
referenced on any one of the mugshots, and leave streaks or
sharp changes in color on the head_base_color texture. You can
patch these areas by toggling off Clone from Paint Slot so the
cloning copies from one location on the object to another, very
much like cloning in a normal image editor program. To set the
location to clone from use CTRL-LMB. The rather disturbing
hybrid of my selfies cloned onto the Person01 head are shown
below in Figure 8-6.

Figure 8-6 Full Texture Cloned from Perspective Images

Using the clone tool to create textures for UV unwrapping in

Blender is powerful functionality, but it does require a bit of care.
Ensuring that the same edges in the UV map line up with the same
physical features on the photos is quite hard, and some trial and error is
often required to get a good result. You probably found that, even with

163
 good lighting, there were shadows and differences in tone when
cloning from different photos.

Images as References

While making the UV maps, it may have occurred to you that the
result would have been much better if the Person01 mesh had been
modified to match the shape of the head in the photos. That is
definitely the case, however leave the mesh unaltered for the moment
as you will use it in Chapter 9 to bake the normal map from the
sculpted head. Instead, you will make a an item of simple furniture
from your home using photos as references.

Don’t choose anything too complicated, your objective is to learn a

few more modeling skills and some techniques for applying textures as
“tiles”. I chose the sofa in Figure 8-7 because it could be constructed
from relatively simple shapes, it was all made from the same fabric, it
didn’t have a shiny surface which would have been hard to photograph
without reflections and because it was light enough for me to be able to
turn it round to face the windows.

Figure 8-7 Images of a Sofa

Reference images typically show the subject in profile and face on,
and can be used in Blender to help construct a 3D model. Ideally the
photographs should have as little perspective as possible, which
normally requires them to be taken from a reasonable distance.
However, the further the subject of the photo is from the camera, the
lower the resolution, so it’s a compromise. For my sofa the two
references are Figure 8-7 image 1 and 2.

For this exercise you are going to apply a photo of a small area of
the sofa as a material for the whole thing and, in order to make the
texture seamless when it’s tiled, the illumination should be as uniform
as possible. Figure 8-7 image 3 shows my image. If you have image
editing software available you can use that to make your image

164
 seamless, alternatively there are websites that offer a conversion
facility for free. Failing that, it is quite possible to make textures
reasonably seamless using Blender either with the compositor or
texture paint cloning and baking.

Modifying the Start Cube

Save the images on your computer and give them meaningful

names before you start building the furniture. I will describe the steps
as if you were making a sofa but, obviously, you will need to interpret
them to match the geometry you’re working with.

Building the Geometry

The approach you will take is to put the front and side images into
the scene as references then subdivide and extrude the start cube the
minimum that will provide a very rough sofa with the right proportions
and angles. You will use a mirror modifier and a subdivision modifier
then add edges and slide them to get the right amount of curvature and
sharpness of edges and creases. You’re not going to try to model fine
detail like piping or casters.

1. Start a new scene, you need to save the work on the Person01
UV unwrapping and cloning if you didn’t do earlier.

2. Bring in your two reference images. First of all switch X-Ray

on then in orthographic view from the Y direction use
Add4Image4Reference and select the frontal image. Scale
the image so the sofa is about the right size, the cube is 2m on
each side so that’s just right for my sofa. Move the image up so
the bottom of the sofa is resting on the origin. Switch view to
orthographic from the X direction and add the side image and
scale it to match the sofa height in the frontal image.

3. You can control how the image is displayed, I find it easier to

work if the image is behind the 3D work and only shows in
orthographic view. With each image selected, in the Object Data
Properties tab (red image icon) set the Depth to Back and un-
tick the Perspective checkbox.

165
4. Rename the cube to sofa and add a mirror and a subdivision
modifier.

5. Do the rough shaping of the cube. You need to divide it up

enough to extrude the back up and the seat down. It will be
easier at this stage to turn off the visibility of the subdivision
modifier in Edit mode by toggling off the second button. With
the cube selected switch to Edit mode and move the vertices so
four are on the mirror plane and the overall size matches half
the sofa up to the height of the top of the arms. Delete the face
on the mirror plane. Figure 8-8 shows the size I used.

Loop Cut

6. Rather than use the subdivide function you’ve used previously

you will add loop cuts where you need them. The whole
modeling exercise will require lots of edge loops so I suggest
you use the shortcuts CTRL-R to add, ALT-LMB to select and
G twice to slide them. In the front orthographic view, add an
edge loop and slide it to the position of the inside of the arm.
Because of the perspective on my image, this is a little hard to
estimate but you can see the overall width of the arm. The
vertical side of the back is about in the middle of the arm, so
add a second edge loop mid way between each side of the arm.
You can see the position on Figure 8-8 stage 1.

7. When you extrude the seat down, you will create some
unwanted edges and faces so it will help if you add a horizontal
edge loop to match the depth of the seat.

8. You need a final edge inserting to allow the back to be extruded,

so switch to orthographic view from the side and add a loop
there, see Figure 8-8 stage 2 and 3.

166
Figure 8-8 Start cube with Edge loops inserted

9. Select the two faces on the top of the backrest and extrude them
upwards. Extrude the seat of the chair downwards. Now you
will see the extra vertex, edges and face so delete these and fill
in the gaps using the shortcut F. Use Mesh4Merge to get rid of
the gap or overlap at the front of the seat. Figure 8-9 shows the
sequence and final result. Extruding the back will also insert an
unseen, unwanted face and edge, temporarily disable the mirror
modifier to see these and delete them. In orthographic view
from the side line up the backrest to match the shape in the
photo.

Figure 8-9 Basic extrusion and tidying up topology.

10. Your challenge now is to add the minimum number of edges to

produce a likeness of the sofa. When you toggle the subdivision
visibility in Edit mode back on, everything will be rather
rounded, but judicious placement of edges will create the
required balance of curvature. Figure 8-10 stages 1, 2 and 3
show the modeling after edge loops had been inserted. You can
see that you need to put three edges close together to form
creases around the seat (that place where all the coins and ball-
point pens disappear). When the overall shape is correct, delete
the extra edges on flat areas where they are not needed, using
dissolve edges, and finally sort out the few faces that are no
longer quads, using the knife tool. In Figure 8-10 stage 4 you
can see on the front of the arm where I had to divide up a six
sided polygon.

167
Figure 8-10 The Sofa after adding edge loops.

Define Seams and Unwrap UV Map

11. To generate a UV without too much distortion you need to

apply the subdivision modifier with one level.

12. Define the seams for the UV unwrapping. Most of these can be
tucked into creases or on edges at the back or bottom, however
you will need one or two in visible places on the arms to
prevent distortion of the UV mesh when it is unwrapped. You
will be able to get a better layout if your seams cut the surface
into islands so you can move and scale each one separately. The
seams in Figure 8-11 produce islands for each arm, the seat, the
back and the bottom. As the bottom will not normally be visible,
it can be shrunk, after unwrapping, to leave more room on the
image for the rest of the sofa.

13. Unwrap the UV mesh and adjust it to make the most use of the
area. Use the tab right at the top of the window to change to the
UV Editing preset for the active workspace, then in the 3D
Viewport area hit UV4Unwrap. You should get a result
similar to Figure 8-11 stage 1.

168
Figure 8-11 Unwrapping the main UV map

14. To scale an island such as the bottom of the sofa and to be able
to move all the islands around to optimize the space, there is a
useful fourth selection mode, island, which has an icon with a
solid and outline rectangle, visible at the top of Figure 8-8. To
select the island for the bottom of the sofa, in the 3D Viewport
deselect everything then select one face on the bottom. In the
UV Editor switch to island selection mode and select the single
face that you can see there. Now, back in the 3D Viewport,
select all the faces so you can see them in the UV Editor. Click
again on the single face from the bottom, which is still selected,
and you will select the whole island.
In Figure 8-8 stage 2 you can see my re-arranged UV layout and
there are a few things worth noting. Not only have I have shrunk the
bottom, I have also shrunk the back slightly, as that will also be less
carefully scrutinized than the rest of the sofa. More importantly, the
layout for the sofa arms look to point in opposite directions, which
might be a problem if the fabric was a print, say, where the design
needed to be the right way up. However if you look carefully you will
see that the UV unwrapping has oriented all the islands so that the
bottom of each area of fabric is on the right of the map. If you use an
image for the material texture covered in arrows pointing from right to
left, on the 3D object, they would go upwards in most locations, see
Figure 8-12.

169
Figure 8-12 Orientation of Panels

The fabric on my sofa has no obvious directionality so orienting

the panels is not an issue, however, as you can see in Figure 8-12, the
scale is wrong. The square of fabric should be about the same area as
the outside of one of the arms of the sofa.

Scaling UV Maps
There are two ways to achieve this. The most obvious, is to scale
the UV map up until the image is the right size to match. This what has
been done in Figure 8-13.

Figure 8-13 Scale the UV up to Reduce the Scale of the Image

This works fine so long as the UV map is only being used for a
single purpose. But in this exercise it would be nice to also feed a
normal map into the shader so that surface details, such as the piping
along the edges, or minor creases, can be represented. The UV map that
the normal texture uses must match the area of a single image
otherwise the creases will be repeated all over the sofa in arbitrary
locations. It would be possible to create a second UV map and feed that
into the input of the texture node, but actually it’s even easier to just

170
 use the default map but scale it up in the shader when it is being used
by the base color texture. Figure 8-14 shows the node layout on the left
and using the brush Stroke set to Curve for drawing piping on the
normal image on the right. I mentioned using curve strokes in
Chapter 5 when I introduced the basic workflow for sculpting.

Figure 8-14 Scaling UV Maps for Different Textures in Shader Editor

15. First of all set up the material. Open the Material Properties tab
and there should already be the default material that came with
the cube, rename it to sofa_material. Change the area on the left
from UV Editor to Shader Editor. Add a new image texture
node to feed into the Base Color input of the shader then click
on the Open button to import the seamless fabric image. Add a
subdivision modifier back to the sofa object to smooth out some
of the angles.

16. In order to see the result of changes to the material, in the 3D

Viewport switch the viewport shading to Material Preview. In
the Shader Editor click Add4Input4UV Map, it will default
to the correct one, but just to make it explicit, click on the little
dot with a drop-down selector and choose UVMap. Connect the
output to the input of the image texture node. This is how you

171
 could set different UV maps to be used by different textures in a
material.

17. Now you need to multiply the values being fed into the texture
node. Select Add4Vector4Mapping and drop the node onto
the connection between the UV node and the texture node, then
change the X and Y scales to 5.0. This node is also where you
could rotate the texture or offset it, if needed.

NOTE For simplicity I have not suggested switching the Type field from
the default Point, and normally this is fine. However, especially
when you want to rotate the UV map about a specific position, you
will normally find it easier to use the Texture option. If you do that
you will need to change your frame of reference for the
translations. i.e. scale of 5.0 with Point mapping becomes 0.2 with
Texture mapping.

Now you can add some surface displacement to the sofa by

painting a texture to represent the piping along the seams, edges or
cushions, or other details to add realism.

Drawing onto a Bump Texture

18. To draw details onto a bump map you need to a new image
textures, and click + NEW to create a new image. Call the
image “sofa_bump” and increase the resolution to 2048x2048,
again you don’t need the Alpha channel. Add a new bump node
to go between the sofa_bump texture node and the normal input
of the shader, with the texture setting the Height input. You will
probably need to fine tune the Distance and Strength of this
node once you’ve done some texture paining.

19. Ensure that the sofa_bump image texture node is selected then
change back to texture painting. If the sofa_bump texture was
not selected as the target you will draw onto the wrong image.
As I mentioned above, drawing using the brush stroke set to
Curve will help with smooth curves like the piping on the
seams. Also, switch to orthographic view using the button under
the camera view button, visible on the right of Figure 8-14. This
will prevent long lines drawn on the 3D object being distorted
by perspective and resulting in tapered lines on the 2D texture.

20. Don’t forget to click Save All Images as you work.

172
Conclusion

You will now see that there are endless possibilities. You could add
areas of wear or dirt by altering the roughness or you could mix in a
different texture, adding a print design with a different repeat spacing
from the basic fabric with yet another sample using a higher repeat as a
bump map for the weave, and so on and so on.

Being able to paint onto a 3D object by cloning from a perspective

image is also amazingly powerful, especially when you can map the
locations from the photo so easily to the object surface. Making quick,
medium poly, models, for use in the distance of a scene, or as game
assets becomes very fast indeed.

Behind both these techniques is a requirement to make a good

quality UV map and that takes a bit of practice. It also helps if you
have an instinct for where to make the minimum number of seams to
give the least distortion. I imagine tailors would be good at it!

You also did some more traditional modeling of a mesh by adding

edge cuts to control the shape of a subdivision surface. Despite the
speed and efficiency of using boolean and bevel modifiers, often you
need to get stuck in and change the mesh. As with many things in
Blender, if you find you are doing a lot of modeling this way there are
add-ons that can help you. However, having the skills to do the work
manually is a prerequisite to finding the best tool for your needs.

In the next chapter you will learn two different approaches to

baking normal textures from detailed meshes.

173
 9
NORMALS, BUMPS AND
DISPLACEMENT

Capturing surface detail as a Normal

Map

Normal maps are used throughout the world of 3D

modeling to render objects with the appearance of much
higher resolution than their meshes actually hold.
At several places in this book I mentioned normals, such as when
describing problems that could arise if faces had more than four sides,
or when explaining the difference between smooth shading and flat
shading. In this chapter I will flesh out some of the principles behind
reflections and shading on surfaces in computer generated images. I

174
 will then take you through two ways to generate a normal map from a
more detailed mesh and discuss some of the ways to combine these
normal maps with other detail such as painted or procedural textures.

Reflecting the Light

When I first introduced the idea of meshes as 3D objects I said that

they generally consisted of lists of vertex locations, along with data
defining the faces and edges in terms of those vertices. In the last
chapter you investigated some additional data stored in the model; the
texture coordinates of each vertex, but the other significant element is
the normal direction.

The normal direction is the way the surface faces at any point and
is used to calculate the way light bounces from the surface.
Superficially each vertex would have a normal so that every point
between them could be calculated by interpolation. However it quickly
becomes apparent that there must be multiple normal directions at each
vertex, one for each face corner, because otherwise it isn’t possible to
represent a change in direction from one side of an edge to the other,
which is needed for flat shading.

Normal maps extend this logic by providing a way to vary the

normal direction between vertices. For low and medium poly models
they add enormously to the realism, as you saw with the rust on the
refrigerator material in Chapter 3. Figure 9-1 shows a low poly floor
and wall without and with normal map applied.

Figure 9-1 Low poly shapes without and with normal maps.

175
Baking a Normal Map

The classic approach to making normal maps is to construct a very

high resolution model, often using sculpting to create fine detail, then
“bake” the detail to an image texture that can subsequently be applied
to a low poly version of the model. This scenario sounds just like
situation you have arrived at with the face you sculpted in detail in
Chapter 5, re-meshed in Chapter 7 and textured in Chapter 8, you have
a ready-made example to work with. Links to the files and videos for
this exercises in this chapter are at get-into-blender.com 9.NORMALS,
BUMPS AND DISPLACEMENT

Part of the difficulty when baking in Blender is in the terminology.

You may feel that you are baking from the detailed mesh to an image
texture on the low poly mesh. However you are required to make the
detailed mesh an active object by clicking on it first then make the low
poly object the selected object by clicking on it second, you must then
tick the check-box that specifies “Selected to Active”. That might
sound to be the wrong way, but it’s what you have to do.

1. Open the blend file from the end of Chapter 8; the head painted
using selfies. Make sure that the armature has all the pose
transforms reset and that, in the Object Properties tab, neither
armature or mesh have moved, rotated or scaled. To avoid
accidentally overwriting your previous work, save the file now
with a different name in the folder you will use for this chapter.

2. The baking process will look in each material of of the selected

object for a selected image texture node that can be overwritten
with a normal map. To set this up, change a window area to the
Shader Editor and for each material in the Person01 object
make sure that there is an image texture node disconnected from
the shader ready to have the normal map baked to it. The head
and body materials will have an image texture node already
created, such as “head_bump”, but you need to create one for
the eyes as well otherwise the baking process will look for
another texture, find the one you painted for the base color, and
overwrite that with a generated normal map. Which you don’t
want. Each material should have just one unconnected image
texture node, so check around to make sure you didn’t leave one
lying around when you were experimenting!

3. In each material, make sure that each of the unconnected image

texture nodes for the normal maps has the Color Space drop-

176
 down set to Non Color and click on each node so it is selected.
The selected node is indicated by a white outline.

4. Now bring in the detailed mesh you will bake from by

appending the sculpted mesh from the blend file saved at the
end of Chapter 5. It will arrive with all its dependencies such as
materials and an armature but you don’t need these. Delete the
armature, which has probably been called Armature.001. There
should be two materials skin and eyes.001, which will probably
also have been renamed to avoid clashes with existing
materials. Delete the eyes.001 material then go through the skin
material setting the base color to plain white and deleting any
texture nodes. This is mainly so you can differentiate the two
models as they will both be in the same location.

5. Baking can only be done using the cycles ray tracing engine so,
in the Render Properties tab set the Render Engine to Cycles
then, lower down, expand the Bake section and set the Bake
Type to Normal and tick the Selected to Active check-box.
Expand the Selected to Active section and set the Extrusion
distance to 0.1. The extrusion distance is the amount to expand
the selected surface that you are projecting onto, in this case the
Person01 object. The extrusion distance needs to be more than
the size of any bumps on the high resolution surface.

6. Select first the sculpted object then the Person01 object, the
order is crucial here, then click Bake in the Render Properties.
So long as no error message pops up this should work, but it
will take a while as images are baked for each of the materials.

7. When the baking has ended switch the Shader Editor area to
Image Editor and look at each of the three image textures set to
capture the normal map. They should all have produce a
generally light blue result with parts tending to purple-red and
parts tending to yellow-green. The results for the body and the
eyes can be ignored as we haven’t really set the system up for
those materials. If your baking has resulted in parts of the
head_bump image having significant patches of orange then
that probably means that the Extrusion distance wasn’t enough.
This is a tedious process on a low powered computer but you
just have to incrementally increase the distance until the amount
of orange drops. I explain why this happens in the next section.

8. To see the results of your normal map on the material of the 3D

object, switch the Viewport Shading to Material Preview. Hide

177
 the sculpted object in the Outliner Editor and change the Image
Editor back to the Shader Editor then, with the head material
selected, use Add4Vector4Normal Map to create a new
node, and connect the the output of the head_bump image
texture to the input of the normal map node, and the output of
the normal map node to the normal input of the shader.

9. It might be hard to see much difference at all but you can

increase the Strength of the normal map node to values greater
than 1.0 to emphasize the effect. Just to verify that it’s doing
what you expect, do some additional sculpting of the detailed
mesh. Switch the visibility of the sculpted object back on in the
Outliner Editor, select it and switch to Sculpt mode. Add some
strong grooves and ridges on the sculpted face then change back
to Object mode, select the Person01 object as well, disconnect
the normal map node from the shader and run the bake again.
When you reconnect the normal map node to the shader you
should be able to see the changes to the surface. It will probably
show better with Viewport Shading as Render, especially if you
set up unflattering lighting striking the object tangentially.

10. Remember to save your image files, either by packing them,

saving them individually or using the button in Texture Paint
mode!
The sequence of operations required to bake a normal map from
one object to another is, in my opinion, very non-intuitive and rather
precarious. If any of the steps are missed or done in the wrong order
then the baking will fail. Sometimes, even if everything is correct, the
baking will fail, in which case the only solution seems to be to save the
blend file, close Blender then start it up again.

Why are Normal Maps Blue?

At this point I should point out that the texture image names used
in the materials, such as head_bump are misleading for this exercise.
Bump maps vary from black to white to represent the amount that the
surface detail varies from the flat plane of the mesh. Where you drew
on details such as the piping on the sofa in the last chapter that was
appropriate. However normal maps are different shades of blue because
they represent something else.

The direction that the surface faces can be represented by an arrow,

or vector with three components X, Y and Z where X is the amount the

178
 arrow points from side to side, Y is up and down and Z is outwards.
These three components can be stored in an image file by allocating X
to the red value of the pixel, Y to the green value and Z to the blue
value. The range of values available for each color on a standard image
file is 0 to 255, so a red value of 0 is interpreted as facing completely to
the left and 255 as facing completely to the right, and similarly with the
green value ranging from straight down to straight up.

The blue value is always adjusted to keep the length of the arrow
the same, with 0 interpreted as straight inwards and 255 as straight
outwards. However, because it only makes sense for the surface to be
pointing outwards, rather than inwards, the values of blue component
should always vary from 128 to 255. If there is an error in the
calculation of the normal directions, such as where a bump on the high
resolution mesh is sticking through the low resolution mesh onto which
it is being projected, then the blue component switches to a value lower
than 128 and the normal map texture has patches of orange. This is a
frequent problem when baking normal map textures and the solution is
usually to increase the separation between the low and high resolution
meshes by increasing the Extrusion distance.

Multiresolution, Another way of Sculpting

There is an alternative method of baking a normal map which

relies on a modifier that you haven’t encountered yet, the
Multiresolution modifier. This modifier has some similarities to
subdivision however it allows the high resolution mesh to be sculpted
and the details then baked to a normal map for use with a lower
resolution mesh.

It may seem illogical at first, but a common work flow is to start

with a basic shape, sculpt it to a good approximation of the final model,
form a new mesh using re-topology over the top of the sculpted mesh,
then finally, use a multiresolution mesh to sculpt all the fine details for
baking to a normal map. Although there appears to be duplication, in
that the sculpting is done twice, in fact the first sculpting concentrates
on larger scale features and can be fairly “broad brush”. The second
sculpting carries on where the first finished, adding more subtle details,
and so there isn’t necessarily any duplication of work.

Lets try using multiresolution to bake a normal map for the

Person01 mesh you created in Chapter 8.

179
1. Save the blend file you worked on earlier, in the folder for this
chapter. Don’t overwrite the version you opened at the
beginning of the exercise because you now need to open that
again.

2. Select the Person01 object and add a Multiresolution modifier

and click Subdivide three times. Each time you press this
button it increases the number of faces exponentially so you
need to be very careful. My model starts at 3k then goes to 11k,
40k, 158k, 627k, 2,507k with each subdivision, so unless you
have a reasonably powerful computer it is easy to crash Blender.

3. Notice that there are subdivision levels shown for Viewport,

Sculpt and Render. This is how the modifier can be used to
generate normal maps; when you bake using bake from multires
the process will make a normal map to represent the difference
between the surface at the level of subdivision in the Sculpt
setting, compared with the level of subdivision in the viewport
setting.

4. Change the 3D Viewport to Sculpt mode and add a few ridges

and creases.

5. Switch the area on the left to a Shader Editor and ensure each
material has a non-color image texture node, such as the
head_bump texture you created earlier, and that it is the selected
node. Baking from multires doesn’t have a problem if the image
textures are connected to the shader but I often disconnect them
before baking, just in case!

6. In the multiresolution modifier turn down the subdivision levels

to 0 for Viewport and Render then in the Render Properties tab
set the Render Engine to Cycles, expand the Bake section, tick
the Bake from Multires check-box and make sure the Bake Type
is set to Normals. Click Bake.

7. Change to Image Editor and check that the head_bump image

texture looks OK then, back in the Shader Editor connect the
output of the head_bump image texture node to a normal map
node and then to the normal input of the shader.

8. Usually, with a fairly coarse mesh such as Person01, a

subdivision surface modifier would be added to the object. In
which case it might make sense to leave the multiresolution
modifier in place after baking the normal map, with the level set
to 1, or 1 in the viewport and 2 in the render. The level should

180
 only be reduce to 1 when baking and, to prevent the blend file
being very large, you should click Delete Higher with the
viewport level set to the highest you want to keep. This will
permanently lose the finer detail from the sculpting so make
sure you have baked the normal map before doing this.

Combining Bump Maps with Normal Maps

So what if you want to combine a bump texture with normal

texture? This can be done but the result is a slight compromise because
of the range that normals operate over, usually it works OK though. As
an example you could add some pores to the skin using a noise texture.
A more flexible way to do this would involve painting the bump map
with the desired texture as a brush, but using a procedural texture will
quickly demonstrate the node setup required.

1. The first thing needed is a series of nodes that will give the right
kind of surface when fed into the normal input of the shader. In
the Shader Editor click Add4Input4UV Map and position
the node to the left. To scale the output of the UV map select
Add4Vector4Mapping then set X and Y Scale to 300.0. The
mapping output feeds in turn to the texture node created using
Add4Texture4Noise Texture with Normalize un-checked,
then add a bump node by selecting Add4Vector4Bump with
Strength set to 0.6 and feed the output into the shader normal
input, replacing the input from the head_bump image texture.
Adjust the settings to give the right level of detail.

2. To combine the two normals you use vector addition. Reduce

the Strength of the normal map node to 0.6 as well then use
Add4Convert4Vector Math to add a node to combine the
two normals. However this result then needs to be normalized
by another Vector Math node with the operation set to
Normalize. The node arrangement is shown in Figure 9-2.

181
Figure 9-2 Nodes Combining Normal Map and Bump

The render on the left of Figure 9-3 shows a material with the
normal disconnected, but with a multiresolution modifier still in place,
set to level 1. The middle render has the normal image texture baked
from sculpted multiresolution level 3. The right hand render has both
the image and the noise texture pores using the shader shown in Figure
9-2.

Figure 9-3 Render showing the addition of baked normal and bump map

Applying a normal map to a material shader works really well if

the surface of the object is reasonably square to the camera, however,
when viewed in profile it is clear that the bumps don’t actually
protrude. Generally speaking this doesn’t cause a serious problem
because the normal maps are confined to small surface details, but for
gross features that need to be seen from the side, an alternative solution
is required.

182
Displacement

An obvious solution would be to model the alternative surface as a

finer mesh using a multiresolution modifier sculpted to provide the
required bumps. This would work fine for many situations but where
finer control is required, for instance if the strength of the effect needs
to be animated, or the bumps are required to move over the surface,
then a good solution is to texture paint a bump map and apply that to a
displacement modifier.

In the following exercise you will draw some scales as a bump

map onto the face of the Person01 object and use a displacement
modifier to show them. In order for there to be enough vertices for the
displacement to act on you will need to add a subdivision modifier, so
this will also re-introduce the importance of the order of modifiers in
the stack. In Chapter 11 you will learn about physics and particles and
it will become apparent why these functions also have an entry on the
modifier stack. Subdivision obviously needs to happen before
displacing bumps but you could apply a surface deform modifier to
follow a soft body simulation, say, after the displacement to allow the
bumps to droop, or before the displacement to get a “crustier” effect.

1. Make sure you have saved your last work, again using a
different name from the one you opened, then load the blend
file from Chapter 8 once more.

2. First add a subdivision modifier by selecting the Person01

object then in the Modifier Properties tab adding a subdivision
modifier with 2 subdivisions. You might need more but you can
come back to increase this when you see the results. Too many
subdivisions will slow the computer down while you work.

3. Now add a displace modifier, click on the + New to create a

new texture and rename it texture_bump. This texture is an
internal reference that then needs to be linked to an actual type
such as noise or image. Open the Texture Properties tab which
will show the settings for the texture you just created and
default Type will be Image or Movie which is what you need.
Expand the Image section and set the linked image as the
head_bump image texture you created in Chapter 8.

4. Back in the Modifier Properties tab adjust the settings to the

desired levels. By default the modifier expects a texture where
mid gray has no displacement, black gives extreme indentation

183
 and white extreme expansion. Turn the Strength down to 0.05
and the Midlevel down to 0.0 you can come back and fine tune
these later.

5. To allow texture painting scales on the 3D model you will need

to use the UV unwrapping you set up in Chapter 8. In the
modifier properties change the Texture Coordinates to UV and
select the main UVMap.

6. Now draw some scales onto the face. Select the Texture Paint
tab from the workspace default configurations, at the top of the
screen then, in the Active Tool and Workspace settings tab
change Mode to Single Image with the head_bump as the
linked image and the UVMap as the specified map.

7. When you draw on the face you might find, as with painting the
particle density in Chapter 5, that the effect isn’t actually visible
until you switch from Texture Paint mode to Edit mode or
Object mode, then back again. For this reason it’s probably best
to draw with viewport shading as Solid and the Color option in
the drop-down as Texture, so you can see where you’ve just
painted. To make the scales tilt, I used a texture for the brush
with Type Blend and set the brush Falloff to Constant. To get a
graduated circle with a single click you will need to set the
Brush Settings, Texture, Mapping to View Plane and to control
the size and direction of the scales set Stroke, Stroke Method to
Anchored.

8. Because the eyes and body are also influenced by any modifiers
added to the Person01 object you will find that they also
become scaly in a haphazard way. You can fix this by confining
the modifier to a vertex group. With Person01 still selected
switch to Edit mode, deselect everything, then in the Material
Properties tab click on the head material then use the Select
button to select just the faces on the head, but excluding the
eyes. Open the Object Data Properties tab and, under the Vertex
Groups section create a new group with + New, change its name
to “scaly_skin” and click the Assign button. In the displacement
modifier setting set the Vertex Group optional field to
scaly_skin.
You should be able to create quite large scales using a
displacement modifier in this way. One additional advantage over
simply sculpting the mesh, is that the texture can be used in the
material to modify the appearance. In Figure 9-4 you can see that the
higher the displacement value, the more reptilian the coloring.

184
Figure 9-4 Increasing scaliness

You will remember how to use a factor to blend two shaders from
Chapter 4, but wouldn’t it be really cool to have one “slider” you could
adjust, or animate, that would alter both the strength of the
displacement modifier and the factor for blending the shaders in the
material. Well this can be done using custom properties and drivers,
both very powerful, but less frequently used features of Blender.

9. First add the nodes to create the reptilian effect on the tops of
the scales. You might like to try figuring this out for yourself,
referring back to the rusty refrigerator material to refresh
yourself.
My node layout is shown in Figure 9-5, if you get stuck. There you
can see that I use a noise texture with a color ramp for the scale color
and and two voronoi textures for the normal. A quick search online will
provide suggestions for excellent lizard skins.

The Color output from the head_bump texture is the factor for
blending the shaders, but it is modified by passing through a multiply
node which has its text input box colored purple. The purple indicates
that the value is being changed by a driver and cannot be altered
manually.

185
Figure 9-5 Scaliness controlling the blend of two shaders

10. Add a math multiply node and use it to multiply the output of
the head_bump image texture before feeding into the Fac input
of the mix shader.

11. With the Person01 object still selected, open the Object
Properties tab, scroll down to the bottom and expand the
Custom Properties section. Click Add and then Edit and change
the name of the new custom property to “scaliness”. The other
defaults are probably fine but you can see it is possible to set
max, min, default and tooltip. Click OK to save the edit.

12. Right click on the value of the scaliness custom property and
select Copy as New Driver, then right click on the value field of
the math multiply node you just added to the material and select
Paste Driver. Open the Modifier Properties tab then right click
and paste the driver into the value field of the displacement
modifier Strength field as well.

13. The chances are that the displacement strength and material
adjustment will need different scales. I’m sure you can already
see how you could adjust the value using an additional node in
the material but it is also possible to perform quite complicated
calculations as part of the driver. Right click on the purple
Strength field and, in the menu options, select Edit Driver and
change the Type to Scripted Expression. In the Expression text
box below change the entry from scaliness to scaliness * 0.05.
Change the multiplication value in the material node in a similar

186
 way to give the desired effect as you move the slider, I
multiplied the scaliness by 4.0 in my example. There is also a
graphical method of achieving the same effect using the Show
in Drivers Editor button at the bottom of the pop-up window.
There isn’t room in this book to explore this in real detail.
In some circumstances drivers are extremely useful, they can link
almost any value in an object to any other value in the same or any
other object. In the exercise you have just completed, you could
actually have made the value in the material node drive the modifier
strength field directly, or visa versa, and skipped the custom property
completely. However it is much better to keep clever mechanisms like
this as clear as possible with all the components visible and explicit and
nothing hidden away. Even when you are comfortable that you know
exactly how everything works, in only a few weeks time you will
struggle to remember where the value with the purple background is
coming from. When that happens you must right click and edit the
driver to rediscover the source.

Conclusion

In this chapter you used two different ways to bake a normal map
from a high resolution mesh to a coarser one. Both techniques have
their applications and which one you opt to use will depend on both the
workflow you are most comfortable with, the resources available for
your modeling and the final destination of the finished product. If you
are making a low poly asset for use in a game engine you will probably
take a different route from the one you would follow for an animation
rendered in Blender.

The displacement modifier can be a very useful tool, providing a

shortcut alternative to sculpting repetitive detail such as tree bark, rock
surfaces or lizard skin. You also used a subdivision modifier to generate
the vertices needed for the displacement modifier to act on, and this
method of stacking physics and particle effects after subdivision is
something you will encounter often in “how to” explanations online.

In the next chapter you will delve further into volume shaders,
using them to investigate four dimensional textures, monsters of the
deep and glowing writing.

187
 10
F O G A N D S E AW E E D

Adding texture and light to volume

shaders

Volumes can be used to add much more to the scene than

simple mist.
In Chapter 2, Chapter 3 and Chapter 4 you added volume shaders
to different shapes or the world properties in order to give the
impression of atmosphere and enhance the final render. However it is
possible to vary both the density and the color across the volume using
textures which enables some really interesting effects.

188
 Volume Shading Under Water

For this exercise you will construct a fairly simple underwater

scene comprising a model of a shark, a cube with a volume shader for
seaweed and murk and a cube with the top removed for the background
sand. It would be possible to model the water refracted through the
surface of the sea using a glass material, but the rendering would be
very slow so instead, you will create dappled light and shade using a
plane with variable transparency. Links to the files and videos for the
exercises in this chapter are at get-into-blender.com 10.FOG AND
SEAWEED

A Low Poly Mesh Based on a Skin Modifier

Modeling the shark will provide you with more practice at the core
Blender skill of building a low poly mesh to create the desired shape as
a subsurface. You could quite simply extrude and scale different faces
of the start cube as you did for the sofa in Chapter 8, however the
challenge here will be to start with an extruded edge and skin modifier
as you used to make the manikin in Chapter 2.

1. First delete the start cube and add a plane, then in edit mode
delete two of the vertices then move the remaining two onto the
Y axis by setting snap to absolute grid. Extrude each end
forward and backwards then outwards for the tail and fins. Add
a mirror and a skin modifier then, finally, use CTRL-A to adjust
the size of the skin at each vertex. The left of Figure 10-1 shows
the rough shape I created.

Figure 10-1 Shaping the shark

2. When you have the proportions about right you need to apply
the mirror and then the skin modifiers. This is unfortunate as
you really want to keep the mirror modifier while you work on
the mesh but there isn’t a way to do that while applying the skin

189
 The reason why the purple scatter produces a green hue is related
to the reason that sun-sets are red. The atmosphere scatters blue light
much more than red, which is why it looks blue during the day. In the
evening, when the sun is low and there is enough atmosphere in the
way that you can look towards it, so much blue light is scattered to the
side that the majority of the light reaching your eyes is red.

Scatter v. Absorption v. Principled

There are two other volume shaders available in Blender and now would
be a good opportunity to see how they differ from each other. The first,
logical, alternative to the volume scatter would be absorption. A
reasonable assumption would be that this shader specifies a color of light
to be preferentially absorbed by the volume material, so allowing the
complementary color to be transmitted, a little like the scatter shader.

21. Check out how the absorption shader works by swapping it for
the scatter shader in your material. Select
Add4Shader4Volume Absorption and connect it into the
place of the existing shader. Set the Color HSV to 0.772, 1.0,
1.0 which produced a good green color in the scatter shader.
You will probably need to the the Value field of the Multiply
node down to 5.0 to allow any light to reach the camera.

22. It will be apparent that the Color value in this shader specifies
the color that isn’t absorbed by the volume. Change the H of the
color to 0.25 then reduce V to 0.9. This should produce a
reasonable green seaweed effect which is rather dull compared
with the scatter shader but significantly faster for the cycles
render engine to calculate. One deficiency of this shader is that
beams of light passing through the volume are not visible at all.
The third shader is the volume version of the default principled
shader. Although this doesn’t scatter light in quite the same way as the
scatter shader, it can represent the passage of a beam of light through
the volume as well as having setting for absorption, emission and
black-body radiation. It is a very flexible shader and can be used to
model most scenarios, however it is the slowest of the three.

23. In the Shader Editor area add a third shader by selecting

Add4Shader4Volume Principled BSDF and connect it in
place of the absorption shader. Change the Color HSV to 0.32,

194
 0.9, 0.56, the Absorptions Color HSV to 0.5, 0.4, 0.75, the
Emission Color HSV to 0.6, 1.0, 1.0, the Emission Strength to
0.001 and Anisotropy to 0.4. Again you will need the Value of
the Multiply node to be reduced to 5.0 to compensate for the
absorption.
You will probably have to adjust all the settings to get a good
result as the amount of attenuation of the light will need to match the
size of the scene you made and how far away you position the camera.

In my render in Figure 10-4 you will see some shafts of sunlight

penetrating the waves above the shark. Although the sun shining
through a screen will produce this effect it is rather subtle and difficult
to set up while keeping the rest of the scene in balance. To achieve this
effect I cheated and added three spotlights just out of view of the
camera.

Figure 10-4 Final Render of Shark with Seaweed and Sunbeams

24. First add one spotlight and set it up, you can duplicate it once
you’ve got it to your liking. Select Add4Light4Spot and
rename it beam1, then position it just outside the field of view
and rotate it so it’s shining near to the shark and from the side.

25. Now increase the power and narrow the cone by opening the
Object Data Properties tab. Set Power to 2000 and, in the Spot
Shape section, set Size to 10.

26. It’s easier to position the spotlight with Viewport Shading set to
Material Preview but you won’t be able to see the result until
you view in Rendered. If you look closely at my render you will
see that I used the cycles render engine without denoising, as

195
 the slightly grainy effect looked more watery than the smoothed
or eevee version. To rotate the beam, use R to rotate
perpendicular to the view plane, and R, R to rotate using
trackball control.

27. Once the sunbeam is leaving a trail in the water, duplicate it

twice and move and rotate the copies to give the classic
spreading pattern. One problem with using spotlights is that
they leave a bright circular patch on the sandy bottom or the
shark. It is possible to control which object is illuminated by
which light, but it’s something to be tackled later, when you
learn about the compositor in Chapter 13. For now aim the
spotlights so they hit the sand somewhere out of sight, either
behind the shark or off the bottom of the frame.

Anisotropy

When you were adjusting the settings of the volume scatter I

suggested that you experiment with the Anisotropy value and probably
found that it had a visible effect at either end of the scale, but seemed to
do little in between. However you should find that this setting does
change the appearance of the sunbeams. This is because it controls how
much of the scattering is back towards the light source and how much
is away from it. If the light is scattered mainly back then little will
penetrate into the volume, if it is scattered almost entirely forward then
the beam will travel further but less light will be scattered towards the
camera.

Anisotropy is the reason that you can often see beams when you
look towards the sun when it is hidden behind a cloud but you rarely
see them when the sun is behind you. Following this logic you may
achieve better sunbeams by shining the spotlights towards the camera
and setting the anisotropy to scatter forwards. Some trial and error will
be required to get the best result.

3D Textures

Throughout this book, whenever you added a procedural texture

node such as noise or voronoi to a material, you may have wondered
why the Dimensions field defaults to 3D. Now that you have used the
texture to fill a volume with seaweed you can see how the extra
dimension could be useful, but is there an application for the extra

196
dimensions when using the noise texture to pattern the shark’s skin or
create the dappled sunlight effect? It turns out that this is a very
convenient way to represent time and can be used to animate textures.

Keyframes
You will learn about animation in Chapter 12 but, as there is so
much to cover, the following exercise is a good opportunity to
introduce one of the most important aspects of animation; keyframes.
Keyframes are a feature of Blender that pervades virtually everything
and, in an astonishingly simple way, allows you to breathe life into
your static scenes. Essentially by holding the mouse cursor over any
variable field and pressing I on the keyboard you can specify a value
for the field to take at the current frame, where frame is a proxy for
time. Blender is clever enough to understand that if you press I when
certain bones of an armature are selected, and the cursor is over the 3D
Viewport in Pose mode, that keyframes should be set for the relevant
bones. You can animate object locations, camera properties, modifier
fields and material settings. There is scarcely anything in Blender that
cannot be animated. For this exercise you will animate the material
used by the screen to cast shadows on the sea bed.

1. Select the screen plane and open a window area as a Shader

Editor. You should move the view point up so you are looking
down onto the surface of the screen. In order for the animation
to happen at a reasonable speed you should switch the Viewport
Shading to Material Preview and make sure that the screen
material Blend Mode in the Settings section is set to Alpha
Clip.

2. In order to change the location of the texture in its third

dimension, to represent time, you need to feed information into
its Vector input. First select Add4Input4UV Map then create
another node with Add4Vector4Mapping and take the UV
map output via the mapping node to the musgrage texture input.
The node layout is visible on the bottom left of Figure 10-5.

197
Figure 10-5 Node Layout and Timeline for Animated Shadows

3. Now you are in a position to set a keyframe, but before you do,
you should check that the current time is set to frame 1. Swap
the editor area to the Timeline and ensure that the vertical blue
line is in the right place. At this point the only reason it might
have moved is if you accidentally pressed the spacebar which
will start the animation playing. You can move the play position
to the start point by pressing the Jump to Endpoint button on
the play controls.

4. Switch back to Shader Editor and, with the mouse over the
mapping node Location X, Y or Z field press I. It doesn’t matter
which of the three location components is under the cursor as all
three will be set at this keyframe. All the location values should
change to a yellow color.

5. Switch to the Timeline where you should see that an orange

diamond has appeared to show the location of a keyframe. If
you can’t see one it’s probably because you haven’t selected the
component that is being animated, in this case the mapping
node, which must be selected and showing a white outline in the
Shader Editor. Move the play head to the extreme right end
point, which will default to frame 250. Change back to the
Shader Editor and where the Location values will still all at 0.0
but they now show green. This coloring indicates a field that is
animated but at the specific frame it is being viewed, the value
has been interpolated by Blender.

6. Change the Location Z to 1.0 and you will see that the field
turns orange to indicate a field that has been changed from an

198
 interpolated value but not yet fixed as a keyframe, it will be
forgotten as soon as you change frame. With the mouse over the
Location fields press I, and you will see them all change to
yellow to indicate that they have been saved as keyframe
values.

7. To see the animated values press LEFT a few times. The

Location fields will be green and the Z should gradually be
changing from 1.0 down to 0.0. Press spacebar to make the
animation play and the water surface should move in quite a
realistic way. Press spacebar again to stop it playing.

8. The first issue you might notice is that the speed of animation is
much slower just after and just before the keyframes. This is
because the default interpolation uses easing to avoid jerkiness
and sharp discontinuities of movement. To fix this, go back to
the Timeline and select both of the keyframes then, from the
RMB menu select Interpolation Mode4Linear.

NOTE In the Timeline area the selection and manipulation of keyframes

has many similarities with vertices in a mesh. Groups of keyframes
can be selected using box selection or circle selection with SHIFT
and CTRL having the same effects. It is possible to move or scale
keyframes using G or S as well as duplicate and delete using the
same shortcut keys as in Edit mode.

9. The second improvement you can make is to give a slight

sideways movement to the ripples by increasing the X Location
to 0.5 on frame 250. The field will turn orange after you make
the change, warning you that the keyframe hasn’t been updated
with the new value until you press I, whereupon it will turn
yellow.

10. Play the animation to verify that it does what you expect.

Using 4D Textures to Animate Volumes

Just as a 3D texture can be used to animate a 2D surface, a 4D
texture can animate a volume. You will now try setting up a little bit of
sway to the seaweed in front of the shark. Your objective will be to add
three keyframes for time in the fourth dimension as well as three
keyframes in the X and Y directions. There will be no movement in the
Z, or up, directions but, hopefully the overall effect will be of gentle
back and forth movement. If you make the time period of each repeat

199
 different then when you copy and paste the keyframes they should
produce a more random looking effect. Figure 10-6 shows the Timeline
with one repeat of keyframes selected.

Figure 10-6 Node Setup to Animate a Volume Texture

1. First of all, on the Timeline move the play head back to frame 1

2. Select the seaweed object and switch a window area to the

Shader Editor. You have already set the noise texture node to
take the position vector from the geometry node via a mapping
node but that only gives you three dimensions, so where does
the fourth one come from in order to change the texture with
time? Click on the Dimension field in the noise texture node
and select 4D.

3. Now you can see that the fourth dimension is available in the
texture node as a field W. With the mouse over the field add a
keyframe by pressing I, then move to frame 48 on the Timeline
and, back in the Shader Editor change the W value to 0.01 and
fix the keyframe. The default animation speed is 24 frames per
second so you are aiming to get a slow change to go with the
physical swaying movement. On the Timeline move the frame
to 96 then add a keyframe with W set to 0.03.

4. In the Timeline Editor you should be able to see the three

keyframes you created. If you can’t, remember you need to
select the noise texture node in the Shader Editor. On the left of
the Timeline is a collapsible hierarchy of the components that
are selected and have keyframes allocated. Expand the
hierarchy so you can see which line of keyframes corresponds
to each animated value, then select the three keyframes you
added to the W value and copy them with CTRL-C. Move the

200
 current frame marker to around 120 then paste with CTRL-V.
Select just the keyframe on frame 1 then copy and paste it to
frame 250 to make the animation loop smoothly.

5. Move the current frame back to 1 then, in the Shader Editor

select the vector mapping node that is feeding into the noise
texture node. Add a keyframe for the Position fields with them
all at 0.0 then add two more keyframes at 30 and 60 with X and
Y positions set to arbitrary values between -0.04 and 0.04.

6. On the Timeline you can drag individual keyframes backwards

or forwards in time. Use this to change the positions of the X
and Y key frames slightly then copy and paste the two variables
separately so they are out of phase. If you look carefully at
Figure 10-6 you will see how the keyframes are arranged to
make the animation appear more random.
When you run the animation for this scene it will probably be
rather hard to get an impression of what it would look like running at
24 frames per second. Unless you are using a powerful computer the
time to calculate each image, even using the eevee render engine, is
quite long. This is a problem with transparent objects such as volumes
and also with particle systems or hair which are notoriously slow to
render. In Chapter 12 you will learn how to render animations as a
series of images and how to then convert those images into a video
using the Video Sequencer. For the time being save your project so you
can render the animation later.

Manipulating Procedural Textures

In Chapter 13 you will look in detail at ways to make Blender

produce non-photorealistic renders, and you will use both material
nodes and the compositor to apply hatching to a scene. Figure 10-7 is
in the style of a 17th Century etching where a vision has been conjured
as a source of light. However, for the following exercise, you will
concentrate on one part of that scene; the apparition appearing above
the alchemist’s table. Your objective is to create a volume in which the
density varies with a random 4D texture that has been stretched radially
and, in addition, the density is high in the center, falling off towards the
edges. As the light shines through the volume it is scattered brightly
from the pattern in the volume give the illusion of rays spreading from
a central point.

201
Figure 10-7The Alchemist at Home in his Study

1. Start a new Blender scene but leave the default cube in place
and rename it to rays. Scale it in the Y direction by a factor of
0.1 then, in the Material Properties tab, in Surface select
Remove and in Volume select Principled Volume.

2. Give the volume a textured density as you did in the previous

exercises by adding a noise texture node and feeding its output
through a color ramp and then a multiplication node into the
shader’s density input. Even with a high multiplication factor
for the density it will be hard to see the pattern so, in the World
Properties tab change the background Color to black. Move the
camera so it’s looking at the rays object square, approximately
along the Y axis, then change the light to a spotlight and move
and rotate it so it’s shining onto the back of the rays object and
towards the camera.

3. In order to create a pattern of spokes you need to calculate the

angle and the distance of each point in the volume relative the
center. For the angle you will use the arctan2 function and for
the distance you will use the length function. In the Shader
Editor create a node to generate the location of each point in the
volume using select Add4Input4Geometry, then create a
node to split the X, Y and Z values with
Add4Converter4Separate XYZ, and finally a node to work

202
 out the angle with Add4Converter4Math. Change the type
of function for the math node to Arctan2 then connect the
Position output of the geometry node to the Vector input of the
separator, then the Z output to the first Value input of the
arctan2 node and the X output to the second Value input.
Connect the output from the arctan2 node to the Vector input of
the noise texture node. Your node setup and result should be
similar to Figure 10-8

Figure 10-8 First Stage of Creating a Radial Pattern

4. You have only done the first part of the exercise, calculating the
radial angle of each point, however there is a problem you
should sort out now. Try moving the rays object and you will
see that the center of the pattern is fixed to the global origin of
the scene. This is because the geometry position is in the global
frame of reference so, to get the position within the volume you
must subtract the location of the center of the object. To do this
use Add4Converter4Vector Math, and drop the node onto
the pipe between the existing geometry position and the
separate XYZ node, then change its mode to Subtract. Create a
node with a Location output by selecting Add4Input4Object
Info, and feed that into the other Vector input of the subtract
node. The node arrangement should look like the left part of
Figure 10-9 and now, if you move the rays object, you should
find that the radial pattern moves too.

203
Figure 10-9 Completed Material Nodes for Radial Rays

5. To make the density vary radially, as well as circumferentially

select Add4Converter4Vector Math, change its mode to
Length then connect the output of the subtraction node to its
input. You also need to feed additional coordinate values into
the noise texture node so use Add4Converter4Combine
XYZ to create a node and drop it onto the input of the noise
texture. Now feed the output of the length calculation into the Z
input of the combine node.

6. The pattern you obtain is probably deficient in two respects; the

spokes aren’t stretched enough and the density doesn’t taper to
the edges, leaving a sharp square outline. To solve the first issue
you can insert a multiplication factor into the Z lookup value
you just set up.

NOTE To improve your workflow you should get into the habit of
duplicating nodes using SHIFT-D rather than creating them with
the menu, especially the common ones like math, vector math and
color ramp.

To make the density high in the middle, reducing to zero at the

edges you could use some math nodes, or a color ramp node followed
by a math node, but in Figure 10-9 you can see that I have introduced a
map range node which can sometimes simplify the node tree. You may
also notice the other change I made; connecting the Y location into the
noise texture lookup. This has the effect of making the density vary

204
 with depth, away from the camera, and adds more complexity to the
pattern.

Conclusion

This chapter used volume shaders as the focus of two ideas that
have uses in other areas of Blender; First you saw how higher
dimensional textures can evolve through time and allow you to add life
and subtlety to many animated scenes. Next you used the modification
of UV maps to generate new patterns from procedural textures, and this
method also creates enormous potential for stunning procedural
textures. Before you move on to the next chapter try experimenting
with materials nodes using different procedural texture to scale or
rotate the UV map of another. Checkout some examples on
eldwick.org.uk/get_into_blender 9.NORMALS, BUMPS AND
DISPLACEMENT.

The occasions when you need to use a volume shader with varying
density might not be many, but when you do, the techniques covered
here can provide unique functionality.

You have seen a little bit of animation so far in the book. In

Chapter 6 when you imported the characters from mixamo.com and, a
little more proactively, in this chapter when you set keyframes for the
material textures. Chapter 12 takes this exploration much further,
including how to rig an armature to a model and ways to constrain
bones while they move. Before that, in the next chapter, you will learn
about the different physics and simulation systems built into Blender.

205
 11
E A R T H , WAT E R , A I R , F I R E A N D
HAIR

Using Blender physics and

simulation systems.

The software has several sophisticated methods available

for producing very realistic special effects and making
your animations behave naturally.
You have used particles already in this book to create the
distribution of grass on the hill and boulders on Mars. When you tried
the quick effects option in Chapter 1, I remarked that the options
sounded exciting but might be frustrating to try out before you had
learned more of Blender. Well that point has now arrived and I will

206
 guide you through the remaining quick effects; explosion, smoke and
liquid. Each one has its own quirks and there are different methods
required to get them to interact with other objects or to render
satisfactorily.

As you work through the interactions of physics, particles and

modifiers used by the quick effects, you will try applying some of the
force fields to them, before moving on to simulations using cloth, soft
bodies, rigid bodies and object constraints. There’s a lot to cover so lets
get started.

Explosion

Surprisingly, out of the remaining three quick effects, the

explosion is the simplest, and produces a visible, if rather
unimpressive, result straight out of the box. The menu option is
essentially a short cut to add two features to the selected object, the
first is a particle system and the second is an explode modifier which
chops the object up. You could create your own explosion very quickly
by applying the two components yourself, then reducing the number of
particles being emitted, setting the start and stop time to the current
frame and changing the render type from Halo to None. However the
critical step is to ensure that the particle reference in the modifier stack
is above the explode modifier. Links to the files and videos for the
exercises in this chapter are at get-into-blender.com 11.EARTH,
WATER, AIR, FIRE AND HAIR

1. Start a new Blender scene and add the explode quick effect to
the default cube. In Object mode, with the cube selected run
Object4Quick Effects4Quick Explode. Press spacebar to
run the animation, and again to stop it. Use SHIFT-LEFT to
return to the first frame, or alternatively, you can navigate using
the Timeline controls in the editor area at the bottom of the
screen. Note that, to see the unexploded cube, you need to move
to the “zeroth” frame by pressing the left arrow again.

2. There are several issues with the explosion; there’s not enough
shrapnel, it’s not moving fast enough, the pieces aren’t spinning
at all and there isn’t anything bouncing off walls or the floor.
The first thing to do is break the cube into smaller pieces and
the best way is to add a subdivision modifier to the stack. In
order for that to work you must move the subdivision above the

207
 explode modifier, set the type of subdivision to be Simple, and
the number of subdivisions in the viewport to 2.

3. To increase the particle speed and make them spin, open the
Particle Properties tab and under Velocity, increase Normal to
10, Tangent to 5 and Randomize to 5. Tick the Rotation check-
box then expand that section and set that Randomize to 1 and
tick the Dynamic check-box. Now when you run the animation
it should look more spectacular but you will find that at the start
of the animation the cube is already starting to explode. In the
Particle Properties tab change the Emission Frame Start and
End to 5. Also, to stop all the pieces disappearing after 50
frames increase Lifetime to 250.

4. To create some bouncing, add another cube and scale it up so

the camera and light are inside it. With the new cube still
selected open the Physics Properties tab and click Collision.
Make the collisions less elastic by increasing Stickiness,
Damping and Friction. The simulation will also be more
realistic if you edit the particle system on the exploding cube
and in the Physics section under Forces change the Drag and
Damp values to 0.02.

NOTE The physics collision used here is a mechanism to control the

interaction of objects in very specific circumstances; it applies to
particle systems, cloth and soft bodies. As you will see later, rigid
bodies collide with other rigid bodies and fluids need to have a
fluid added to the physics properties of any objects with which
they interact, with its type set to effector. An issue you might have
is that the simulation won’t reflect changes to the physics. This is a
side effect of the efficient caching that Blender uses, and you
sometimes need to trick it by changing a value or adding or
removing a keyframe!

5. Once you’ve got the exploding cube behaving as you want,

fine-tune it and experiment with some force fields. The pieces
of shrapnel would look better with some thickness which you
can achieve by adding a solidify modifier, adjust the Thickness
to suit the size of the chunks. To create a force field add a new
empty object then, in the Physics Properties tab select Force
Field. Leave the Type as Force but change the Strength to -15.0
to create a black hole effect. Try the other types, some will need
more setting up to work, but Wind, Vortex and Magnetic should
have dramatic effects, though you might have to rotate the

208
 empty or change the Strength value. As I mentioned in the last
chapter it’s possible to add keyframes to almost everything in
Blender so you could make the black hole start after the initial
explosion. Try setting the force strength to 0.0 and insert a
keyframe at frame 30 then add another keyframe at frame 50
changing the strength to -15.0.

Smoke

The next quick effect is smoke, and, although it might be tempting

to add this to the existing explosion, everything will be easier to follow
if you start a new scene. Smoke uses Blender’s fluid simulation system
which is very flexible but, unfortunately, requires you to set it up in a
different way from the particle system used by the explosion. The main
things to watch out for are; setting the physics fluid properties for
objects to interact with the fluid, and baking the simulation after you’ve
made changes.

1. Start a new scene, save your explosion so you can make a video
using it when you cover how to do that in Chapter 12. With the
start cube selected run Object4Quick Effects4Quick
Smoke. This should produce some reasonably realistic smoke
when you play the animation, even with the viewport set to the
default solid shading.
A new object has appeared named “Smoke Domain” which defines
the volume in which the fluid simulation will be calculated. When you
run the animation the smoke might appear to be deflected from the
edges of the domain as if by invisible walls and ceiling but that is an
effect caused by the simulation having a sharp cut-off at the edge of the
domain. The fluid domain is divided up into a number of resolution
divisions which defaults to 32, so if you make the domain larger it will
make the fluid simulation coarser. However if you increase the number
of divisions you will slow down the simulation and use much more
memory. So your objective should be to make the domain as small as
you can get away with and only increase the number of divisions once
you have run the simulation and can see, for instance, that the droplet
size is too large.

If you are using an older version of Blender, prior to 3.3, you will
run into a problem getting the simulation to re-run and bake in any
changes you have made. Run the simulation right through to the end
then stop it and return the play head to frame 1. With the Smoke

209
 Domain object selected open the Physics Properties tab then scroll
down and expand the Cache section. You will see that the Type is set to
Replay, which, in theory, will bake the cache as the animation is
playing; the first time through might be slow but subsequent runs will
be fast. The problem is that the fluid domain doesn’t always “know”
when it needs to re-bake the simulation. If you have an old version of
Blender you will see what this means by selecting and scaling the
domain object by a factor of 2. The cube of smoke should have grown
in proportion to the domain, and when you run the animation you will
simply see a scaled up version of the original animation. To force a re-
bake you need to change one of various settings in the domain, so in
the Cache section Type select Modular then re-select Replay. You
should see the box of smoke shrink back to match the start cube, and
when you play the animation the simulation will be re-calculated.

2. In order to deflect the smoke you need to add a mesh object and
define it as an Effector. Add a new cube object to the scene,
scale it up so it is the same width as the fluid domain then scale
it down in the Z direction so it’s about half the height of the
start cube. Move the new slab object up so it’s about a third of
the way between the start cube and the top of the domain. In
Figure 11-1 you can see the size and location I used.

Figure 11-1 Smoke Simulation Rendered with Cycles

3. To make the slab into an obstruction, open the Physics

Properties tab and click on Fluid, then select Effector as the
Type. Select the domain object and, if using an old version of
Blender, force a re-bake by changing the cache type from, and
back to, Replay. Make sure the play head is back at frame 1 then

210
 run the animation and you should see the smoke hit the new
baffle.

4. Now make a hole in the slab about a quarter of its overall width
and over to one side, again, you can see the proportions I used
in Figure 11-1 and Figure 11-2. The simplest way to make the
hole is to add a cylinder to the scene, reduce the Vertices to 8 in
the Adjust Last Operation panel then, when it is suitably scaled
and positioned, apply it as a boolean difference modifier to the
slab. In order to see the smoke flowing around the baffle add a
new material to it, change the shader to Glass BSDF and reduce
the IOR to 1.01. Duplicate the slab, move it up and rotate it 90
degrees about the Z axis. You may once again have to force a
re-bake of the fluid simulation to see the effect of these double
baffles.

5. We haven’t yet looked at the settings on the original cube which

is generating all the smoke. Select the cube and open the
Physics Properties tab where you can see that the quick smoke
function has enabled fluid physics and set the Type to Flow, the
Flow Type to Smoke and the Flow Behavior to Inflow. Most of
the settings here are self explanatory but the three Flow
Behavior options; Inflow, Outflow and Geometry need a little
more explanation. Inflow and Outflow are continuous processes
that either flow into the simulation, as a source of fluid, or out
of the scene as a drain. Geometry is a one-off inflow of fluid in
the shape of the mesh so this would be suitable for an object
disappearing in a puff of smoke, or to simulate the spillage of a
cup of coffee. Try changing the Flow Behavior now to
Geometry then trigger a re-bake of the simulation, if needed,
and run the animation.

6. Finally, see what happens when you tick the Border Collisions
check-boxes in the domain object and try adding an empty to
the scene with a force field set to Wind, Vortex or Magnetic

Liquid

At this point, as an alternative to using the quick effect to make the

default start cube into a liquid, it might be more instructive to try
converting your existing smoke simulation.

1. First of all save the blend file so you can use it to make a video
later, then delete the empty with the force field as this will be an

211
 unnecessary confusion to start with. It would be a good idea to
save the blend file now with a different name so you don’t
accidentally overwrite the smoke version at the end of this
exercise.

2. Because a liquid is going to flow downwards you need to move

the start cube, with type flow, up to just below the top surface of
the domain. Move the two baffles down into the lower half of
the domain below the cube.

3. The start cube still needs to have Type set to Flow but the Flow
Type should be changed to Liquid. Keep the Flow Behavior as
Geometry for the moment.

4. The domain also needs to be changed from Domain Type Gas to

Liquid. This might make it appear solid again in the viewport
display, in which case open the Object Properties tab and under
Viewport Display change Display As to Wire. Move the
Timeline to frame 1, force the simulation to re-bake if required,
and you should see a box-full of generally blue particles flow
out over each baffle, draining through the holes and gradually
flooding the bottom of the domain. The left of Figure 11-2
shows the simulation part way through.

Figure 11-2 Stages to Achieve Liquid Simulation

5. To make the liquid actually look like a liquid you need to

specify it as a mesh by ticking the check-box in the domain
object Physics Properties tab, and you also need to go back to
frame 1 and re-run the animation. The middle of Figure 11-2
shows the same frame but with the liquid as a mesh. At this
point you may decide, as I did, that the droplets are too coarse,
so increase the Resolution Divisions to 64. Changing this value
will trigger a re-bake of the simulation so you don’t need to do
anything apart from return the frame pointer back to frame 1.
Notice how much slower the first run through the animation is.

212
 6. Once the animation has run and the cache has been re-baked,
you need to do a few more things to make the liquid look like
water. First, in the Object Properties tab for the domain change
the Display As setting back to Solid, it only needed to be wire
to allow you to see the blue particle simulation stage. Next open
the Material Properties tab and change the surface shader to
Glass BSDF, giving it an IOR of 1.33. When you view this as a
render preview, with the cycles render engine, it should look a
little bit more watery but rather lacking in reflections and
somewhat turbid. That’s because the quick smoke effect added a
volume shader to the material, so remove that. You should also
set the object shading to smooth and, to provide some more
interesting reflections, change the world background color to an
environment texture. The right side of Figure 11-2 shows the
final render.
These two fluid simulations give you an idea of how to start
setting up your own projects, and the kinds of problems you might
encounter, however they barely scratch the surface of the potential
available. Viscosity, surface tension, foam, spray and bubbles are
possible and allow all manner of effects from molten metal to tidal
waves.

In the next sections you will move onto a form of simulation mid-
way between fluids and rigid body physics; cloth and soft bodies.

Coth

This basically does what it says on the tin. You make a mesh,
define it to be made from cloth by adding a modifier and when you run
the animation it crumples and deforms like cloth. OK, so there are a
few details that you need to set up as well.

1. Save the liquid project and start a new scene, then with the start
cube selected, add a cloth modifier to it. When you start the
animation playing the cube will fall and disappear off the
bottom of the screen.

2. To give the cloth cube something to crumple on add a plane,

scale it up then move it down below the cube and add a
collision modifier. Now when you animate the scene the cube
drops onto the plane and stops, but it doesn’t look like cloth.

213
3. In order for a mesh to deform like cloth it needs to be
subdivided to a relatively fine level of detail. So add a
subdivision modifier to the cube, select Simple and increase the
subdivisions to 4. The animation still won’t work yet because
the subdivision is happening after the cloth, but when you move
the modifier up you should see a crumpling cube.

4. Now you can have some fun by adding internal pressure to the
cube. Tick the check-box for the optional Pressure section in the
Physics Properties tab then increase the value of Pressure to 5.0.
Now when you run the animation the cube will inflate and
bounce realistically on the plane. Duplicate the cube twice and
move the duplicates above and slightly to one side. If you try
running this you will see that you need to add a collision
modifier to two of the cubes as you did for the plane in order to
make them interact with each other.

5. Cloth simulation is a very quick way to model objects for static

scenes as well as for creating animations. With a little practice
cushions, clothes and drapes can all be made very easily this
way. Try adding a curtain to your scene in a position where at
least one of the bouncing cubes will hit it. Add a plane then
scale it up, rotate it 90 degrees about the X or Y axis and move
it to the side of the falling cubes where a cube will roll. In order
to hang the curtain along its length subdivide it in Edit mode so
there are 5 vertices on each edge. Select the vertices on the top
edge then, in the Object Data Properties tab, create a new vertex
group, rename it pins and assign the selected vertices. Switch
back to Object mode and add a subdivision modifier using the
Simple method and 3 levels of subdivision. Below this modifier
add a cloth modifier then open the Physics Properties tab and, in
the cloth settings, expand the Shape section. Select the pins
vertex group as the Pin Group and, in order to create some
vertical pleats, set the Shrinking Factor to -0.1. This will make
the cloth expand between the pinned vertices to provide some
“fullness”.

6. When you run the animation you should see the curtain drape
and then deform realistically as it is hit by the cubes. As with
the fluid simulations there are many setting to adjust and create
anything from a rubber blanket to silken gauze. Mix up some
materials as well to create an interesting scene as in Figure 11-3.
Note that this render was done using the eevee engine with
ambient occlusion, bloom, screen space reflections and motion
blur all turned on. Also I added a solidify modifier to the

214
 curtain, after the cloth modifier. Sometimes the Blender physics
simulation system may struggle with multiple different collision
objects and not produce the expected distortions on the cloth.
For the example shown in Figure 11-3 I had to make sure that a
cube with a collision modifier hit the curtain and that the curtain
did not have a collision modifier.

Figure 11-3 Cloth Simulation with Inflated Cubes and a Curtain

Soft Bodies

Unsurprisingly soft bodies lie part way between cloth objects and
rigid bodies and one of their most useful applications is to animate
parts of a character automatically. Details such as hair, jowls, bellies,
loose clothing or antennae should all be animated in time with the
movements of the character and this would be very time consuming to
do convincingly by hand.

You will apply a method using a soft body to animate the belly of
the peasant man salsa dancing FBX model you downloaded in Chapter
6.

1. Start a new scene, saving the cloth simulation for later, don’t
delete the start cube, this will become the soft body, then import
the salsa dancing FBX file following the instructions in Chapter
6. Select the armature then open the Object Data Properties,
which has changed to a green stick-man icon, and under
Skeleton select Rest Position.

215
 2. The peasant man will be submerged inside the cube so select the
cube that will become the soft body, and scale it down so it’s
about the size of the belly area then move it in the Y and Z
directions so one side is just inside the man’s body as on the left
of Figure 11-4, this is most easily done in orthographic view.
Just to preempt any confusing behavior later, select
Object4Apply4All Transforms.

Figure 11-4 Making a Soft Body to Deform the Peasant Man’s Belly

3. In Edit mode add an edge loop then scale and move the vertices
to more closely match the shape of the belly as in the middle of
Figure 11-4.

4. You need to define a vertex group to use for attaching the soft
body to the armature, in a very similar way to creating the pins
for the cloth curtain simulation. Still in Edit mode select the
four vertices nearest to the man, create a vertex group, rename it
pins and assign the selected vertices.

5. To attach the the soft body to the armature you will need to
parent it and add an armature modifier. The quick way of doing
that, which you used before, was to use the parent with
automatic weights option, however that would add a vertex
group for every bone in the armature and would try to guess
what weight to use for each. In this case you want to just use
one bone, the Spine, so the easiest route is, still in Edit mode,
select all the vertices then create a new vertex group and rename
it Spine. Note that it has to match the name of bone exactly,
then assign all the vertices to that group. Switch back to Object
mode and SHIFT select the armature as well, the soft body
should be highlighted in orange and the armature in yellow, then
select Object4Parent4Object. The final step is to add an

216
 armature modifier to the soft body and specify the armature in
the Object field.

6. When you change the armature Skeleton back to Pose Position

from Rest Position you should see the soft body move with the
armature as shown on the right of Figure 11-4. Of course you
haven’t actually made it into a soft body yet, so select the soft
body object and add a soft body modifier. All the settings for
soft bodies are in the Physics Properties tab so open that and
expand the Goal section. In order to attach the pins vertex group
to its start position but allow the rest of the body to deform, set
the Vertex Group to pins and the Strengths Default to 1.0. To
prevent the soft body being too soft, increase the Strengths Min
to 0.8 and set Settings Stiffness to 0.9 and Damping to 4.0. It
will also be more realistic if you increase the Object Mass to
2.5. All of these values, plus others will probably need
adjustment depending on the kind of object you are hoping to
simulate. Run the simulation to see how it looks.

7. When the soft body is behaving reasonably, it doesn’t need to

be perfect, you should set up another modifier to transfer some
of that movement onto the peasant man object. Select the
character skin then enter Edit mode and, in orthographic mode
with X-Ray enabled, select all the vertices in the belly area, as
in Figure 11-5. Create a new vertex group called belly, but
before you assign the vertices soften the edge of the selection by
reducing the Weight to 0.25, click Assign then reduce the size of
the selection by pressing Select4Select More or Less4Less,
or use the shortcut CTRL-MINUS, then change the Weight to
1.0 and assign the vertices.

Figure 11-5 Selecting the Vertex Group for Belly Movement

217
 8. To make the vertices move with the gyrations of the soft body
you need to switch back to Object mode and add a Surface
Deform modifier. Set the soft body, still named Cube, as the
Target and, in the Vertex Group field select belly. Change the
Strength to 0.3 then click Bind. In the outliner hide the soft
body object in the viewport and the render. The right hand
image in Figure 11-5 shows the two views superimposed; with
and without the modifier.
Using soft bodies to animate parts of characters can be very useful
but sometimes they behave a little too like jelly and it can be hard to
reduce the amount of wobbling. For things like hair, which are
relatively light and floppy, it is normally better to follow almost the
same procedure as that outlined above, but use a cloth modifier instead
of the soft body. One additional problem for dangling features, such as
pony tails, purses or sporrans, is that they need to be prevented from
swinging inside the character’s body. Rather than add a collision
modifier to the whole of the character mesh, which will have thousands
of polygons and a complicated geometry, it is much more efficient to
put the modifier on a simple mesh with only a few vertices and parent
that mesh to the armature as a deflector.

Rigid Bodies

In some ways rigid bodies are the simplest concept in the range of
Blender physics processes. Rigid bodies collide with each other in a
realistic way and can be connected together using different types of
links. However the construction of the rigid body constraints is not
very intuitive and takes a little practice to get the hang of. More
importantly, there can be conflicts in Blender’s internal machinery for
applying modifiers and other physics systems at the same time as rigid
body simulation. For instance, it is much harder, and less reliable, to
construct a system of rigid bodies connected by damped springs to
simulate the swaying belly of the peasant man than to use a soft body
or cloth simulation.

To finish off this chapter you will do one more exercise modeling
six spheres rolling into a bowl. This simulation will be used for an
animation in Chapter 13 in which the spheres turn into bugs that uncurl
and start crawling out of the saucer. There will be five bug spheres
which will be joined together into a kind of train using rigid body
constraints and there will be a heavier sphere that will be launched
around the bowl to jostle them.

218
 A problem you will encounter with rigid bodies, sooner or later, is
their tendency to pass through each other. Apart from forgetting to
define them as rigid bodies, the most frequent cause of this behavior is
due to scaling meshes and not applying the transform to the vertex
locations.

1. Start off by making a bowl from a sphere by inverting the top

half of it. Start a new scene, delete the start cube then add a new
sphere object. Scale the sphere 10 times then scale it down in
the Z direction by a factor of 0.5. In Edit mode use orthographic
and X-Ray view to select the vertices above half way, scale
them down slightly then invert the curve by scaling them in the
Z direction by a factor of -1.0. Move the selected vertices down
and adjust and scale them, then by selecting edge rings at the
top expand and flatten the lip so create a bowl as shown in
Figure 11-6.

Figure 11-6 Simulating Balls and a Bowl

2. To make the bowl interact with the balls when they’re added
you must make it into a rigid body. With the bowl selected open
the Physics Properties tab and click on Rigid Body. There are
several things you need to correct with this setup but you will
pick them off one by one, that way you can see the symptoms
and, if they occur later, you will know how to fix them.

3. Add the first ball as a UV sphere and in the Adjust Last

Operation panel change the Segments to 16, Rings to 8 and
Radius to 0.2. Enable rigid body behavior for this sphere too,
then move it up, higher than the rim of the bowl and to one side,
so it will drop onto the curved part of the surface.

219
4. When you play the animation you will see that the bowl and the
ball both fall under the effect of gravity. In order to stop the
bowl falling you need to tick the Animated check-box in the
Settings section of the rigid body. The bowl should stay where it
is now but the ball will stop abruptly just above the bowl, level
with the top edge. That’s because the collider shape defaults to
Convex Hull, but the bowl is, by its very nature, concave. If you
position the ball below the rim of the bowl it will be ejected
dramatically by the simulation.

5. To make the ball bounce off the actual surface of the mesh
change the Collisions Shape to Mesh, for the bowl, but for the
ball set it to Sphere. Now when you run the animation the ball
will very likely drop through the mesh, if it doesn’t, try moving
the start point around, the majority of locations should cause the
ball to fall through.

6. You can probably already guess the solution to this problem

because it’s cropped up so many times before; you need to
apply all the scaling you did on the sphere that got turned into
the bowl. With the bowl selected hit Object4Apply4All
Transforms and now you should see the ball interacting with
the bowl correctly. If you still have problems it could be that the
mesh of the ball is no longer around its center point, in which
case use Object4Set Origin4Origin to Center of Mass
(Volume)

7. There is one other change you should make now, before you
start carefully positioning the balls to create an interesting
animation. The smooth or flat shading option changes the
normal directions that are used by rigid body physics, as well as
for light reflection, so you should set the option to smooth
shading now.

8. The ball you just made will be the heavy ball to knock into the
bugs so keep its weight at 1.0, but to keep it rolling round and
round the bowl you should reduce the friction. In the rigid body
properties, under Surface Response reduce Friction to 0.1 and
Bounciness to 0.5, under Dynamics reduce the Damping
Translation to 0.0 and Rotation to 0.0.

9. To create the bugs, duplicate the ball and change its mass to 0.1
then duplicate the lighter ball four more times. Position these
five balls on the flat edge of the bowl so they are just above the
surface and separated by about one diameter. The idea is that

220
 the bugs will slowly fall into the bowl and the heavier ball will
be fired horizontally so it hits them. You can see the locations in
Figure 11-6. Make sure that the balls are not intersecting the
surface of the bowl otherwise they will experience a strong
repulsive force in the first frame of the animation!

10. To join the bug spheres into a train you need to add four
constraints. In Blender each constraint is expected to be added
to a new empty object and the constraint specifies the two other
objects to be linked as well as all the physical properties of that
link. For most rigid body constraints the location of the empty
object relative to the two linked objects is critical to its behavior
and behaves most reliable if placed mid way between them.
Fortunately there is a menu option to automate much of this.
Select the first and second bug then use Object4Rigid
Body4Connect, then select the empty object that has been
generated and in the settings for the rigid body constraint
change the Type from Fixed to Point. Repeat the process until
all five balls are joined together.

11. When you run the animation you should see the first ball slowly
roll away from the flat edge area and drag the other four with it.
You might need to adjust the balls’ positions or tweak the mesh
of the bowl, but if you do move a ball you should move the
empty object that holds the constraint so it’s approximately mid
way between each of the balls. Notice that the empties do not
move with the balls but the constraints are constantly re-
calculated as if they had done.

12. To give the heavier ball some starting velocity you will animate
its position for the first few frames and then switch back to
simulation. Select the heavier ball and, in the rigid body settings
tick the Animated check-box under Settings. Make sure that the
animation position in the Timeline is set to frame 1 then, with
the cursor over the check-box press I to insert a keyframe. With
the cursor over the 3D Viewport press I again and select
Location. Now move the position on the Timeline to frame 5
and move the ball in the 3D Viewport in the direction you want
to fire it so it will hit the five bug balls. Insert a keyframe here
by pressing I and again selecting Location. In the rigid body
un-tick the Animated check-box and insert a keyframe for that
by pressing I with the cursor over it. If that was all done
correctly then, when you play the animation you should see the
ball set off circumferentially round the bowl.

221
 13. It is very unlikely that you will hit the bug balls first try. Often,
if you do hit them, either the heavy ball or the bug balls fall out
of the bowl. I found the easiest way to get the right direction
and speed was to view the bowl from straight above and keep
moving the first frame location of the heavy ball, remembering
to reset the keyframe each time by pressing I. Because you have
a keyframe to turn animation off on frame five you will not be
able to move the location of the ball at that point unless you
temporarily tick the Animated check-box. To increase the initial
speed increase the distance between the first frame position and
the fifth frame position. I also had to move all five bug balls a
small amount, along with their constraint objects, to get the
timing right.

14. In the end I managed to get two collisions within 250 frames. If
you want to run your animation longer then you can change the
End field in the Timeline, however there is a separate location
where the end of the simulation is set, and this defaults to 250
frames. To change this, open the Scene Properties tab, which
has a cone and sphere icon, expand the Rigid Body World
section and then the Cache section where you can set the
Simulation End field to match the end of your animation.

NOTE The Scene Properties tab also contains settings that allow you to fix
problems with physics in the rigid body world. Increasing the
accuracy of the simulation can prevent fast moving objects
“tunneling” through other objects and similar effects, but at the
cost of slowing down the calculations. There are similar controls
available for fluids, cloth and soft bodies in the Physics Properties
tab.

You will use this scene in Chapter 13 as the basis for a short video
using cartoon style shading so there’s no point working on the materials
and other details now, just save the project along with the other blend
files for this chapter.

Conclusion

You have glimpsed each part of the wide range of physics engines
built into Blender. You should have an appreciation for the potential
traps, and how to avoid them but you probably feel there are several

222
different mechanisms all nearly doing the same thing, so it might be
worth taking a step back and reviewing how they fit together.

There are strengths and weaknesses to each of the different

simulation systems and knowing which one to use in a given
circumstance is an important skill to learn. At a superficial level there
are two categories of physics; one uses vertices as its components and
one uses individual objects. The vertex level simulation includes
everything apart from rigid bodies and a characteristic is that there are
different forms of baking, but the baked information is not then
accessible, it exists to speed up subsequent animation runs. For rigid
bodies the animation information consists of object locations and
rotations and this can be integrated more easily with animation
keyframes. You used this facility to apply an initial impulse to the ball
in the last exercise and it is possible to convert the whole simulation
back to an animation by baking keyframes, which you will do with the
bugs in Chapter 13.

For cloth and soft body physics the distinction between the two
systems becomes much less distinct. It is possible to create bouncing
cubes using rigid bodies, cloth or soft bodies and your choice will
depend on how important it is to see realistic deformation of the mesh,
the complexity of the scene and the requirements for other constraints
on objects’ movement.

The relationship between physical simulation and animation is

very close. You have been adding keyframes to animate aspects of
Blender quite extensively, both in this chapter and the previous one,
and in the next chapter you get to grips with animation in a much more
comprehensive way. You will learn how to build an armature from
scratch and how to rapidly fit an armature to an existing mesh in order
to animate it.

223
 12
A N I M AT I O N A N D R I G G I N G

From building and animating

armatures through to rendering and
video production.

Creating 3D models that can be posed quickly and easily

is a first step to making impressive animated scenes. But
the process of controlling all the components including
the camera, and then re-building the rendered frames
back into a finished video are also critical stages.
By this stage of the book you have seen objects being animated in
the 3D Viewport several times. What’s more you are now fairly

224
 familiar with the process of adding keyframes to a property, such as
when you applied an initial velocity to the ball in the last chapter. This
chapter will provide some depth to your understanding of keyframes
and introduce alternative editors that can make editing scenes simpler.
However, before that, you will learn some basics of rigging by making
an armature to control your own model and then by adding a standard
armature to a previously un-rigged figure.

Rigging Hard Surfaces

The armatures that you have used so far have all been controlling
soft surfaces. The manikin you made in Chapter 4 and the peasant man
and lola that you imported as animated .fbx models in Chapter 6 all
deformed smoothly at the joints as you rotated the different bones.

Sometimes you need to move parts of an object without inducing

any bend or stretch, such as where the material is supposed to be steel
or plastic. In the following exercise you will make a robot arm
consisting of six very simple meshes, see Figure 12-1. You will then
add an armature and see how to constrain and then animate its
movement. Links to the files and videos for the exercises in this chapter
are at get-into-blender.com 12.ANIMATION AND RIGGING

1. Start a new scene and keep the cube, renaming it “robot_arm”.

The cube will be the platform of the arm, so switch to Edit
mode and move the bottom four vertices up so they are level
with the origin and scale the top four vertices in a little.

2. You will now add five more meshes while in Edit mode so they
will be part of the robot_arm object but the vertices will not be
connected. First make the two arm sections by selecting
Add4Mesh4Cube then scaling the mesh down in the X and
Y direction with S then SHIFT-Z then 0.2. Move the new cube
in the Z direction so it’s above the platform then duplicate it and
move the duplicate in the Y direction. Now create the cylinders
for the joints, after adding the mesh open the Adjust Last
Operation panel and change the Vertices to 16, Radius to 0.2
and Depth to 0.4 then move the cylinder in the Z direction until
it is almost completely protruding from the top surface of the
platform. Duplicate the cylinder, rotate it about the Y axis by 90
degrees, move it in the Z direction so the axis is about level
with the top surface of the original cylinder then duplicate the

225
 second cylinder and move it in the Z direction so it’s clear of the
other components.

3. Now you should arrange all your components into a robot arm
shape using the orthographic view along the X axis as shown on
the left of Figure 12-1. So long as you specified the Z or Y
directions when you did the movements in the last paragraph,
everything should lie in the plane X = 0, but check this by
viewing along the Z or Y axis. You can select all the vertices in
each of the meshes by first selecting one vertex then running
Select4Select Linked4Linked. But note that you will need to
do that selection many times during this exercise so you might
want to adopt the shortcut CTRL-L.

Figure 12-1 The Robot Arm Showing Distortion of the Mesh

4. You will now add an armature to control the robot arm. Switch
back to Object mode and select Add4Armature4Single
Bone. Keep in orthographic view from the side and switch back
to Edit mode. The starting single bone of the armature will be
the “root” bone for positioning the whole robot arm assembly
and you are now going to extrude three more bones for the
remaining moving parts. The first bone will be for rotating the
arm about the vertical axis, the second will be for rotating the
first stage of the arm about a horizontal axis and the third bone
will be for rotating the last part of the arm. The cylindrical
hinges represent the mode in which the robot arm is intended to
bend.

226
NOTE Armature bones are directional, in that each has a root, about which
they will rotate when posed, and a tip. In Edit mode it is possible
to select either the tip of a bone, in which case extrusion will cause
a new bone to branch from that tip, or the whole bone can be
selected, in which case a new bone will also be created from the tip
of the previous bone in the chain, if one exists.

The default bone representation is the octahedron, as used here, to

make it clear which end is the root and which the tip. There are several
other options to represent bones, including the stick method you saw in
Chapter 4, and in addition it is possible to create custom shapes to use
for each individual bone as you will see below when you use rigify.

5. Select the top end of the initial bone and extrude in the Z
direction until the tip of the new bone is at the center of the
horizontal cylinder. Extrude the next bone and guide its tip to
the center of the next hinge cylinder. Finally extrude the last
bone so its end is at the end of the robot arm. On the left of
Figure 12-2 you can see the positions of the bones.

6. To set up the parenting, modifier and vertex groups, switch back

to Object mode and select first the robot_arm object then the
armature. Use Object4Parent4With Automatic Weights,
then select the armature, switch to Pose mode and try rotating
various bones.

7. You will find that there are two problems as shown in the
middle of Figure 12-1. The first is that the joints all behave like
ball and socket with no restriction of how far, and about which
axis, rotation takes place. The second issue is that the cylinders
become distorted as the arm is posed, and move relative to the
cuboid section of the arm. Switch back to Object mode, select
the robot_arm object and switch to Weight Paint mode. In the
Object Data Properties select each of the vertex groups and see
how the color varies. On the right of Figure 12-1 the weights
are being shown for vertex group Bone.002, some vertices are
being affected by more than one bone and, for any given bone,
some vertices are being affected more than others. This isn’t
what you want for a supposedly solid component.

227
 Resetting Vertex Groups

8. First remove all the vertices from all the groups. Switch to Edit
mode and select all the vertices, then, in the Object Data
Properties tab select each bone in turn and click Remove. Now,
select a vertex on the platform mesh and extend the selection to
all linked vertices, then select the Bone vertex group in the
Object Data Properties tab and click Assign. Repeat this process
with the vertical axis cylinder and Bone.001. For Bone.002
select a vertex on the next cylinder as well as the first part of the
arm before selecting linked vertices and assigning them. Finally
select the last cylinder along with the last part of the arm to
assign to Bone.003.

Bone Constraints

9. Now when you pose the armature there is no distortion but the
hinges still have too many degrees of freedom. You need to fix
that by setting bone constraints. With the armature selected and
in Pose mode select Bone.001, the one that should only rotate
about a vertical axis, then open the Bone Constraints Properties
tab and select Limit Rotation. It might seem logical to set the
X and Y limits to 0 for Max and Min then limit Z to -45 to 45.
And this would appear to work, but only so long as the platform
wasn’t tilted at all. That’s because the default frame of reference
for the rotation limits is World Space and you want your bones
to rotate relative to the position of the previous bone, which is
Local Space, so you must change that for each bone. However
before you start trying to set the limits for each of the bones you
should open the Object Data Properties tab, and in the Viewport
Display section tick the check-box next to Axes. Now you can
see that Bone.001 should only be allowed to rotate about its
local Y axis and the next two joints should only rotate about
their local X axes. The actual values of the Max and Min will
depend on the starting rest pose and the configuration of the
joints so you will need to figure those out by a process of trial
and error.

10. The robot arm is now behaving like a piece of machinery and
can be posed by rotating each bone. However it’s quite a long-
winded process if you wanted, say, to make the tip of the robot
arm follow the outline of a square on the floor. To do tasks like
that, very quickly and easily, Blender has an inverse kinematics
solver available as a bone constraint, which you will now add to

228
 your robot arm. Inverse kinematics, or IK, is widely used in
game engines and other posing software to attach a hand or foot
to a given location, notoriously to prevent feet sliding over the
ground as a character walks. To act as a target for the inverse
kinematics, add an empty object to your scene and move it
somewhere in front but within reach of the robot arm. Select the
armature, change to Pose mode and select the last bone,
Bone.003. In the Bone Constraints Properties tab add an inverse
kinematics constraint and set the Target to the new empty
object. The tip of the bone should reach to the empty but you
will probably see the whole robot arm tilt. This is because you
haven’t limited the IK solver to just the last three bones. Change
the value of Chain Length to 3.

Inverse Kinematics

11. When you move the empty object around, the robot arm will
follow it but the bone constraints for the joints have stopped
working. This can be rather confusing at first, however the
reason is that there are additional IK controls within each bone’s
Bone Properties tab. Not only is it possible to add constraints
for joint rotation but joints can have variable stiffness to resist
rotation in some directions rather than others. For this exercise
just go through the Inverse Kinematic settings for the three
bones in the chain and set the rotation constraints to match those
set up as bone constraints. You will see that there are useful arcs
drawn in the 3D Viewport to represent the range of movement.

12. On the right of Figure 12-2 you can see that I added three
different materials using the vertex groups of my robot arm and
also a bevel modifier, the lighting is using an environment
texture. Animate the robot arm by moving the empty object and
setting keyframes for its location. You could do all kinds of fun
things such as following the outline of lettering or following the
seams of a simple mesh that was “appearing” one face at a time
using a build modifier.

229
Figure 12-2 Setting the Robot Arm Joint Constraints

Once you have set up some materials and an interesting animation

for the robot arm it’s time to render the animation.

Rendering an Animation

1. In the Output Properties tab under the Output section open the
file browser and create a new directory in the location for this
chapter’s files then click Accept to set the directory. Unless you
have a powerful computer it’s probably worth setting
Resolution percent to 50 in the Dimensions section as well.
Leave the Render Engine as Eevee in the Render Properties tab.
Note that the rendering is still set to produce png images rather
than a video format and this is the preferred option, partly for
quality reasons but also, if something goes wrong or you need
to stop the render part way through, you can carry on rendering
from the next frame at a later stage.

2. Using the main menu at the top of the window select

Render4Render Animation. You will see each frame as it is
being rendered with the time taken for each, so you can
calculate how long the whole job will take. Normally it takes
long enough to make a cup of tea, sometimes long enough to go
to bed and check again in the morning!

The Video Sequencer

3. Switch one of the editor windows to a Video Sequencer and

from the menu choose Add4Image/Sequence, then browse to
select all the images rendered previously and create a strip.

230
 4. At this point it would be possible to do various adjustments and
compositing but for this exercise simply save the sequence as a
video. Open the Output Propterties tab and browse in Output
the same folder you saved the frames and change the File
Format to FFmpeg video. Finally, from the top menu select
Render4Render Animation again.
The FFmpeg video format is the most flexible to use and the
default encoding is Matroska, however, because of the mess of
different standards and intellectual property claims you might find that
your particular version of your particular operating system won’t play
the video. You can either find a video player that will play the .mkv file
you created or you can choose from a wide variety of different
encoding formats such as MPEG-4, AVI, Quicktime or WebM. If you
get stuck there is lots of help available online.

Now you know the basic way to generate a video you can convert
the simulation exercises from Chapter 11 to artistic masterpieces, and
start to build up your youtube following.

Rigging a Standard Armature to an Existing Mesh

Being able to find a free 3D model online and quickly add an

armature to modify its pose, or to animate it, can be a very useful skill.
For instance, if the angel you imported in Chapter 6, looked just right
but you wanted it to be sending a text.

The job of rigging is actually very easy to do in Blender with only

a couple of potential gochas. Normally it is done on a character mesh
that is symmetrical, standing in standard pose and the benefits of that
are two-fold. Firstly it’s only half the work as positioning bones on one
half of the armature can be reflected onto the other side. Secondly the
rotations of joints and distortions created are much more predictable
when starting from the standard position. However the process you
need to go through will be essentially the same as the steps for this
exercise, and, as I said above, being able to “tweak” an un-rigged
model can be very useful on occasion.

1. Start a new scene, save the robot arm project and delete the start
cube. Use the top menu File4Import4Wavefront (.obj) and
navigate to where you downloaded the files for Chapter 6.

231
 2. Convert the model to the correct size and orientation by scaling
to just under 2 grid lines and rotating about the X axis by -90
degrees. I’m sure you know this by now, but if you scaled and
rotated in Object mode you need to apply the transformations.

Rigify

3. The rigify addon is installed and enabled in Blender by default

after version 2.8 and this includes some ready-make skeletons.
Add one now using Add4Armature4Human (Meta-Rig)
then scale the armature down so it matches the angel, as near as
you can. You will delete the bones for posing the face so the
main points of reference are the location of the shoulders and
length of the arms. You have probable already done it, but apply
the transformations.

4. With the armature selected switch to Edit mode and roughly line
up the armature with the mesh. For this first stage I suggest you
stick with box select mode and select parts of a limb to move
and rotate as one block, positioning in stages. For instance, first
select the whole of the right arm and rotate it to about the
correct orientation and position it so the elbow is in the right
place, then select just the forearm and hand and adjust it so the
wrist is nearly in the right place, and finally select just the hand
to rotate the wrist to match the angel’s hand. Obviously you
need to move the view position around to help with this but you
will also benefit from switching between normal rotation, in the
plane of the screen, and track-ball rotation using shortcut R
twice. On the left of Figure 12-3 you can see the final
positioning of the right hand using this technique.

Figure 12-3 Positioning the Bones of an Armature

232
5. The detailed face features of the human meta-rig are too
difficult to match to the angel’s face, partly because the pose is
not symmetrical, but you can do without them for this exercise.
Carefully select all the face bones, without selecting the head or
neck bones (the bones you need to keep will be called
Spine.006, Spine.005 and so on) and delete them. The Face
bone may be hidden by the Head bone.

Snap to Volume

6. You should be able to get the feet lined up by selecting the three
bones involved and rotating and moving them as described
above. To position the knees you need to select just the ball
representing the joint between the femur and the shin bone
which can be moved without disturbing the position of the ankle
or hip joints. On the right side of Figure 12-3 I have selected the
first joint from the tip of the right ring finger. However notice
too, at the top of the image, that I have switched snapping on
and set the type of snapping to Volume, this is crucial for
rigging animals. The other setting you may find helpful is to
change the selection mode to Tweak, the shortcut for quickly
changing from one selection mode to another is W. You should
now find the rigging of all the finger bones really quite easy.
When you’ve finished, don’t forget to turn snapping back off
otherwise you will be scratching your head when you come to
pose the armature later.

7. When all the bones are lined up you might be tempted to parent
the angel object to the armature using automatic bone weights,
as you have done before, and you would then be able to change
the pose. However this would be hard work, especially to get
the finger joints to move realistically. So, instead of this, you
will convert the armature to a rigify armature and then use that
to pose the model. Select the armature in Object mode then
open the Object Data Properties tab, expand the Rigify Buttons
section and click on Generate Rig.

8. Sometimes there is an error while the rig is being generated,

usually caused by a gap in the chain of bones that rigify expects
to be continuous. If you look at the human meta-rig armature
you will see that there are some bones, such as the collar and
thighs, that are separated from their parent by a dotted line. This
allows extra articulation in the skeleton, for instance the upper
leg will rotate about its root but the root can also be moved

233
 slightly without causing any connected bones to rotate.
Unfortunately there are several bone ends that rigify expects to
be coincident, some in the face and one between Spine.003 and
Spine.004. The easiest way to make one end of a bone coincide
with the end of an unconnected one is to set snapping to Vertex,
so, if you get an error of this type when you tried to build the
rig, use snapping to position the bone ends and try again.

THE 3D CURSOR

If you search online for help with problems converting an armature to a rigify rig,
you might well find a suggestion as above, but with instructions to first snap the 3D
cursor to the position of one bone end and then to snap the second bone end to the
position of the 3D cursor. If you look in the menu for the armature in Edit mode you
will see, under Armature Snap, a series of options including the ones involving the 3D
cursor.

Up to this point in the book I haven’t mentioned the 3D cursor, not because it
isn’t sometimes useful, but the number of occasions where you need to use it are
much reduced compared with the earlier versions of Blender. The 3D cursor is the
default spawn point for objects added to the scene and it is used by several of the
mesh transforms such as Bend, Warp and Spin.

9. Once the rigify armature has been created you can parent the
mesh to it for posing. First hide the metarig armature object
which will not be used any more, then select first the angel
mesh object, then the rig object, and select
Object4Parent4With Automatic Weights.

10. Try posing a few bones to ensure everything worked correctly.

First double-check you turned off snapping and proportional
editing, then make sure you can see all the bones by select the
rig object, and in the Object Data Properties tab, under the
Viewport Display section tick the In Front check-box. Switch to
Pose mode, select the left hand IK bone, which is a rounded
rectangle and colored red, and move and rotate the hand. You
should find that the angel bends their arm and shoulder in a
fairly smooth way. Now try moving and rotating the head, neck,
abdomen, hips and so on. Some of the bones will create more
severe distortion than others, so you have to be careful with
those. The fingers have a strange wire with a spatula sticking
out of them. This can be rotated and, as with the other bones,
the fingers will rotate about the point where they are attached to
the palm, however if you scale these bones the fingers will curl

234
 and uncurl which is a very useful feature for posing. Remember,
in order to rotate in the plane of the screen use a single press of
R and to track-ball rotate press R twice.

11. The next step in this exercise is to construct a short animation in

which the angel pulls a mobile phone out of their back pocket,
glances down at it, touches the screen, puts the phone back in
their pocket and returns to the start pose. First of all you need to
make the phone. Add a cube the scene, scale it down to an
appropriate size and shape, add a bevel and then a subdivision
modifier, add a new material with black base color and reduced
roughness and finally set to smooth shading. Position the phone
behind the angel’s back, there’s no need to model the pocket or
other details, the hand will just reach around out of sight and
reappear holding the phone. You should also make sure that the
camera angle doesn’t view the screen of the phone or the
angel’s face in too much detail.

12. Now add some keyframes for the movement of the angel’s arms
and head. Select the rig object and switch to Pose mode, then
select the right hand IK bone and insert a keyframe by pressing
I and selecting Location, Rotation & Scale. Move the current
frame pointer in the Timeline on to 50, which is about 2 seconds
then position the right hand about where the right, back pocket
would be. Rotate the hand so the fingers are pointing
downwards and the palm is twisted towards the angel’s back.
The hand will be hidden from the camera but the IK process
might propagate twist back up the chain of bones which would
impact on the rotation of the elbow and shoulder. Insert a
keyframe here. Move to frame 100 and position the right hand
in front of the angel’s chest, somewhere it would be comfortable
to read a text message, and insert a keyframe. Don’t worry
about positioning the fingers at this stage, you will do those
later.

13. To return the phone and strike the original pose you will copy
these three keyframes. Make the Timeline area a little larger
then select the first keyframe, duplicate it using SHIFT-D and
move the duplicate to frame 250. Duplicate the keyframe at 50
and put a copy on frames 75, 175 and 200, the same position
from frame 50 to 75 and from 175 to 200 represent the pause
while the phone is being taken out of, and reinserted into, the
back pocket. Duplicate the keyframe at 100 to frame 150 which
will give the angel about two seconds to see the message and
press the acknowledge button.

235
NOTE The Timeline shows the keyframes for each of the selected objects
in the scene that have had any keyframes previously added. On the
left is a collapsible hierarchy of detail, and if you expand that, you
will see that there are keyframes for each component, enabling, for
instance, the X rotation to take place prior to the Y rotation. When
you copy or move keyframes, especially if more than one object or
bone is selected, you need to be careful that you only select the
ones you intended.

14. Run the animation, and where the movement is awkward, or

attempts to pass through the angel, position the hand correctly
and insert a keyframe until you have created a plausible
sequence of movements.
It turns out to be rather hard to animate humanoid armatures
without making them look a bit like the robot arm in the first exercise.
For many animation sequences, such as the two Mixamo models you
downloaded in Chapter 6, the positions and rotations of bones have
been inferred from motion capture of human actors. This results in very
realistic movement, but it creates lots of keyframes, and is extremely
difficult to reproduce by hand. You already know one way to improve
the result by deforming the mesh using physical simulation of cloth or
soft bodies, but for this animation you will use the Graph Editor and
see the functionality that is available there. Before that, however, you
need to make the phone attach itself to the angel’s hand while it’s out of
sight, you will then be able to animate the left hand and the head.

The Child Of Constraint

15. To make the phone move as if attached to a bone in the

armature is actually quite easy using object constraints and there
are several ways to make an object copy the location, rotation or
transforms of a bone in an armature. The problem is you don’t
want the phone to be positioned exactly where the hand bone is,
it needs to be offset into the palm of the hand and getting that to
work is much easier using the Child Of constraint. Move the
frame pointer along to 70, then in Object mode select the the
mobile phone object and open the Object Constraints Properties
tab. Select Child Of and set the Target field to rig and Bone to
hand_ik_R. Add a keyframe for the Influence field set to 1.0 by
pressing I with the cursor hovering over it. Move back one
frame to 69, change the constraint Influence field to 0.0 and set
a keyframe for that.

236
 16. You now need to set the position and rotation of the phone for
these two frames. At frame 69 the phone should still be behind
the angel’s back, so press I with the cursor over the 3D
Viewport and select Location, Rotation & Scale. Move
forward to frame 70 and position the phone correctly relative to
the hand. You will sort out the finger positions later so try
holding your own phone and see where it lines up relative to
your palm, it only needs to be fairly approximate. Insert a
keyframe here as well, then, in the Timeline duplicate the
keyframe at 70 to 180 and 69 to 181, the reverse order for
letting go of the phone.

17. Sort out the fingers on the right hand by moving to frame 69
then selecting the rig armature and entering Pose mode. Select
the spatula controls for all four fingers and the thumb then enter
a keyframe so they keep their starting positions. Move on to
frame 70 and pose each finger by scaling and rotating the
spatula controls, once each finger is correct save a keyframe for
it. It might be easier to see what it looks like by moving forward
to frame 140, say. If you enter the keyframes at this later time
then you must move them back to frame 70 by dragging them in
the Timeline area, see Figure 12-4. When it looks good,
duplicate the keyframes in the Timeline in reverse order to 180
and 181.

Figure 12-4 Keyframes for Gripping Phone

18. Now that you have a phone with which to interact, you should
animate the left hand, the head and the neck. I did a little role
play to get the timings of when the angel might start to move
their left hand, and glance down. It seemed more natural to do
most of the phone extraction and lifting up before moving the
left hand to press the button and glancing down at the screen.

237
 You could add some slight shoulder or torso movements as well
if you thought they made the animation more realistic. The left
hand is already held in a pose suitable for touching a screen but,
again, you might like to add a little movement to add some life.

The Graph Editor

19. The Graph Editor area extends the functionality of the Timeline
by representing not only the time of each keyframe but also the
value at each point and the shape of the curve between the
keyframes, see Figure 12-5. Change the Timeline area to a
Graph Editor then, with the rig armature in Pose mode, select
the right hand IK bone. As you can see there is a lot of info
squashed in here so there are two ways you could reduce the
clutter; the first is to hide the channels you’re not editing by
clicking on the eye symbols on the left, the second is to adjust
the ends of the sliders to set the range of the area to match the
graph.

Figure 12-5 Adding Noise Using the Graph Editor

20. This editor is a good place to make subtle, accurate changes to

an animation, it would have been ideal for aiming the ball in the
last chapter, but for this exercise you will just use it to add a
little noise to the arm movement. Each curve can have a series
of modifiers applied to it in a stack which is accessed using the
right sidebar, open that now either by dragging the
inconspicuous arrow or using the shortcut N. Select the channel
for the Z Location, and from the Add Modifier drop-down, click
Noise. Slow the the noise down by increasing Scale and reduce
the amplitude by reducing Strength. Increasing Depth to 2 will
make the noise more detailed, and you can make the shaking
start after frame 1 and finish before frame 250 by setting values
in Restrict Frame Range. Adjust the settings and run the

238
 animation until you get the right level of variability to stop it
looking too much like a robot arm but short of delirium
tremens!

21. Once you have all the animation working well, with the right
amount of noise, you can add some artistic style and create
another video. These were the features I added, but you
shouldn’t feel constrained by this list: A marble material for the
angel, mainly different scales of noise texture with Normalize
un-checked for subsurface color and bump map. A shape key to
rotate the eyes and eyelids, animated to flick downwards when
the head tips forward. An environment texture of a suitable
indoor space, a cloth plane to animate the right wing, and
finally, camera animation. To add keyframes to the camera,
select it in the outliner then switch to camera view in the 3D
Viewport, set the frame in the Timeline area, move the camera
using the walk navigation controls, and save keyframes by
pressing I and choosing Location & Rotation.

Creating a More Complex Video Sequence

For the third exercise of this chapter you are going to take the two
characters you downloaded from Mixamo and use their animations in a
creative way. The first problem is to stop Lola’s hands from moving
while they should be gripping a hold, which you can do using an IK
constraint. You then need to apply the climbing up the wall animation
to the peasant man, and when they both reach to top of the wall you
need to apply the salsa dancing animation to Lola so the two characters
can celebrate together. You will use the wall factory add-on to build the
castle walls and a cloth simulation to create a rope.

1. Save the last exercise and start a new scene. Delete the start
cube then import the .fbx file of Lola climbing a wall which you
downloaded for Chapter 6. Add two empty objects to the scene
and rename them left_hold and right_hold.

Making Hands and Feet Stick

2. You now need to find where would be a good place to fix the
holds. Run the animation one frame at a time by pressing
RIGHT, and when the left hand looks to be latching onto a hold
move the empty up so it lines up with the tip of the hand bone.
You will need to change the Viewport Display option for the

239
 armature to In Front in order to see all the bones. Use
orthographic view from the side and the front to help position
the empties. As you move the animation on, the hand will slide
slightly but that’s the point of doing this exercise so don’t
worry.

3. You will add an IK constraint for each hand to make them stick
to each of the holds, and you will animate the strength of the
constraint from 0.0, a few frames before the hand lands on the
hold, to 1.0 at the start of holding on, then 1.0 at the end of
holding on, to 0.0 at the end of letting go. Because the
animation loops very nicely, the constraint for both hands
should be 0.0 for the first frame and the last frame, otherwise
there will be a discontinuity in the hand positions. Move back to
the first frame, select the armature, switch to Pose mode and
select the right hand bone. In the Bone Constraints Properties
tab select Inverse Kinematics then set the Target to
right_hand and Chain Length to 4. The right hand is actually
on the hold in the first frame but set the Influence to 0.0 and add
a keyframe, move forward three frames, set the Influence to 1.0
and add the next keyframe. The frame to start letting go is when
the left hand starts to curl over its hold. Move forward to that
frame and insert another keyframe with Influence still at 1.0
then allow the motion to blend smoothly into the animation by
setting the Influence to 0.0 ten frames later. Figure 12-6 shows
the sequence of keyframes.

Figure 12-6 Keyframes to Control IK Influence

4. Repeat the procedure for the left hand. The first keyframe
should be for Influence 0.0 ten frames before the hand reaches
the hold and, at the end of the loop the influence should drop
smoothly to 0.0 again by the last frame. When you run the
animation the hands should stay in place once they’re grasping
a hold.

240
 Ideally the animation would last for several loops, to allow you to
record a few seconds of continuous climbing, and you can probably see
one way to do that. It would be quite easy to copy and paste all the
keyframes so they were repeated a number of times, then in the Graph
Editor, increase the values of the Z component of the armature root
bone so that the second loop continued from the end of the first, and the
third continued from the second. However there are deficiencies with
that technique, the first one being that the empty objects controlling the
IK constraints would have to be moved up at the end of each loop. The
second problem is more practical, where you are using a short
animation repeated many times, for instance the legs of a bug scurrying
along, it becomes very hard to manage a scheme based on copying and
pasting.

Baking Constraints to an Animation

The best solution is to bake the animation, including the IK
constraints, into a series of posed positions, which is essentially the
form of the animation imported in the .fbx file. Blender then gives you
the option of creating an action from the animation, which you can can
use to make a strip to repeat multiple times. You can also apply the
action to other armatures, so long as they have the same bones.

5. To bake the animation, select the armature and enter Pose mode.
Select all the bones then from the menu click
Pose4Animation4Bake Action and tick the four check-boxes
Only Selected Bones, Visual Keying, Clear Constraints and
Overwrite Current Action. If you don’t tick the last option, you
will preserve the animation imported with the .fbx, and Blender
will generate a second action with the generic name “Action”,
“Action.001” and so on. This might be useful if you were going
to make different versions of the animation, each requiring
tweaking, but in this case it will be fine to just overwrite it.
Bake Data should default to Pose but make sure the Start Frame
and End Frame are correct, then click OK.

6. After a pause the keyframes should have been overwritten and,

when you run the animation, it should look no different.
However if you check in the Bone Constraints Properties tab
with either hand bone selected you will see that the IK
constraints have gone and the two empty objects can now be
deleted. If you select both of the hand bones and look in the
Timeline area you will see that the two channels of keyframes

241
 are still there, but underlined in red to show there is an error.
Delete each of the lines.

The Dope Sheet and Nonlinear Animation

7. There are two more editor types that you will need to use for
more complicated animations. The first one to look at is the
Dope Sheet which allows you to see keyframe positions for a
variety of different types. Change the area at the bottom from
Timeline to Dope Sheet, then change the Editing context drop-
down from Dope Sheet to Action Editor. As shown in Figure
12-7. At the top of the area some options have appeared; Push
Down, Stash and a drop-down to Browse Action to be linked to
the object. At the moment there is just one action, but now is a
good time to rename it to climbing_baked_ik. When you import
the Peasant Man .fbx there will be another action and this is
where you will be able to swap them, to make the Peasant Man
climb and to make Lola dance.

Figure 12-7 The Dope Sheet Editor

8. In order to make an action strip which can be repeated a number

of times change the editor type to Nonlinear Animation. This
view has a different section for each animated object in the
scene, and each section has a list of actions in orange, for this
scene there is just one object and one action. On the left of
Figure 12-8 you can see that there is a button; Push Down
Action that will convert the action to a strip as shown on the
right of Figure 12-8, press that now for the climbing_baked_ik
action.

242
Figure 12-8 The Nonlinear Animation Editor

9. The bright orange block can now be moved, split, duplicated

and generally used for building more complicated animations.
The idea is that other animations can be overlaid on the basic
one. As an example, you could make a short animation of just
the neck and head to make Lola look over her shoulder for a
fraction of a second and if you placed that strip above the
climbing it would override the default climbing animation, just
for those bones. If there were face bones, or shape keys
available you could then make Lola’s face either smile, or
grimace in horror, as a third layer of composition. The potential,
and complexity, of this animation editing facility is enormous
but, for now, you simply need to duplicate the existing strip
twice using the RMB context menu. The copies will be placed
in a new track above the existing one so drag them down so
they are all in a line and delete the newly created, but now
empty, tracks by clicking the track name on the left and pressing
DELETE.

Animating Parent Objects

10. If you now change to the Timeline view and increase the End
frame so you can see all three strips playing you can see that
you still need to correct the start position for the second and
third strips. The tidiest way to do this is to have one empty
object that can be moved around the scene to provide the start
location for the animation, and a second empty that will move
up by the correct distance at the end of the first and second
cycles. Create two empty objects and rename them lola_start
and cycle_start. Select first the armature object then the
cycle_start object and hit Object4Parent4Object, then, in a
similar way parent the cycle_start object to the lola_start object.

243
 11. In the Timeline move to the last frame of the first animation
cycle and with the cycle_start object selected add a keyframe
for its location by pressing I with the cursor over the 3D
Viewport. Use the orthographic view from the side and zoom in
so Lola’s head is about half the height of the screen, then pan
the view so the tip of Lola’s nose if exactly in line with the top
edge of the window. Move forward one frame then move the
cycle_start in the Z direction until Lola’s nose is in the exact
spot it was at the end of the previous cycle. Add a location
keyframe for the cycle_start here. Repeat the process at the end
of the second cycle, adding a keyframe in the same location,
and on the next frame a location that keeps her nose in the
correct place. Now you can move the lola_start empty to
different locations on the castle walls and she will climb there.

Simulating a Rope
Now that the climbing animation is sorted it’s time to construct the
rest of the set. Having a rope dangling from Lola’s waist will add a
dynamic element so do that first.

1. Add a plane to the scene and rename it rope, then scale and
move it so it’s where the floor would be behind Lola’s back,
you can judge the size from Figure 12-9. In Edit mode
subdivide the plane four times then, with edge selection
enabled, delete rows of edges so you are left with a serpentine
shape. Move four or five vertices so the end vertex is just
behind Lola’s waist. Finally create a vertex group containing
just the end vertex called “mixamorig:Hips”, this has to match
the bone name in the armature but you can also use it as the pin
group in the cloth.

244
Figure 12-9 The Rope

2. The rope object needs to be parented to the armature object as

an object, you don’t want the bone weights to be generated.
Then you need to add modifiers in the order; 1 Armature,
targeting the armature object. 2 Subdivision, set to Simple and
Levels to 2. 3 Cloth. 4 Skin with smooth shading. Reduce the
skin to a realistic thickness by selecting all the vertices in Edit
mode and using Mesh4Transform4Skin Resize. In the
Physics Properties tab open the Shape section of the cloth
simulation and set the Pin Group to mixamorig:Hips. Check
that the simulation runs correctly, you should also change the
shading to smooth and set the material to a more rope-like color.
Having a one dimensional rope like this is quick to simulate but
difficult to get collisions working well, but for the purposes of
this short video it will be fine.

3. To make the castle wall, enable the Extra Objects add-on then
hit Add Mesh Extras Wall Factory, then move and scale the
section of wall to match Lola’s climbing. Add two array
modifiers to the wall object, one repeating in the X direction
and the other repeating in the Z direction. You might need to
tweak the factors from exactly 1.0 or -1.0 so you don’t have
gaps. To make the blocks expand to touch each other without
leaving cracks you can select all vertices in Edit mode then
Mesh4Transform4Shrink/Fatten. You will probably want
to add a bevel modifier and create a suitable stone material. To
avoid distortion of textures on the end of the blocks you should
smart unwrap the UV map and feed that as an input to the
texture nodes.

245
4. Use an environment texture for the background and adjust the
strength to give it night-time lighting. I used the eevee engine
for speed but set ambient occlusion and motion blur on. In
retrospect the motion blur was a mistake because of the jump
frame of the cycle_start empty object. I also reduce the frame
size to 50 percent which rendered each frame in about one and a
half seconds on this desktop computer with a reasonable, if old,
graphics card. To add a little more realism I also added a tiny bit
of noise to the camera X and Y rotation, for a hand-held feel.
Render two or three videos of Lola climbing quite high on the
wall, as image sequences that you can mix later.

Animating a Second Armature

5. The final product will include the Peasant Man character

climbing up the wall after Lola so make sure the lola_start
object is located reasonably up the wall then import the .fbx as
you did for Lola. Open the Dope Sheet in Action Editor mode
and rename the convolutedly named animation to
salsa_dancing. From the Browse Action to be linked drop-down
select climbing_baked_ik. To make your life easier, at this
point, also rename the armatures in the Outliner Editor to
lola_armature and man_armature.

6. Change to the Nonlinear Animation view and follow the same

process you used for Lola. Push down the climbing_baked_ik
action to create a NlaTrack then duplicate the strip twice, drag
them in line and delete the empty tracks to keep it tidy. You also
need to create two empties, as you did for Lola and parent the
man_armature to the man_cycle_start empty and parent
man_cycle_start to the man_start empty. Follow the same
procedure to animate the movement of the man_cycle_start
empty with four keyframes so the animation runs without
discontinuity for three cycles.

7. Record some video of the man climbing. It would be nice to

attach the bottom end of the rope to his Hips bone but the
simulation with collisions becomes rather slow so it’s probably
sufficient to increase the Air Viscosity in the rope Physics
Properties tab, to stop it swinging around so much.

8. When you have enough footage of climbing, in the Dope Sheet

change the animation for both armatures to salsa_dancing. In
order to stop the characters jumping upwards part way through
the dance you also need to disable the animation for both the

246
 cycle_start empty objects. Do that in the Dope Sheet either by
clicking the X next to Browse Action to be linked or deselect
the check-box on the channels list. In the Timeline change the
end frame to match the length of the dance then move both
characters to the top of the wall and record some video of them
dancing.

9. When you have all the footage you need, create your final
masterpiece using the Video Sequencer as shown in
Figure 12-10. There are many controls available here, and it’s
hard to do more than scratch the surface in this exercise,
however two things may be useful. The first one you may want
to use is the fade effect which you can can control using the
right Sidebar in the Compositing section. You can set keyframes
for the Opacity in the same way you have used to animate other
Blender values. The second one is to slow down a strip, or part
of a strip. For this you need to click Add Effect Strip Speed
Control from the area menu, again, there are many controls
available in the Sidebar.

Figure 12-10 The Video Sequencer

Before closing the chapter there is one job that belongs here. When
you used rigid body simulation in Chapter 11 to bounce balls off each
other as they rolled around a bowl, I mentioned that you would use
them later in the book, and that the balls would turn out to be bugs that
would uncurl themselves and scuttle back up the sides. Well now would
be a good time to create the meshes, armatures and animations for
those bugs.

247
 Creating and Animating a Bug

This will be another opportunity to do some bread-and-butter mesh

work, building on the lessons learned when you made the furniture in
Chapter 8 and the shark in Chapter 10. This time you will make most of
the geometry by extruding faces of the start cube. As the bug will be
symmetrical you can use some shortcuts to build the armature and you
will animate the legs using the Follow Path object constraint.

1. Start a new Blender scene, saving your previous work, but leave
the start cube as the basis for the bug mesh, and rename it bug.
The balls in the simulation were radius 0.2 so the bug should be
about 0.6 long, 0.2 high and 0.4 wide. You can see the
proportions in Figure 12-11 stage 2, and when curled up in stage
3. With the cube selected enter Edit mode and scale the vertices
down to the size of the middle leg section. To help with the
animation later, design the bug so it is walking on a surface
running through the origin as shown in Figure 12-12. Extrude
and scale sections in the Y direction for other body sections, the
head and tail. Once you have enough divisions along the length
of the bug use mesh bisect as you did in Chapter 10 then add a
mirror and a subdivision modifier and set shading to smooth.

Figure 12-11 Crafting and Rigging the Bug

2. Extrude, scale, move and rotate the faces from the sides of the
bug to form the legs and add extra edge loops to help define the
shape and sharpen edges as needed. The modeling doesn’t need
to be too detailed but, as the bug will be curled into a ball,

248
 involving large deformations, you should make sure that the
visible parts are composed of quads. When it looks sufficiently
bug-like, switch back to Object mode and apply the two
modifiers.

3. Because the bug is symmetrical you only need to construct one

half of it. Start by adding a new armature object, set it to display
in front of the mesh, then switch to Edit mode in orthographic
view from the X direction, and scale the initial bone down so
it’s about a third of the height of the bug, as visible in Figure
12-11 stage 2. Extrude this forwards four times and backwards
three times to make the spine – if that’s the right term for an
invertebrate. In orthographic view from above extrude three
bones from matching nodes on the spine to the points where
each leg joins the carapace, renaming each bone “leg1.001”,
“leg2.001” and “leg3.001”. Now, when you extrude these bones
the suffix will increment, and if you set snapping to volume,
you can position the bones to the center of the legs as you
extrude them. Remember to turn snapping off afterwards
otherwise you will run into problems when you try to move
control points or objects later.

Symmetrize an Armature

4. Once you have three legs rigged you can quickly generate the
mirror image but first you need to append “.L” on the name of
each leg bone, and there is a handy tool for that too. Select all
the leg bones then from the menu run
Armature4Names4Auto-Name Left/Right. Now, with all
the bones selected, click Armature4Symmetrize to generate a
symmetrical rig with the left sided bones named with an L
suffix and the right sided bones an R suffix. Finally generate the
rig as you did previously with Parent4With Automatic
Weights

5. Before you set up the scuttling animation, create the pose with
the armature curled up, changing to the unposed state as an
uncurling animation. With the armature selected and in Pose
mode, view the bug from the side in orthographic view. Move
the spine bones to make the mesh as spherical as possible then
view from different angles and fold the legs in. In Figure 12-11
stage 3 you can see the two orthographic views, and also that I
added a rigid body with a spherical collider to help position the
bug inside, though you will need to remove the rigid body when

249
 you run the animation as it will make the mesh fall off the
armature! Select all the bones then add a keyframe for the rolled
up pose at frame 1, move to frame 30 and select Pose4Clear
Transform4All and add a keyframe for all bones in the flat
pose. Check that the animation runs as expected then, with the
Dope Sheet set to the Action Editor, rename the animation to
uncurl and push it down onto the NLA stack.

The Follow Path Constraint

6. You will now create a second strip to make the legs move, so
click on + NEW to create the action and rename it scuttle.
When you move the frame pointer you will find that the
NlaTrack for uncurl is still being followed because there is no
action yet set above it in the stack. Move to frame 1, in Pose
mode select all the bones, clear the transforms and add a
keyframe for the unposed state. To set up leg movement, switch
back to Object mode and add a Bezier circle object then, in Edit
mode scale, rotate and move it to the path that the end of the
front left leg should follow, as shown in Figure 12-12. Add an
empty and reduce its size in the Object Data Properties tab so it
matches the scale of the bug, then add an object constraint
Follow Path setting the Target to the Bezier circle you just
added. Click the button Animate Path to automatically create
the modifier in the curve. You now just need to select the
armature and, in Pose mode select the end bone of the front left
leg. Add an inverse kinematic bone constraint setting the Chain
Length to 2 and the Target to the empty.

Figure 12-12 The Bezier Circle to Control an IK Constraint

250
 7. Check that the leg follows the curve without any awkward
bending, you will probably need to adjust the position and
shape of the curve. There is a 50 percent chance that the
animation will run in the wrong direction, if that’s the case, with
the curve in Edit mode, select all the control points and run
Segments4Switch Direction. The default speed of the
animation is also rather slow so select the curve object and open
the Graph Editor then in the Sidebar on the right open the
Modifier tab. Increase the Coefficient for X^1 to 12.0 which
should make the legs spin round a bit more frantically. Once the
leg is behaving properly make sure you have renamed both the
curve and empty to something related the the front left leg, such
as BezierCircle.L.001 and Empty.L.001 then duplicate them and
position them to drive the other legs. Remember to change the
L.004 to R.001 when you duplicate the first leg and empty for
the opposite side. In order to make the legs move out out of
phase, so that there are always three points of contact on the
ground, for each empty change the Offset value in the follow
path constraint. You need to get the timing right so the lifting
leg doesn’t hit the leg that’s about to land, though the legs will
be whizzing round, and the bugs will be small enough, that it
will be hard to see details on the final video.

8. Once you are happy with the animation, it’s time to bake the
constraints to actual bone rotations, but first save the blend file
just in case. It will be much harder to adjust once it’s been
baked, so don’t tick the overwrite option when baking. When
you’re ready, select the armature, and in Pose mode, with all the
bones selected, run Pose4Animation4Bake Action and tick
just the three check-boxes; Only Selected Bones, Visual Keying
and Clear Constraints. In the Dope Editor rename the action to
scuttle_baked_ik and push it down as a new NLA strip.

9. To check it all worked correctly change to Nonlinear Animation

view, select the scuttle_baked_ik track by clicking on it and
drag it so it starts at the end of the uncurl strip. When you play
the animation the bug should uncurl and then scuttle, the
empties will still go round and round the curves but they no
longer control anything. You could delete them but I suggest
you keep them until you have the whole sequence in the can.
Save your file!
There is one more step needed for the bugs in the bowl exercise,
which is to make the bugs follow paths as they scuttle up the sides.

251
 However, as that will be part of the final rendering process, you will
implement it in Chapter 13, when you work on cartoon style materials.

Conclusion

You’ve covered a lot of ground again in this chapter. Building an

armature by hand, from scratch, has given you a little more insight into
how they work and, hopefully, will enable you to spot problems when
they occur. You also saw how you could use armatures to bring your
machines to life by making wheels rotate, levers swing and pistons
slide, and how bone constraints simplify so much of the difficulties of
posing and animating armatures.

Fitting an armature to a mesh is a requirement for both soft bodies

and hard machines, so being able to do this rigging quickly and easily
has enormous benefits to your productivity. The exercise with the
imported .obj file of the angel showed you how this can be done and it
also introduced you to posing with the rigify system. As well as rigging
and posing, you used object constraints as yet another way to
synchronize different parts of your scene and automate parts that would
be difficult to construct manually.

The castle wall exercise took some existing animations that were
imported to Blender as part of an .fbx file and showed you how you
could modify and re-apply them in different ways to different objects.
You used four different representations of animation keyframes;
Timeline, Graph Editor, Dope Sheet and Nonlinear Animation, all with
slightly different applications. To start with you will need to keep
referring back to your notes to remember what functionality is where,
but in time you will learn your way around them and value their
flexibility.

In the next chapter you will concentrate on non-photorealistic

rendering and the compositor. You will add the finishing touches to the
bug animation and learn about outlines, cell and hatched shading. The
compositor tends to be undervalued by people starting to use Blender
but it allows a great deal of different effects and optimizations that
would be impossible to do otherwise.

252
 13
U N R E A L I S T I C A S P I R AT I O N S

Creating non-photorealistic
materials and the many uses of the
c o m p o s i t o r.

Using 3D models to create cartoon style images gives a

massive productivity boost to artists compared with
manual methods, but achieving a reproducible, hand-
drawn, effect turns out to be quite hard to do. Fortunately
there are several features of Blender that help the
process.

253
 In this chapter you will use a few new techniques of combining
nodes in the Shader Editor to create “cell” shading, which is the term
used to describe blocks of uniform color applied as if by a simple
printing process or by a color wash. It is used in comic books and many
animations, including manga and anime. If you search online for help
on this subject you will find a plethora of clever systems, most of
which are more complicated than the ones you will use here. However I
suggest you start simple and get an understanding of what’s going on,
you can then add extra features as and when you find they are
necessary. You will then look at ways to add hatching such as might be
used for pencil or ink drawing, and also, the grease pencil line art
system that Blender has for adding outlines.

The second half of the chapter will cover some of the uses of the
compositor. You will see how different parts of the scene can be
rendered onto different view layers, and how the layers can then be
recombined. The use of nodes in the compositor has many overlaps
with their application in material shaders, but you will briefly sample
some of the filter and distortion nodes and see how they can be
introduced into the data flow. Links to the files and videos for the
exercises in this chapter are at get-into-blender.com 13.UNREALISTIC
ASPIRATIONS

Cell Shading

The first method of cell shading described here is essentially to

take the output of the standard principled BSDF shader material and
feed it through a color ramp converter to create a hard division between
two areas of solid color. It will be easier to see how this works using a
simple example.

Color Ramp

1. The shading effect will be more obvious on a rounded shape so

start a new file in Blender and add a subdivision modifier to the
start cube with two levels of division.

2. To set up a very simple arrangement of material nodes as shown

in Figure 13-1, increase the size of the bottom window area and
switch it from Timeline to Shader Editor. Change the name of
the material from Material to comic. To see the results of
changes to the node arrangements you will need to change the

254
 Viewport Shading to Rendered with the Render Engine set to
Eevee.

3. The default material has just two nodes, the shader and the
material output, so move them further apart to leave room for
three more nodes to be inserted between them. Now select the
following menu options placing each node onto the connection
between the shader and the output node;
Add4Converter4Shader to RGB,
Add4Converter4Separate Color,
Add4Converter4ColorRamp and
Add4Converter4Combine Color The nodes to separate and
recombine the color should be set to HSV with Hue and
Saturation connected directly from one to the other but the
Value output passed through the color ramp node. Set the ramp
interpolation drop-down to Constant and adjust the slider to
make the cut-off between black and white part way up the side
of the rounded cube.

4. You can probably see how this method could easily be extended
to having three or four divisions rather than this simple black
and white one. Try adding another three stops to the color ramp
node and edit the base color of the BSDF shader to different
brightnesses of yellow, say.

Figure 13-1Simple Cell shading using the color ramp node

This simple color ramp method is perfect for many situations, and
might produce quite satisfactory results if used for the bug uncurling
animation that you will tackle later. However there are disadvantages
that arise from the inability of the color ramp node to take variable
inputs for the colors and positions of its stops, which means that each
color ramp needs to be adjusted individually for each material and
lighting conditions. This arrangement also produces blocks of uniform
color that tend to blot out surface detail, apart from in the area where
one cell changes to another.

255
 The Map Range and Mix RGB Nodes
Node groups tidy up the layout, so when you make complicated
materials using lots of nodes, there is an obvious visual improvement.
However the real power of groups is that, if designed properly, they can
be reused many times. This will be one of the real benefit of switching
from the color ramp node to the map range and the mix RGB nodes, as
they can both use variable input values.

1. Edit your existing material by deleting the color ramp and the
combine color nodes then add two more to replace them by
selecting Add4Converter4Map Range and
Add4Color4MixRGB. The output of the map range node is
used for the factor input of the mix node. You need to set the
From Min and From Max values very close together to give a
sharp cut-off at the edge of the cells. In the mix node change the
colors to two shades of yellow, putting the darker one above the
lighter one.

2. The first, obvious, deficiency of this system over the color ramp
node, is that there isn’t an obvious way to insert additional stops
with different shades. One solution is to change the
interpolation of the map node from Linear to Stepped Linear
when you do that you will see an extra input value; Steps,
which you can change to 3.0 to give intermediate blended
colors. You will need to change the From Min and From Max
values to make the cell divisions fall more evenly.

3. The advantage of this technique over the simple color ramp is

when you want to build up layers of shading, such as for the
cross-hatching you will use later. For those situations the more
general solution, using grouped nodes will be required, but in
order to see the potential of this technique, quickly try moving
the color input out of the mix node. Create a new input by
selecting Add4Input4RGB, place the node to the left of the
shader and feed its output into the Base Color input. Change the
color of the new input RGB node and use the eye-dropper to set
it equal to the lighter value in the mix node. Now drag another
two connections from the input RGB node to both color inputs
of the mix node. You need to darken the Color1 input by adding
a node with Add4Color4Bright Contrast, dropping that
onto the connecting pipe, and reducing the Bright value. Finally
add a magic texture and a bump node to set the normal of the
shader. Reduce the Distance value of the bumps to 0.05 to give

256
 a more subtle effect. The full node layout and the resultant
material are shown in Figure 13-2.

Figure 13-2 More complicated Cell shading using map range and mix RGB nodes

As you tried to fine-tune the settings you probably found that it

was difficult to get the correct range of colors, especially where the
roughness was low, as highlights or reflections are typically a different
color from the bulk material. The neatest solution to this problem is to
use node groups.

Node Groups
As mentioned above, groups are a way to organize nodes but, by
doing so, they help you to think through the logic of the material as you
build it. For instance, in the cell shader you are making now, it would
make sense to have a node group that took inputs of: First the threshold
value for the edge of this boundary between cells. Second the color to
use use where the material is darker than the boundary. Third the color
to use if lighter than the boundary. These nodes could then be stacked
one after the other, each taking the result of the previous node for the
darker color and introducing a lighter color for the other. In this way
each group is equivalent to a stop of the color ramp node, the
arrangement is shown in of Figure 13-3 where the three node groups
are arranged on the right side of the image.

257
Figure 13-3 Using multiple node groups to produce a cell shader.

4. Select the map range node as well as the mix RGB node then
from the menu Add4Group4Make Group. The original
network of nodes has now been replaced by the contents of the
new group with the additions of an input and an output node.
On the right, in the Sidebar, you can alter properties of the
group such as adding new inputs, renaming them and setting
default or limiting values.

5. For the general purpose version of this group you need to add
an input to control the cut-off point. In the sidebar select the
Node tab then add a new input by clicking on the relevant +.
Change the name of the new input to Cut-off, the default to 0.5,
Min to 0.0 and Max to 1.0. Connect the Cut-off input to the
From Max input of the map range node, set the From Min to
0.0 and the Steps to 1.0. The layout of nodes within the group is
shown in Figure 13-4.

Figure 13-4 The Nodes inside the cell_shader_stop node group.

6. To finish editing the group node click on the button on right of

the header with an icon of an upward pointing arrow. To return
to editing the contents of a group you click on the connected
nodes icon in the top right corner of the node, which you can
just make out in Figure 13-3. When you select on a node that is

258
 a group, if you open the Properties section of the Node tab in
the Sidebar, you can change the name to something more
meaningful, in this case call it cell_shader_stop. In the Node
section above you can also change the unique name for each
instance of the group, as well as the label to be displayed at the
top of the node. Using meaningful names will make your life
much easier in the long run. Duplicate the prototype node twice
using SHIFT-D and connect them up as in Figure 13-3
There is much sophistication you could add to this basic node
group. For instance, if you wanted to give the cell areas slightly
blended edges, as if created using a water color wash, you could
change the interpolation method to linear and calculate the From Min
value as a fraction of the Cut-off using a multiply node. For the
moment, however, there is sufficient control and complexity for the
current exercise, so we will move on to look at some inking techniques.

Hatched Shading

There are two more mechanisms to investigate before finishing off

the bugs-in-a-bowl animation, the first is how to add hatched shading
to a material. In the node group you just developed you mixed the two
colors in a fairly intuitive, pro-rata way. The effect was the equivalent
to changing the transparency of the top color so it blended with the one
underneath. However for hatched shading you need the black lines of
each layer to accumulate on top of each other, and this can be achieved
by changing the mode of the mix node to darken.

1. Make a duplicate of the cell_shader_stop group node and drag it

down below the other three. To the right of the drop-down and
name of the node group there is a number and, to the right of
that, a shield icon. The number represents the occasions that the
node group is being used anywhere in the current blend file,
which should be 4. Click on the number to make a new copy of
the node group, and to allow you to edit it without altering the
three already being used in the cell shader. Rename the node
group from cell_shader_stop.001 to hatched_shader_stop.

259
 USERS IN BLENDER

Many components, including materials, image textures, animation actions and

node groups can be used in multiple locations in a blend file. Bender keeps track of
which components are required by other objects, and shows the number of users at
the side of each component wherever it is listed. When you save a blend file, Blender
strips out unused clutter to save disk space and improve running efficiency when you
reopen it. More often than not, this is just what you want, and you don’t have to think
about it, but sometimes you need to keep unused resources available. One example
of this would be where you created a blend file as a library of animations, materials or
node groups that you could append into other projects when you have a requirement.

The way that Blender allows you to control the users, and consequently whether
the resource will be saved, is through the series of buttons beside the resource
name. The first button shows the number of users, if there are more than zero, and
clicking on this will create a new instance of the original. The second button has a
shield icon and this will create a fake user for the resource to prevent it from being
discarded when you save the file. In some locations there are additional buttons such
as one showing two overlapping squares which will create a new resource, a button
with a folder icon for browsing files to load or one with an X icon for deleting the
resource.

2. The hatched shader will operate in a similar way to the cell

shader, it will take an input of background color and it will
overwrite it with a hatched texture wherever the value of the
color output by the principled shader is below a cut-off value.
Click on the top right corner of the hatched_shader_stop to edit
its contents and change the mix node from Mix to Darken. In
order for the hatched texture to be blended at darker values, you
need to invert the output, on the map range node change the To
Min value to 1.0 and To Max value to 0.0.

3. The hatching will be provided by a noise texture node but in

order to build up layers of shading on top of each other you
need to rotate the direction of stretching. This means that you
won’t need a second color input but you will need one for the
rotation angle. In the Sidebar Group tab remove the Color2
input and add one called Rotation with Type Float, Default 0.0,
Min -pi and Max pi.

260
NOTE Within Blender, angles are held as radians but in some entry fields,
such as the Object Properties tab, the entry and display is provided
in degrees. For those situations you can specify which format you
are using by following a number with the letter D or R. It is also
possible to enter formulae into any value field in Blender so long
as they are legal python expressions. You could have typed
radians(-180) and radians(180) and Blender would have
calculated the range using the python radians function.

4. In order to construct the hatched texture you need to add some

more nodes: UV map, vector mapping, noise texture and a color
ramp, all of which you’ve used before several times. You need
to use the rotation as the Z value of an XYZ vector as the vector
mapping node requires a vector input, add this by selecting Add
Converter Combine XYZ. To simplify the vector rotation
operation, set the Type field of the vector mapping node to
Texture. The node layout for the group is shown in Figure 13-5.

Figure 13-5 The Nodes inside the hatched_shader_stop node group.

5. Close the edit of the node group by returning to the parent node
tree and position the new hatched shader to take the output of
the last cell shader as well as the value output from the original
shader. Adjust the rotation and the cut-off to give a pleasing
result.
At this point you could further enhance the hatching by, for
instance, allowing the thickness of the lines to be controlled with
another map range node instead of the color ramp. Additionally the
rotation of the lines could be made relative to the curvature of the
surface, by using the normal output from a texture coordinates node.
But for the moment leave the material as it is and save the blend file so

261
 you will be able to append the node groups into the bug animation
project.

Line Art

At the start of the previous section I alluded to two mechanisms,

and the second one is Blender’s Grease Pencil Line Art. This is a
convenient way to add lines to specified edges by creating grease
pencil objects with various modifiers. In previous versions of Blender
this would have been done using freestyle lines, but they have the
disadvantage of being harder to control and can’t be previewed on the
3D Viewport, only appearing when the scene is rendered. If you are
constrained to an old version of Blender then the previous version of
this section, using freestyle lines, can be viewed in Appendix A.

1. First of all you need to add a new Grease Pencil object with a
Line Art modifier. In the file you’ve been working on for the
cell shading shading exercise, make sure that the rounded cube
is selected then in Object Mode select Add4Grease
Pencil4Object Line Art and a new object called LineArt will
appear in the Outliner. In the camera view the line runs round
the outside of the Cube object but from other view positions
you can see that it’s a 3D object stuck to to the rounded cube’s
surface.

2. To change the line style and thickness, select the LineArt object
and open the Modifier Properties tab where you can see the
modifier that was generated automatically. Try adjusting the
Line Thickness to different values then set it back to 25.

3. In the modifier expand the Edge Types section and you will see
that the default is to draw lines around contours but none are
apparent on the bumpy surface. This is because the surface
lumps are only being created by adding a normal texture rather
than deforming the mesh surface. To fix that, select the Cube
again then open the Shader Editor and delete the magic texture
and bump nodes feeding into the principled shader. Open the
Modifier Properties tab and increase the subdivision levels to 4
then add a displace modifier and click + New to make a new
texture to control the distance. The default new texture will be
called Texture so open the Texture Properties tab and change the
Type to Magic then, back in the Modifier Properties tab change
the displacement Strength to 0.2.

262
 4. You should see that some lines have appeared around the new
bumps. However the lines are rather thick and have ugly square
ends. To improve this situation we can use a different Grease
Pencil object for the lines drawn inside the object, away from its
outline, with a smaller thickness and with tapered ends. Select
the LineArt object and rename it LineArt_silhouette, then in the
Modifier Properties tab change the Edge Types to Silhouette in
the first drop-down, and un-tick the other options.

5. Add a new line art object using Add4Grease Pencil4Object

Line Art as above, and rename it LineArt_contour. If you didn’t
first select the Cube object you will need to first change the
Object value in the Modifier Properties tab to Cube. Add a
second modifier with type Thickness, tick the Uniform
Thickness check-box and reduce the Thickness to 15. Expand
the Influence section, tick the Custom Curve check-box and
create a profile that goes from small to big to small again. The
settings and resultant lines can be seen in Figure 13-6

Figure 13-6 Line Art settings.

As you added the thickness modifier you saw that there are
multiple features of line drawing that can be controlled by stacking
multiple modifiers. The permutations are enormous and, if you are
interested in using Blender for graphic art, you will find many amazing
works online to inspire you. I will leave you to explore this topic
further on your own, for the moment just save the blend file so you can
append the material for the bugs in the next section.

263
Assembling the Project Elements

At last it’s time to put together the physics, animation and

materials to create a cartoon style video of the bugs rolling down then
climbing out of the bowl. The first job that needs to be done is to
substitute the bug mesh and armature from Chapter 12 for the simple
spheres used in Chapter 11.

1. Open the the last blend file from Chapter 11 which should be
the physics simulation of the balls rolling around the bowl.
There are three ways you could approach switching the bugs for
the balls. The first would be to snap the cursor to the location of
each ball, appending the bug, snap its location to the cursor, add
a rigid body and constraint to the bug then remove the original
balls and run the simulation again. The next logical approach
would be to keep the existing simulation but hide the balls and
add an object constraint for each bug to copy the location and
rotation of the corresponding ball. The third option, and the one
I will guide you through here, is to bake the physics of each ball
to a keyframe animation and then, after appending the five
bugs, add the corresponding animation in the dope sheet action
editor. Before you bake the animation you probably need to do a
bit more fine-tuning. Ideally you want all the bugs to have
stopped rolling by the time the uncurling animation starts so
you might need to increase the friction of all the surfaces in the
Physics Properties tab. I added keyframes to the friction values
so they increased from 0.1 to 1.0 between frame 149 and 150. I
also increased the Rigid Body World, Cache to 350 frames in
the Scene Properties tab. After the friction changes you will
probably have to re-aim the heavy ball, try using the Graph
Editor to tweak the location of the ball at the second keyframe.

2. When you have the physics running well and the balls nearly
stationary by the end of the animation, select the five bug balls
and the heavy ball, then from the menu run Object4Rigid
Body4Bake to Keyframes. Save the blend file now to the
folder used for this chapter. The baking process removes the
rigid bodies from the objects involved, however there are still
some hanging animated properties for the removed constraints
which might cause Blender to crash if you trigger certain
processes. Make the Timeline Editor pane larger and expand the
left hand outliner. If you see any entries underlined in red then
delete them now, and then delete the four empty objects. In the

264
 Outliner Editor, toggle the visibility and rendering off for all the
bug.00x balls.

3. Append the bug object from the last exercise of Chapter 12.
Even though the object you import is a mesh, the armature,
animations and materials used by it will also be brought into
this file. To make Blender help with renaming as you duplicate
the bug, rename the armature to bug_armature.001 and the mesh
object that it parents, to bug_mesh.001. Now select the armature
and the mesh and duplicate them four times, you should end up
with sequential suffixes.

4. Select bug_armature.001 then open the area at the bottom of the

screen as Dope Sheet with the Action Editor context selected.
From the drop-down for choosing actions select bug.001Action
that was generated by baking the keyframes for the first ball.
Switch to the Nonlinear Animation and scroll down to find the
tracks for bug_armature.001, move the two imported tracks;
uncurl and scuttle, to the right of the baked rolling ball
animation. Repeat this procedure for the other four
bug_armatures. The order of the tracks for the uncurl and and
scuttle animations is important as the positions of bones in a
higher track will override a lower one. If the bugs roll down the
bowl uncurled or don’t scuttle after uncurling move both the
actions into one strip. The alternative is to toggle a track off
while recording the animation before it, then toggle it back on
to record the remaining animation.

5. When you run the animation you should see the bugs replicating
the collisions and rolling just as the balls did. You may find that
the bug diameters are not quite the same as the rigid body
colliders and the bugs either float above, or sink into, the
surface of the bowl. It would be quite easy to move the
positions of the animation keyframes up or down using the
Graph Editor, however it is even easier to adjust the Z location
of the bowl object. Getting the correct height will be easier if
you use Viewport Shading as Rendered and check the position
of the cast shadows.
Putting together the full animation in one unbroken shot would be
pretty cool but it will be much easier to assemble the video from
several components. To achieve that you will first set some keyframes
for the camera location and rotation so it follows the action and finally
zooms in close, as the bugs come to rest. You then need to set the
materials for the bugs, ball and bowl to produce anime style rendering,

265
and subsequently render a sequence of images for the first part of the
video.

For the next stage you want the bugs to start in frame 1 in the same
locations and rotations that they finished the previous animation. You
then want the bugs to play the uncurl action at the same time as they
transition from whatever rotation they end up in, to the right one for the
scuttling sequence. For the scuttling you will add five Bézier curves
drawn onto the surface of the bowl and set an object constraint on each
bug to follow a curve.

6. Select the camera and set the current frame to 1 then use walk
navigation to position it at the start and insert a keyframe for
position and rotation. Move forward through the animation so
you get a good balanced view of the action without dramatic
camera movements that will give the viewer sea-sickness. Add
enough keyframes to prevent the camera swinging the wrong
way but not too many that you have sudden transitions. By the
final frame you want the bugs fairly filling the frame.

7. Append the material named comic from the blend file you
worked on at the start of this chapter. You could also append
two LineArt objects but there are several modifications needed
so it simpler to create two new Grease Pencil objects as
described below. Select each bug mesh and set the material to
the appended comic. For the bowl and the heavy ball, also select
the comic material but then click on the overlapping squares
button just to the right, to create a new material for each object.
You can now change the base color, roughness, metallic and
other properties of these two objects. I used a magic texture for
the base color of the bowl but set the cell input color as blue, as
you can see in Figure13-7.

8. In the Outliner select the default collection called Collection

then in the 3D Viewport add two line art objects with
Add4Grease Pencil4Collection Line Art, rename them and
adjust the modifier properties as before but, for
LineArt_silhouette set the Edge Types to Individual Silhouette.
Adjust the Line Thickness values and do a few trial renders to
make sure everything works as expected, then render the whole
animation as an image sequence into a sub-folder for this
exercise. Depending on the power of your computer you might
want to reduce the resolution to 50% in the Output Properties
tab, it could take quite a while to render 350 frames.

266
 9. When you have the required shot in the can, save the blend file
with a different name. For projects like this where something
might go wrong or get accidentally deleted, it makes a lot of
sense to do lots of incremental saves so you can go back to a
working restore point. Put the current frame marker on the last
frame, then in the Nonlinear Animation area, for each bug, push
down the baked physics animation onto the stack. Mute the
physics animation track and un-mute the uncurl and scuttle
tracks. Now when you run the animation the bugs should stay
where they are but uncurl and scuttle their legs.

10. You need to add keyframes for the location and rotation of each
bug armature at frame 1 as it is now, and at frame 30 rotated
onto its stomach, disentangled from any adjacent bugs and
pointing outwards from the group. The two keyframes are
shown in Figure 13-7.

Figure 13-7 Keyframes for curled and uncurled locations and rotations.

11. The bugs each need a path to follow. Stay on frame 30 and
check which bug is number 001 then, in Object mode add a
Bézier curve, rename it bug_track.001, switch to Edit mode,
delete all the vertices then select the Draw tool. In the Sidebar,
Tool tab set the Depth option to Surface. Draw a curve starting
just in front of the bug, in a circuitous route up to the rim of the
bowl. Change to the selection tool, select the vertex nearest to
the bug and move it as close as possible to the bug’s origin; the
yellow dot visible when the bug is selected. Adjust the handles
of the vertices near the bug so the curve is in line with the
direction the bug is pointing.

267
12. Now you have a line, the bug needs an object constraint to make
it follow it, but only after frame 30. Return to Object mode,
select the bug_armature.001and, in the Object Constraint
Properties tab, add Follow Path. Set the Target as
bug_track.001, the Forward Axis as -Y, tick the Follow Curve
check-box and click Animate Path. Change the Influence value
to 0.0 and set a keyframe for this at frame 30. Move on one
frame to 31, change the Influence value to 1.0 and set a second
keyframe here.

13. If you run the animation you should see the bug suddenly jump,
at frame 31 to an offset position part way along the curve. There
are two corrections needed; first, delay the start of the path
animation so it commences at the end vertex on frame 31,
second, the location and rotation of the bug_armature need to all
be set to 0.0 between frame 30 and 31. Select the bug_track.001
curve object then change the bottom window area to the Graph
Editor. Open the Sidebar and select the Modifiers tab where you
can see the details for the generator function created when you
clicked on the animate path button. The function is a straight
line starting from -1 with a slope of 1.0, so by frame 31 it has
moved 30 percent of the way along the path. Fix this by halving
the slope, change the x^1 value to 0.5, and delay the start by
changing the Coefficient to -15. The bug will now start moving
at frame 30 and get to the end of the line at frame 230.

14. To make make the bug actually stay on the track instead of
being offset, move to frame 30, select the bug_armature.001
object and open the Object Properties tab. There should already
be a keyframe here giving a yellow background for the location
and rotation values. Advance on one frame to 31, change all the
location and rotation components to 0.0 and set a keyframe
here. If you use the shortcut, you will need to press I over one
of the location values as well as over one of the rotation values.

15. Watch the bug carefully all along its path. If you see the bug
floating above or sinking into the surface, or alternatively,
rotating to one side, then select the curve object and switch to
Edit mode. Carefully move the relevant vertices up or down and
rotate the bug by adjusting the tilt. For exercises like this you
will probably find the shortcut CTRL-T more convenient than
the Toolbar or menu options.

16. Add paths and constraints for the other four bugs following the
description above. When they’re all animating to your

268
 satisfaction you should set up the camera animation. In the
Timeline delete all the keyframes you set up for the rolling
animation, apart from the last one and move that to frame 1.
Run the animation on for fifty frames or so then position the
camera using walk navigation and set another keyframe. Try to
get some close-up views of the bugs as well as making sure to
see that there are lots of bugs all running around independently,
take care that the camera always moves smoothly.

17. Finally render the frames to a different sub-folder of this

chapter’s storage location. Save the blend file then open a new
project using File4New4Video Editing and import the two
sequences then render in video format. When you import the
images make sure that they are sorted by file name as Blender
will keep them in the order they appear in the file browser
window.

The Compositor

Many users of Blender never venture far into the intricacies of the
compositor, which is a shame because, not only does it allow many
enhancements to the finished render, but sometimes it can speed up the
process significantly.

To use the compositor beyond the simple application of filters,

requires the use of view layers which create different sources of image
information to composite together. It is possible to render a scene to
multiple view layers with each layer being used for different
collections of objects and collecting different information such as
lighting, depth and masks. One of the options that can be varied
between layers is the number of samples used for rendering with the
cycles engine, and this allows the background objects to be rendered at
relatively low resolution, but the central focus of the scene to use a
much higher sample rate. Different layers can have noise reduction,
blur, sharpen, glow, color correction and many other filters applied.

View Layers
The creation and switching of view layers is controlled using an
inconspicuous drop-down at the very top right of the Blender window.
By default there is just one view layer called ViewLayer and, very
often that will suffice, but in this exercise you will use two layers; a
background containing some glass objects and a foreground containing

269
 a large diamond. The problem with transparent and reflective objects is
that they can produce a lot of noise when rendered, which requires a
high number of samples to resolve. The denoiser in Blender version 3.0
onward is amazingly good, as you will appreciate if you look at Figure
13-8, but in some circumstances it can blur details, such as the
highlights on the top of the red cut-glass block. In this exercise the
foreground will be rendered with a high number of samples and the
background will use a low number of samples and denoised in the
compositor, then a small amount of the noisy render will be mixed back
in.

1. Start off by adding a few objects to the scene. I used the add-on;
extra objects, which was still enabled from Chapter 12, and has
a diamond mesh option, but a normal ico sphere would be fine.
Add a cube with beveled edges and some five sided cylinders as
table and background objects. Change all the materials to Glass
BSDF with low roughness and light tint, for the cube use a
procedural texture and bump node to create a slight normal
pattern. Finally in the world settings add an environment
texture. You can see in Figure 13-8 a detail from the scene I set
up, but the actual layout isn’t too critical.

Figure 13-8 Comparison of different sample numbers and effect of denoising.

2. Once you have the scene set up it’s time to create the two view
layers and configure some collections to use them appropriately.
At the top right of the window rename ViewLayer to

270
 background_layer, then create a new layer by clicking on the
button to the right with the overlapping squares, and rename the
new layer as foreground_layer.

3. You now have two layers but there will be no real indication of
how they effect anything. Open the Render Properties tab and
set Render Engine to Cycles, change Max Samples under the
Render section from 4096 to 50 and un-tick the Denoise check-
box. When you render an image you will see that the whole
scene is rendered twice and at the top of the Blender Render
pop-up window you can click on a drop-down to see the results
of rendering each layer. At the moment the two layers are the
same.

Collections

4. Divide the objects into two collections. At the moment

everything is in a collection called Collection, rename that to
background_objects. Right click on the Scene Collection and
add a new collection, rename it foreground_objects then drag
and drop just the diamond object into it.

5. At the top of the window select the background_layer as the

current view layer and un-tick the check-box next to the
foreground_objects collection. Then select the foreground_layer
as the view layer and un-tick the check-box against the
foreground objects. Now, when you render you will see a
difference between the render results, but it’s not really what
you want. The lighting is missing from the foreground render
and the background render has no shadows cast by the
foreground object, besides which, there is no composite result
recombining the two layers.

Holdout and Alpha Over

6. A neat solution to this problem is to use holdouts to enable the

renderer to skip over the pixels where objects in a given
collection should be. With the foreground_layer still selected as
the view layer tick the check-box next to the
background_objects collection to re-enable it, then right click
on that collection and select View Layer4Set Holdout. In
order to get reflections and shadows from the foreground object,
change the view layer to the background_layer and re-enable
the foreground_objects collection here too.

271
7. Before you render there is one more setting you need to make;
open the Render Properties tab and, near the bottom, expand the
Film section and tick the Transparent check-box. When you
render you should get the background with all the objects
shown but the foreground will just have the diamond with
everything else the gray checker pattern to indicate
transparency. You are in a position to reassemble the image in
the compositor.

Compositing

8. Enlarge the area at the bottom of the window, change the editor
type to Compositor then tick the Use Nodes check-box in the
header. The default is a single render layers input node
connected to a composite output node. Select the render layers
node and duplicate it then create a new node with
Add4Color4Alpha Over and drop it onto the connection
leading to the composite output node. Connect the image output
of the duplicated render layers node to the other input of the
alpha over node, then using the drop-down list at the bottom of
each node, select the background_layer for the input to the top
input of the alpha over node and the foreground_layer for the
bottom input.

9. In order for the composite result to be any different from the

one for the background layer, you need to increase the samples
when rendering the diamond. Make sure you have the
foreground_layer selected as the current view layer, then in the
View Layers Properties tab scroll down to the Override section
and change the value of Samples from 0 to 1000. You also need
to transfer some more information for the background layer to
use in the compositor later, so select the background_layer as
the current view layer then in the View Layer Properties tab,
Passes section, under Data, tick the Denoising Data check-box,
and under Light tick the Environment check-box. When you
render you should see that the background is noisy but quick,
and is overlaid by a less noisy diamond. However the
environment texture background is missing.

10. Back in the compositor area there is an addition that will allow
adjustments to be checked much more quickly. From the menu
select Add4Output4Viewer, put the node next to the
composite output node and connect the image output from the
alpha over node to this node as well. In the Sidebar on the right,

272
 open the View tab and tick the Backdrop check-box, you can
zoom and position the result using the controls in the tab. When
something produces an unexpected result it is often useful to
direct intermediate results into the viewer node until you find
the step in your network that is causing the problem.

11. In order to denoise the background image select

Add4Filter4Denoise and drop the node onto the output from
the background_layer. The quality will be improved by
connecting the normal and albedo inputs as well. To blend the
denoised image with the noisy one use Add4Color4Mix and
use the first input from the denoiser node and the second input
from the background_layer directly. Adjust the factor to give the
best result, and try different mixing mode such as darken or
lighten, to reduce blurring without showing too much noise. The
environment texture needs another alpha over node taking its
first input from the Env output of the background_layer and the
second input from the color mix node. The output of this node
goes into the first input of the original alpha over node. You
node layout should look like Figure 13-9 below, with the
exception of the sun beams and lighten nodes applied to the
foreground layer.

Figure 13-9 The compositor node layout.

The sun beams effect added in this example is just one of a vast
number of filters, distortions, masks and other effects available in the
compositor. The full scope is similar to the functionality of specialist
image processing software such as Photoshop, indeed the compositor
has an input node to read an image or movie from file and an output
node to write one, so it would even be possible to use Blender to
process photographs or film clips. However there isn’t enough space in
this book to investigate all the details, so I suggest you do some

273
 experiments on your own. Where a filter or color processing node has a
Fac input, you can use the output of a matte mask node, often blurring
the mask to soften the edges, which makes it very easy to do things
such as reducing saturation towards the edges of the image.

Back in Chapter 10 I mentioned that the correct way to avoid the

spotlight shining onto the shark or the sand was to use the compositor.
You might now have some idea how that could be achieved, it is a little
more complicated than the example you have just completed. There
need to be two view layers and the spotlights need to be moved into
their own collection. In the main view layer the spotlight collection is
completely disabled and creates no illumination or shadows. In the
spotlight view layer the main collection has holdout set so that the
compositor can just use the version without any spotlight effects. The
seaweed object needs to be illuminated by the spotlights as well as the
sun, and it should cast a shadow over the sand and the shark. To
achieve this the seaweed should be put into a third collection which is
enabled and rendered in both layers, but the version with the spotlights
will be composited in front of the version without.

In situations where the foreground object is large, and the

background objects are being rendered at a high sample rate, the time
spent rendering the foreground twice might become quite significant.
In those cases, to avoid duplicated rendering, you can set holdout for
the foreground object in the background view layer, however the alpha
over node will then leave a faint, one pixel wide line around the
foreground object. To avoid that you must add the colors of the two
view layers and then set the alpha value to the sum of their individual
alphas. Figure 13-10 shows the node layout for a simplified version of
the previous exercise without denoising or sun rays.

Figure 13-10 The node layout for compositing layers with complementary alpha values.

274
 As you can see on the left of Figure 13-10, the area of the diamond
is this scene is small relative to the rest of the background objects, so
the benefit of not rendering it is only a few percent, but sometimes the
speed increase will be significant, so this is a mechanism to keep in
mind.

Conclusion

Making non-photorealistic images is an important creative use of

Blender, and a significant proportion of the software development is
directed in support of animator workflow. The approach you took for
the cell and hatching materials was just a simple starting point, and
when you search online you will find of a host of different methods,
but now you have an understanding of the way they work, you will be
able to pick and choose the best ideas.

The final assembly of the cartoon video, with the curled up bugs
climbing back out of the bowl, used several sophisticated techniques
and, if you managed to get everything to work perfectly first time, you
should feel proud of yourself. Even if you had to re-read sections, or
use the support information for this chapter, you will have learned a lot
from this exercise, probably more than if you hadn’t encountered
problems.

The compositor, by its very nature, is the logical final stage of

producing an image with Blender. To use it effectively you need to
build on many other features, and it is no coincidence that this chapter
has come at the end of the book. As I mentioned above, there are many
filters and effects that I don’t have space to describe, but you should
spend some time investigating them on your own. You transferred the
environment texture and denoising information to the background view
layer, but you will have also seen that it is possible to pass freestyle
lines, cryptomatte masks, normals, shadows, UVs, lighting information
and much more, all of which enables you to do some amazing things.

The next chapter, the last one in this book, is dedicated to all the
things in Blender that I didn’t have time to cover in detail. Of course
there wouldn’t be space in a single chapter to do more than itemize all
the subjects that I missed out so, instead, I will concentrate on two
important features that are being actively enhanced with each Blender
release; geometry nodes and grease pencil.

275
 14
W H AT N E X T

Geometry nodes, grease pencil and

all the features that couldn’t be
squeezed into the book.

There are so many useful and effective ways to use

Blender that there just isn’t space to explore them all
properly. In this chapter I provide a basic introduction to
two significant areas not covered elsewhere.
Both geometry nodes and grease pencil are relatively new arrivals
for Blender. Although version 2.79 provided the ability to draw with
grease pencil, the power and flexibility has increased dramatically with
each release since. Geometry nodes were introduced in version 2.9 but

276
 they have since been completely overhauled with many new nodes
available, and a very different method of linking the nodes together. As
at version 3, both geometry nodes and grease pencil have reached a
level of stability where minor details may change, and usability will
certainly improve, but the core functionality should stay the same.
Links to the files and videos for the exercises in this chapter are at get-
into-blender.com 14.WHAT NEXT – Geometry Nodes and Grease
Pencil

Geometry Nodes

A stated aim of the Blender developers is to transfer much of the

configurable aspects into graphical, node based systems, as you
experienced with material shaders and the compositor. The grand name
of this project is everything nodes and, although nodes have already
been used for many years in other parts of Blender, the first release of
functionality under the new heading is geometry nodes.

The number of things you can do with geometry nodes is growing

all the time, but it embodies many of the characteristics of modifiers
along with the particle system ability to scatter instances of other
objects over surfaces. The scope of geometry nodes is already quite
amazing and, as more of the existing mesh and curve modifier become
available as nodes, it will undoubtedly become the default way to
accomplish several of the exercises in this book.

In the first exercise you will use geometry nodes to scatter rocks
and boulders over the surface of another object. This is essentially the
same result achieved by the particle system in Chapter 4 and Chapter 5
but with greater flexibility, control and, once you get the hang of it,
much easier to use.

The Geometry Nodes Modifier

As with physics and particles, the starting point for working with
geometry is placing a modifier in the stack, and, as with those other
two systems, the order of the modifiers will make a difference to end
result.

1. Start a new Blender file, then, to set the screen up with a 3D

Viewport, Geometry Node Editor and Spreadsheet, use the
Active Workspace menu at the top of the window and select

277
 Geometry Nodes. Delete the start cube and add a plane mesh
object. In the Adjust Last Operation pane change the Size to
20 m.

2. In the Modifier Properties tab add a Geometry Nodes modifier,

then click on + New to create a group for this object. There will
be no apparent change in the 3D Viewport but a simple node
tree will appear in the Geometry Nodes Editor.

Mesh Objects Created Within the Node Tree

3. The first divergence from the workflow using particles is the

ability to create meshes within the node tree. For more
complicated meshes, such as the tufts of grass, it is better to pull
the geometry into the node tree from external objects, but for
simple shapes such as rocks, they can be generated
procedurally. From the menu in the Geometry Nodes Editor
select Add4Mesh Primatives4Ico Sphere and drop the new
node onto the pipe running form the input to the output node.
There are a couple of things to note here. The original plane has
vanished, replaced by the ico sphere, and you can see that the input
node is now disconnected from anything. In order to see the plane as
well as the rock the two geometries need to be joined back together,
which you will do later using the Join Geometry node. The other thing
you may have noticed is that, although there are some similarities, the
options under Add in the menu are much more extensive and arranged
differently from the options for material nodes or compositor nodes.
Although I will give explicit menu selections in my instructions for this
exercise, you will find that you need to use the search option when you
are working on your own. Another useful feature that was added in
Blender version 3.1 is the ability to drag a colored connecting dot on
the side of a node into a space, when released it will present you with a
list filtered by compatibility.

4. You are now going to add some nodes to make the regular ico
sphere a little more random as shown in Figure 14-1. First of all
add a new node by selecting Add4Geometry4Set Position
and drop it onto the output connector from the ico sphere node.
Notice that there are three other diamond shaped inputs to the
new node which indicate that they use a different type of data
from the circular inputs and outputs. The diamond connections
relate to a finer level of detail within the geometry, in this case
each vertex in the mesh can be selected, positioned or offset

278
 individually. The introduction of this mechanism was a major
change from the first release of geometry nodes and, as a result,
much of the online help is rather confusing. Hopefully, in time,
this situation will remedy itself.

5. Add a node to create a random offset using

Add4Utilities4Random Value, change the Type to Vector,
the Min to -0.2 and the Max to 0.2, then connect it to the Offset
input of the set position node. Finally position a node to
transform the resultant mesh, after the set position node, using
Add4Geometry4Transform, changing the Scale value in the
Z direction to 0.6 to flatten the rock.

Figure 14-1 Using geometry nodes to create a rock

6. In Figure 14-1 you can see that I increased the ico sphere
subdivision to 2. Try doing that then change the Seed value in
the random value node to see the variety of results. A better
alternative is to return the ico sphere subdivisions to 1 then
insert a new subdivision surface node after the set position node
using Add4Mesh4Subdivision Surface.

7. Before looking into how the rocks can be distributed over a

surface you need to learn a little more about the usage of data
relating to vertices within a mesh. At the moment the surface of
the rock is flat shaded and this cannot be changed to smooth by

279
 toggling the setting on the original plane object. However, there
is a node that can be used so add that after the transform node
using Add4Mesh4Set Shade Smooth.

8. For the sake of this exercise, imagine that, because of the

extreme temperature difference between night and day on Mars,
only the upward facing surfaces should be smooth, while the
downward facing surfaces remain flat. This feat can be achieved
relatively easily using nodes. The first node you need is one that
can generate data about the direction that each face is pointing.
As you have seen before, this property is the normal direction,
so select Add4Input4Normal. It may seen unclear how this
general sounding node can know which mesh it is supposed to
providing the data for; and, if you mouse over the output of the
node you will see the prompt “The socket value has not been
computed yet”. This is the clever aspect of the geometry nodes
system in its final evolution, the specific sources of data are
only set once the node has been connected to other nodes.

9. In order to create a selection criterion based on the normal

direction you need to find whether the Z component of the
vector is positive which you can do by adding two more nodes
by selecting Add4Vector4Separate XYZ and
Add4Utilities4Compare setting the Type to Float, Operation
to Greater Than and the second input Value to 0.0. Connect the
Result output to the Selection input of the set shade smooth
node. The layout is shown in Figure 14-2

Figure 14-2 Applying selection criteria to a node

280
10. You’ve now made a set of nodes that can be converted into a
group, and used to generate a variety of rocks for scattering
over the terrain. Box select all the nodes apart from the input
and output, then hit Add4Group4Make Group. The editor is
now showing the node tree for the new group, so, to give
control over the rock being generated by the group, drag the
Scale input from the transform node to the unconnected socket
of the input node and drag the Seed input from the random
value node to the next unconnected socket. Return to the parent
node tree using the up arrow in the header bar and change the
name from NodeGroup to something meaningful such as
rock_generator.

11. Now it’s time to construct the mechanism for scattering the
rocks over the plane. First of all disconnect the rock_generator
node from the output node and reconnect the input Geometry
output, you should now see the original plane again. Insert a
node to generate locations for the rocks using
Add4Point4Distribute Points on Faces and, provided the
viewport shading is still set to solid you will see a large number
of diamonds appear. Insert a second node to create instances at
each point with Add4Instances4Instances on Points, at
which point all the points will vanish. Finally connect the
Geometry output of the rock_generator node to the Instance
input of the instances on points node.

12. There are two obvious problems with this initial result; the
rocks are too large and they overlap each other, so, before
introducing variation in shape, rotation and size, you should
correct this first setup. Change the Scale values in the
rock_generator node to 0.1, 0.1 and 0.06 , and change the
scattering method on the distribution points on faces node from
Random to Poisson Disk. Now gradually increase Density Max
value of the same node until there are no gaps. Many of the
rocks will be overlapping each other again, but Poisson disk
distribution has another variable that can stop that happening;
gradually increase the Density Min until all of the rocks just
stop overlapping.

13. At this stage you need to combine the original plane with all the
generated rocks, you want to produce a variety of rocks, the
rocks should be rotated randomly about their Z axis and that
axis should be rotated so that each rock sits flat on the surface.
The latter point is not a problem while the plane is flat but for
an undulating surface, it will be important. Insert a node after

281
 the instances on points, using Add4Geometry4Join
Geometry then drag the Geometry output from the group input
node to the join geometry node. Notice that there is a single
input to the join node but it expands to take as many geometries
as you want.

14. You should now see rocks spread over the original plane, but
they are embedded too far in. Open the rock_generator group
and in the random value node change the third Min value to 0.2
and the third Max value to 0.8. This will bias the Z offset
upwards. Return to the parent node tree and duplicate the
rock_generator four times using SHIFT-D then insert a node
onto the connection from the rock_generator node leading to the
Instance input of the instance on points node using
Add4Geometry4Geometry to Instance. Tick the Pick
Instance check-box on the instance on points node then connect
all of the rock_generator outputs into the geometry to instance
node. Finally change the scale and seed values for each of the
rock_generator nodes and you should see the scattered instances
change too.

15. To make the rocks lie flat on the surface connect the Rotation
output of the distribute points on faces node to the Rotation
input of the instances on points node. Then, to make the rocks
rotate randomly around their Z axis insert a rotate node into the
connection you just made, select Add4Utilities4Rotate
Euler and change its method of rotation to Axis Angle with the
Axis 0.0, 0.0, 1.0. You now need to feed a random value into the
Angle input of the rotate Euler node, use
Add4Utilities4Random Value, setting the Min to -pi and the
Max to pi. The node tree and resulting 3D view are shown in
Figure 14.3

Figure 14-3 Scattered rocks on a plane.

282
 Image Textures in Geometry Nodes
To finish off this exercise introducing geometry nodes you will add
textures, allowing you to paint hills and valleys onto the plane object,
as well as painting the density of rocks and boulders. In many ways the
process is simpler than using the particle system but there are one or
two quirks that are hard to figure out from scratch.

1. Although you know how to use normal modifiers to subdivide

and displace a mesh, you will now do that process within the
geometry node tree. Select Add4Mesh4Subdivide Mesh,
add the node to the Geometry output pipe from the input node
and increase the subdivision Level to 4. Add a second node after
the subdivision using Add4Geometry4Set Position this node
will use the Z offset from an image texture.

2. It would be possible to feed image data directly into the Offset

input of the set position node and the red value of a pixel would
be used for the X offset, the green for the Y offset and the blue
for the height. However the standard practice is to draw the
height texture in gray-scale, so a better option is to use a
combine XYZ node to confine the offset to the Z axis only and
to use a multiplication node to allow the height scale to be
controlled. In a space to the left of the node tree position a node
using Add4Texture4Image Texture click New to create a
new image and set the Color to mid gray, to allow valleys as
well as ridges, with HSV 0.0, 0.0, 0.5, un-tick the Alpha check-
box and rename the image to height_map. Create a second node
with Add4Utilities4Math, change the mode to Multiply, use
the Color output of the image texture node as one input and
change the other value to 10.0. Create the third node with
Add4Vector4Combine XYZ, set the X and Y values to 0.0,
connect the output of the multiply node to the Z input and the
output vector to the Offset input of the set position node.

3. At this point the plane will have probably floated up above the
rocks which has happened because both the distribute points on
faces node as well as the final join geometry node both need to
take their input from the same displaced mesh. Reconnect the
nodes so they match the arrangement shown in Figure 14-4.
Now it’s time to paint some terrain.

283
Figure 14-4 Using an Image Texture to displace a plane.

4. In order to see what you’re painting set a material for the plane
object in the Shader Editor add an image texture node using the
height_map image as the base color. You also need to change
the viewport shading to Material Preview. Make sure that the
image texture node is selected in the shader then, using the
workspace tabs at the top of the screen switch to Texture Paint.
As soon as you start to paint you will see that the UV map used
in geometry nodes is incorrect, this is because the UV data has
to be fed into the node tree as an additional input.

5. Switch the workspace back to Geometry Nodes and drag a link

from the Vector input of the image texture node for height_map,
back to the empty socket on the group input node. In the
Modifier Properties tab on the right, you will see that a new
input variable has appeared here called Vector with a
spreadsheet icon and three components all set to 0.0. The icon is
a button that toggles between a single value and an array of
data, click on the button then click in the empty input field that
appears next to it, and select UVMap. To give the variable a
more meaningful name go back to the Geometry Node Editor,
select the group input node then open the Sidebar and in the
Group tab, change the Name from Vector to uv_map. If the
node tree was intended for reuse then you would connect other
variables to the input, such as the scale factor for the height
offset.

6. Now, when you change back to texture painting you should see
the terrain rising or falling with the value of the color you apply,
the rocks rotating to rest flat against the surface however the
surface slopes. To add an image texture for controlling the
density of the rocks, switch to editing the geometry node tree

284
 and duplicate the the height_map texture node, then create a
new texture for it by clicking on the little button with an
overlapping squares icon next to the image name. Set the
default color of the new texture as black and name it
rock_density. Connect the Color output directly to the Density
Factor input of the distribute points node and connect the
uv_map from the input node to the Vector input of the
rock_density image texture node. Go back to texture painting
and choose the rock_density as the referenced image, paint
where you want rocks to appear. Crucially, when you have
finished painting, remember to click the Save All Images
button!
There are so many different things you can do with geometry
nodes that it is hard to limit the scope of this exercise, but from now on
you should be able to find your way a little more easily, you will also
be able to follow the excellent examples available online.

Grease Pencil

Blender’s grease pencil is a bit of an oddity. In many ways it can

be viewed as a rather clunky 2D drawing application, except that it
draws in the normal 3D space, with which you are now familiar. Also,
once you understand how grease pencil works, your general knowledge
of Blender converts the clunkyness into flexibility and productivity as
you can use many methods you already know from editing, sculpting
and modifiers.

Using grease pencil for 2D, hand-drawn, animations is one of its

big selling points and there is much information available online for
using the specialist tools such as onion-skinning and interpolation.
However, in this chapter, I will concentrate purely on basic steps
needed for creating a single image, which is a prerequisite skill for
anyone setting out to make an animation.

Strokes
The grease pencil object consists of a number of strokes, each of
which is a series of points. The points can be shown either as lines
connecting the positions of each point, or as dots, or as image textures.
Each point along a stroke has a size and opacity and the points can be
used to define the outside of a polygon which can be filled with a color.

285
 As well as being situated in the 3D space of the Blender scene,
there are many settings to control the material, the brush, the layer and
vertex color of strokes. Learning where the settings are, and how to use
them will the main focus of this exercise.

The Drawing Plane

Before starting to draw, it’s a good idea to set up Blender to make
your life as easy as possible. To make everything explicit I will go
through all the steps, starting from a new general purpose Blender file.

1. Open a new file and delete the default cube. Set the view to
orthographic along the Y axis by clicking on -Y on the gizmo.
Set the camera to match this view so you can easily return here
later, select View4Align View4Align Active Camera to
View. In order to see the colors as you draw them you need to
set the viewport shading to Rendered and, in the World
Properties tab, under the Surface section change the Color to
white. Finally, select the light object and in the Object Data
Properties tab under the Light section change the type to Sun
and the Strength to 3.0. In the 3D Viewport rotate the direction
of sun light so it shines onto the XZ plane as viewed from the
camera. Later, when you learn about grease pencil layers, you
will find that there is an option to use the actual stroke color
without being effected by lights, but for the moment, this setup
will provide a good starting point.

2. After rotating the light return to the camera view then add a new
empty grease pencil object by clicking Add4Grease
Pencil4Blank. Switch to Draw mode, but before you start
drawing anything look up at the header where you will see a
selection of new options. For the moment just concentrate on
the two that define the location of strokes, the first shows a set
of axes and the word Origin, the second show a computer
monitor and the word View. Change View to Front (X-Z) to
ensure that strokes are laid down as if drawing onto paper held
in the X-Z plane through the origin.

Brushes

3. You will now draw three different lines in order to see the first
place you can control grease pencil strokes. With the mouse
LMB held, draw a rough circle, don’t leave a gap between the

286
 start and the end as that will prevent the flood fill working later.
Notice that the line is rather fine and not completely opaque. To
make the line broader press F while the cursor is over the 3D
Viewport, then increase the diameter of the brush and draw a
second closed loop overlapping the first. From Blender version
3.5 onward, the line that is drawn matches the diameter of the
brush, but on previous version you just have to estimate the size
of the brush.

4. To change other features of the brush, open the Active Tool and
Workspace settings tab. Although all these controls are
duplicated in the editor header, the layout there is cramped and
some options are rather cryptic so it will easier, to start with,
using the tab on the right. You will see that the default brush is
the Pencil, expand the Brush Settings section below the brush
image and check how it is set up. The first variable is the
Radius which you increased before drawing the last stroke.
Notice also that the Use tablet pressure button, to the right of
the value, has been enabled and a graph is shown below. If you
use a tablet to draw, then the harder you press the thicker the
line, up to a maximum at the value of the Radius field. Below
the Radius value is the Strength, which controls the alpha, or
opacity of the stroke. For the pencil brush, once again, the
opacity is pressure controlled but note that the maximum value,
which will be used when drawing with the mouse, is 0.6. If you
look where the lines you drew overlap you will be able to see
the first one through the second.

5. To draw the third line click on the pencil icon and choose Ink
Pen Rough then in the Brush Settings section change the
Radius to 200. The thickness of this brush has non-linear
sensitivity to tablet pressure but the Strength is 1.0, you will not
be able to see through it. Draw another closed loop overlapping
the other two.

Materials

6. Below the Brush Settings is the Color section, where the color
picker is set by default to light green. However the lines you
drew were black because the default setting is to use Material,
change that now by clicking Color Attribute then draw a fourth
closed loop overlapping the other three.

287
NOTE It is very easy to find that the lines you draw look black for two
other reasons. The first is not having sufficient illumination shining
onto the face of the strokes visible from the camera. The second is
forgetting to switch the viewport shading to Rendered.

7. Select the Material Properties tab and expand the Surface

section. There are two sub-sections; Stroke, which is enabled,
and Fill which is not. The default material produces solid lines
and the color is black. Select the color picker for the stroke Base
Color and change it to a mid orange HSV values 0.02, 1.0, 1.0,
say, and you should see the change reflected in the 3D
Viewport. The strokes drawn with the brush set to material have
changed color but the one drawn using vertex color has
remained green. However if you now reduce the alpha value of
the material it will also change the transparency of the stroke
drawn using vertex color. There are effectively four places that
the color and alpha of the grease pencil object can be controlled:
at each vertex, the material settings, the layer and finally by
using modifiers. There is great flexibility here, but there is also
scope for confusion, so adopt each new feature sequentially
after you have mastered the previous one. For the moment leave
the alpha for this material at 1.0.

8. To see the effect of fill on the stroke, tick the check-box in the
Materials Properties tab. The result will not be visible because
the default alpha value for the color is 0.0 so edit the color to
increase the alpha and try changing the HSV values to a color
other than black. There are a few things to note here; where the
line is not fully opaque, the fill is visible through the stroke,
also the fill color for the material is not used for the stroke that
was drawn using vertex color, although the alpha value from the
material setting is used in a similar way to the line itself.

The Flood Fill Tool

In the grease pencil draw mode toolbar on the left, the next option
is named Fill and the icon is the same as the standard flood fill used by
all digital painting apps. However the behavior of this tool is different
in that it calculates an area that would be flood filled, starting from
wherever you click, but then draws a stroke around the perimeter using
the currently selected material.

288
 9. To see the effect of this tool, first change the Black material
stroke color back to black, with full opacity, and un-tick the fill
option. To clarify things later rename this material to black_line.
Now create a new material for filling only, without a line
around the edge, by using the + on the right then clicking +
New below. Rename the new material to gray_fill then un-tick
the Stroke check-box, tick the Fill check-box and change the fill
Base Color HSV to 0.0, 0.0, 0.1 with alpha 1.0.

10. With the Fill tool enabled, click on an area of the drawing
enclosed by one of the pencil or black pen lines, and you may
be surprised to see that, rather than filling with the gray_fill
material, the same vertex color was used as for the last stroke
you drew. This is because the toggle between material drawing
and vertex color persists as you switch between brushes or
between the draw tool and the fill tool. In the header of the
editor area you can see there are two cryptic buttons, just to the
left of the vertex color picker, that can be used to switch
between material and vertex color drawing and filling. See
Figure 14-5.

Figure 14-5 Vertex Color button in the header.

11. You will see how to edit strokes and fills that you have made, in
the Tint section below, but for now just change to material
drawing and fill a different, but adjacent, area of the drawing.
The fill has used the gray material you specified and you can
appreciate two other aspects of drawing with the grease pencil:
The fill tool has various settings to avoid gaps around the edge
which make it overlap other strokes. Each stroke or fill overlays
all the previous ones. This latter feature is the behavior you
would expect if you were using physical paint on paper,
however for digital art it is very convenient to be able to work

289
 23. The grease pencil layer is also where you control whether to use
Blender lights for illumination. Rotate the view so you are
looking at the back of the sketch then toggle the Use Lights
check-box off for each layer. Finally, expand the Adjustments
section and change the Tint Color, Factor and Stroke Thickness
to get a feeling for how they could be used to enhance a work of
art. Figure 14-6 shows the effect of tinting in Draw mode with
only the ink_line layer unlocked.

Modifiers
There is much of grease pencil drawing that I have had to skip
over, however one thing will be familiar to you from every previous
chapter in this book; modifiers. Many of the usual mesh modifiers are
not relevant to grease pencil objects, and are not available, but some,
such as armature or build, can be applied in very powerful ways to 2D
artwork, especially animations.

The modifiers specific to grease pencil include ones such as

Hue/Saturation, Opacity, Tint, Thickness, Dot Dash and Mulitple
Strokes. Their effect and potential usefulness is easy to guess, but an
additional feature of grease pencil modifiers is that each has an
Influence section. The influence is very similar to the use of vertex
groups in mesh modifiers, and, indeed, many grease pencil influence
sections include a vertex group option. However all grease pencil
modifiers can be confined to, or excluded from a specific layer, and
most can also be made to operate on a single material.

24. Try a few of the modifiers to see what they do. First of all
change the interaction mode back by selecting Draw Mode,
then add the following three, confining the influence for each to
the ink_line layer: Multiple Strokes with Fade check-box ticked
and the fade Opacity increased to 1.0. Followed by Noise with
Thickness increased to 0.5 and Noise Scale increased to 0.25.
Followed by Dot Dash with Gap increased to 3.

Sculpt Mode
It’s nearly time to clear all the doodles and start a proper drawing,
but before you do that, there is one more interaction mode that often
proves very useful. Sculpting may seem an unlikely fit with the other
tools required for 2D art, however, in Blender, it provides a familiar
way to make adjustments and corrections to strokes laid down in Draw
mode.

294
 25. Using the interaction mode drop-down select Sculpt Mode. You
will see that the modifier effects have changed, the multiple
lines modifier doesn’t show in edit, vertex paint or sculpt mode
but the other two do. However the fact that they show interferes
with your ability to modify the strokes. From the toolbar on the
left choose the Push tool then attempt to move some of the
strokes. The ink_line layer stroke will not move though the
strokes on the other layer do. In the Modifier Properties tab
deselect all three Realtime option buttons to prevent displaying
modifier in the viewport.

26. When you use the push tool now all the strokes in the grease
pencil object are moved, but that isn’t always what you want.
To the right of the interaction mode drop-down are the three
buttons to use selection to mask the effect of each brush. Select
the second one again to select whole strokes, deselect
everything with ALT-A then use B to box select a part of just
the ink_line stroke. The push tool now only moves the one
stroke, leaving the others unaltered.

27. The same modifier keys work when sculpting a drawing as a

mesh. SHIFT will smooth lines and CTRL will reverse the
normal effect of the brush. On the ink_line stroke try thickening
and thinning, increasing and decreasing strength then
randomizing and smoothing. The size and strength of the brush
can be controlled just the same way as other brush based
interfaces using F and SHIFT-F. Remember to turn the
modifiers back on and switch to Draw or Object mode to see
their effect.

Making a Drawing
You may, or may not be a natural artist, but working through the
sequence of steps to produce a picture will give you some practice
using the techniques outlined above. As a subject I chose some
sketches of Don Quixote and Sancho Panza by Daumier as my
inspiration, but you should feel free to pick whatever you like.

If you have access to a drawing tablet with pressure sensitive pen

then that will make this whole drawing experience much more
satisfying. However it is still quite possible to work using a mouse, you
just might have to make more corrections to wobbly lines in sculpt
mode.

295
 The proposed workflow will use layers in a systematic way, from a
pencil sketch, adding the paint lines over the top, then adding filled
areas below the paint lines and finally adding painted textures over the
fill, but still beneath the lines.

1. In the Object Data Properties tab delete the two existing grease
pencil layers then add four new ones, renaming them as sketch,
flood_fill, paint_blocks and paint_lines.

2. Select the sketch layer, the Black material and the Pencil brush.
The sketch can be fairly low resolution and lightly drawn to
simply act as guides for the stronger painted lines, or it can be
allowed to show through later layers to give additional texture
and life to the finished painting. To reduce the resolution of the
Pencil brush, and avoid ending up with a very high vertex
count, open the Active Tool and Workspace Settings tab then
expand the Post Process section within the Brush Settings and
change the Simplify value to 0.1. Expand the Advanced section
and change the Hardness to 0.95. There is also a Pencil Soft
brush that produces a line of overlapping dots and can be
effective, however if you simplify these lines it will create long
gaps on straight sections of line, which would not be
appropriate here.

3. Draw using vertex color mode but using a neutral color and
create a rough layout as shown in Figure 14-7, using Sculpt
mode to correct the proportions and positions of the main
features as you work.

Figure 14-7 Pencil sketch lines on first grease pencil layer.

4. When you’re happy with the layout, use the paint_lines layer
and Ink Pen brush with the Black material, but using
appropriate vertex colors, to draw over the sketch.

296
 5. On the flood_fill layer your aim is to create areas with solid
color using the gray_fill material. You will find that you can’t
flood fill open areas, so initially create some enclosed areas by
drawing with the gray_fill material to link existing pen and
pencil lines. You can then flood fill the gaps, creating a
patchwork of smaller adjacent areas. There is arguable an
advantage to this as it will enable you to tint them later to vary
the color slightly. Note that when you draw using the gray_fill
material you might be tempted to release the mouse button
before completing the loop around the edge. The fill will then
join the last point to the first with a straight line, which often
appears fine, especially if it is a relatively short distance.
However for later flood filling there will be a gap in the stroke,
through which the flood can enter and overwrite the original fill.

6. Finally, add some more areas of color by drawing onto the

paint_blocks layer, use the pencil brush set to a large diameter.
Draw with the Black material and vary the vertex color, you can
tint over the top but remember to lock the other layers to
prevent them being tinted as well. Figure 14-8 shows my
painting as it nears completion.

Figure 14-8 Four layers used as the drawing nears completion.

As with many techniques in Blender, practice will help you gain

familiarity with the tools at your disposal, and online help videos will
become much easier to follow.

There are some aspects of pixel oriented, digital drawing

applications, such as Krita, Adobe or Autodesk offerings, where grease
pencil will always be at a disadvantage. This is because it works with
strokes rather than a simple image texture for each layer, and so cannot
easily manipulate the drawing pixel by pixel. However, at some stage,
using any multilayer drawing application, you will find that you
accidentally had the wrong layer selected while you were drawing. This

297
 means you either have to undo the brush strokes, or if you didn’t notice
your mistake in time, just live with it. With grease pencil, on the other
hand, you can easily select individual strokes and move them from one
layer to another.

Other Features

There is much of Blender that this book doesn’t have space to

explore; some are recent additions and some have been there for a long
time.

The Texture Node Editor can be used to design more sophisticated

brushes and line styles. The way that textures are used in Blender as
well as the baking process are things actively in development so this
functionality will undoubtedly change after version 3

Asset Libraries build on Blender’s data system which underlies

some of the things covered in this book, such as the effect of users
when saving a blend file. When you set up files and folders for use as
asset libraries, it makes the process of appending assets into a much
simplified drag and drop action.

The Movie Clip Editor allows the integration of existing video

with animated 3D scenes created using Blender.

When adding new objects to a scene you will have noticed several
options we never touched. There are choices under Add4Surface,
Metaball, Volume and Lattice, all of which have rather specific
applications, but can be real time-savers when needed. More
importantly, we didn’t explore Light Probe objects and their associated
indirect baked lighting maps, which can dramatically improve realism
when using the Eevee render engine.

Among the modifiers, none of the left hand column under Modify,
such as Data Transfer were explored, and neither was Texture Paint.
With these features, as with the ones listed above, the day to day use
might be limited, but for certain situations they are very useful.

298
Conclusion

Each day, when I check what’s new in the world of Blender, I find
more and more amazing things that people have conjured from their
imagination, using geometry nodes. Whole landscapes, castles, ruins,
flowers or pipe networks have been generated as if by magic. Once you
have mastered the basic use of geometry nodes for yourself, I suggest
you set out to discover what the latest version of Blender is capable of
doing.

With grease pencil there is also a wide range of stunning art-work,

including many unique styles. But the majority of the really interesting
works of art are grease pencil animations, so, if you get inspired by any
of those, you will need to build up an understanding of the relationship
between grease pencil objects and animation frames. As with every
other part of Blender there are a host of videos and discussion forums
to help you on your way.

This chapter is the final one in the book. Rather than tidy up all the
loose ends it has opened up further areas to explore and created more
questions, the answers to which you will have to discover on your own.

By this stage, though, you have mastered much of Blender.

Discovering different techniques and solutions to new problems will be
more rewarding and much less frustrating than at the start of your
journey.

299
2D Curve 47 .. Modifier 28
3D Curve 47 .. Transformations 56
.. Text 49 .. Transforms 87, 95
.. Textures 196 Arctan2 201
.. Viewport 4 Areas 4
3D Cursor Snap 233 Armature 223
4D Textures 198 .. Auto-names Left/right 248
Absorption Volume 193 .. Automatic Weights 137
Active 17 .. Rigging 98
.. Spline 57 .. Single Bone 225
.. Tool And Workspace Settings 101 .. Skin Modifier 88
.. Workspace 19 .. Symmetrize 248
Add Face 141 Array Modifier 71
Add-on 112, 119 Assets Appending 87
.. Bsurfaces 146 .. Importing 87
.. Download 123 Assign Vertices To Material 60
.. Enable 120 Auto Merge 138
.. F2 146 .. Smooth 33
Add To Selection 16 Auto-names Left/right Armature
Adjust Last Operation 8 248
Aligned 47 Autodesk FBX 113
All Deselect 12 Automatic Number Suffix 248
.. Select 12 Automatic Weights Armature 137
Alpha Blend 115 .. Parent 98
Alpha Clip 115 Avoiding Triangles 136
.. Blend Mode 190 Axis 9
Alpha Over 270 .. Deform 76
Ambient Occlusion 62 Backface Culling 139
Ambient Occlusion 121 Background 43, 54
Animation 223 .. Surface Color 43
.. Bake Action 240, 250 Bake 209
.. Nonlinear 6 .. Extrusion Distance 176
.. Render 229 .. Multiresolution 178
Anisotropy 195 .. Normal 175
Aperture F-stop 62 Bake Action Animation 240, 250
Appending Assets 87 Bake To Keyframes Rigid Body
Apply All Transforms 56, 75 263
Banana 127 .. Principled 18
Base Color 18 .. Transparant 190
Basis Shape Keys 150 .. Velvet 20
Beautify Faces 50 Bsurfaces Add-on 146
Bevel 30, 48, 51 Build Modifier 33
.. Edge 133 Bump Node 67
.. Edit Mode 73 Bump Texture Texture Paint 171
Bisect Mesh 189 Bumps 173
Blend Brush Strength 161 Buttons Mouse 8
.. Brush Texture 183 Bézier 48
Blend Mode Alpha Clip 190 Cache 208
Blendmode 115 Camera 7, 10
Blob Brush Sculpting 94 .. Fly Mode 40
Bloom 62 .. Focal Length 160
Bone Extrude 226 .. Location 10
.. Root 89 .. Positioning 22
Bone Constraint 227 .. View 10, 160
.. Inverse Kinematics 227 .. Walk Mode 40
.. Limit Rotation 227 .. Walk Navigation 160
Bone Root 226 Cell Shader 253
Bone Tip 226 Celtic Knots 121
Book Generator 123 Chain Inverse Kinematics 239
Boolean Modifier 26, 38 Chevron Drop-down Menu 28
Box Select 15 Child Of Constraint 235
Bridge 59 Circle 48
.. Edge Loops 59 .. Mesh 77
.. Edge Loops 139 .. Select 16
Bright Contrast Node 255 .. Select Size 16
Browser File 5 Clay Strips Brush Sculpting 94
Brush Grease Pencil 285 Clean Topology 118
.. Texture 103 Clean Up Mesh 99
.. Texture Transparency 103 Clear Transform Pose 89
.. Texture Mapping 183 Clone From Perspective Image 158
Brush Size Sculpting 92 .. Paint Slot 161
Brush Strength Blend 161 Clone Tool Texture Paint 161
Brush Texture Blend 183 Cloth 205
BSDF Glass 20 .. Modifier 212
.. Pin Group 213 Convex Hull 219
.. Pressure 213 Copy 15, 39
Clouds Texture 36 Copy As New Driver 185
Collada DAE 113 Correcting Normals 50
Collection 37, 80, 270 Crease Brush Sculpting 94
.. Grease Pencil Line Art 265 Create Face 58
Collider 218 .. UV Map 159
Collision Modifier 212 CTRL-TAB 11
.. Physics Properties 207 Cube 7
Collision Shape 218 .. Start 3
Color Base 18 Culling Backface 139
Color Attribute Grease Pencil 286 Curl Finger 233
Color Ramp Node 69 Current View 10
Colorramp Node 254 Curtain 213
Columns Outliner 54 Curve 45, 46
Combine Normal With Bump 180 .. 2D 47
Combine Color Node 254 .. 3D 47
Combine XYZ Node 203 .. Draw 52
Compose 22 .. Extra Objects 120
Compositor 5, 252, 268 .. Modifier 49
Compositor Node Denoise 272 .. Radius 53
.. Sun Beams 272 .. Segments Switch Direction 250
.. Viewer 272 .. Stroke Method 93
Cone 26 .. Tilt 55, 268
Connect Rigid Body 220 Curved Surfaces 50
Connected Only Proportional 150 Custom Properties 184
.. Proportional Editing 160 Cycles 22
Constrain Movement 13 .. Render Engine 39
Constraint 9 Cyclic 49
.. Child Of 235 .. Toggle 49
.. Follow Path 249, 267 Cylinder 27
.. Rigid Body 220 DAE Collada 113
Control Fine 13 Decimate Modifier 32
.. Point 47, 59 Deform Axis 76
Controls Game 10 .. Surface 217
Convert 46 Delete 31
Converter Node 19 .. Dissovle Edges 135
Denoise 42, 55 Edge 11
.. Compositor Node 272 .. Bevel 133
Depth Of Field 62 .. Loop 58, 129
.. Reference Image 164 .. Loops 16, 72
.. Surface 53 .. Slide 148
Deselect All 12 .. Subdivide 12
.. Vertices By Material 60 Edge Loop Select 165
Difference 27 .. Slide 165
Displace Modifier 36, 182 Edge Loops Bridge 139
Displacement 173, 182 .. Select 16
.. Texture 103 Edges Sharp 50
Display As Solid 212 .. Split 50
.. Wire 211 Edit Driver 185
Dissovle Edges Delete 135 .. Hide Mesh 92
Distibution Poisson Disk 280 .. Mode 11
Distribute Points On Faces .. Show Mesh 92
Geometry Node 280 .. Vertex Weights 99
Divisions Resolution 208 Edit Mode Grease Pencil 291
Domain Fluid 208 .. Modifier 132
Dope Sheet 6 Edit Tool Smooth 91
Dope Sheet Editor 241 Editing Proportional 13
Download Add-on 123 Editor Dope Sheet 241
Drag 9 .. Graph 6, 237
Draw Curve 52 .. Image 5, 101
.. Surface 53 .. Nonlinear Animation 241
Draw Brush Sculpting 94 .. Shader 5, 19, 42
Draw Curve Surface 266 .. Timeline 33, 117, 196
Driver Copy As New 185 .. UV 5, 156
.. Edit 185 Editors 4
.. Paste 185 Eevee 22
Drivers 6, 184 .. Render Engine 39
Duplicate 39 Effect Quick 20
.. Keyframe 234 Effector Fluid 209
.. Material 105 Empty Object 74
.. Objects 26 Emulate Numpad 16
Dyntopo 129 Enable Add-on 120
.. Sculpting 91 Endpoint Jump To 197
Engine Render 39 .. Surfaces 50
Environment Texture 54 Flip Image 161
Environment Texture 82 .. Normals 141
Erase Grease Pencil 289 Flood Fill Grease Pencil 287
Even Sided Polygon 136 Fluid Domain 208
Explode Quick Effects 206 .. Effector 209
Export 112, 118 .. Modifier 208
Extrude 14, 35 .. Simulation 205, 208
.. Bone 226 Fluid Flow Behavior Geometry
.. Face 14 210
.. Text 49 .. Inflow 210
Extrusion Distance Bake 176 .. Outflow 210
Eye Visibility 54 Focal Length Camera 160
F-stop Aperture 62 Focus 62
F2 Add-on 146 Follow Path Constraint 249, 267
Face 11 Force Field 207
.. Add 141 Free 47
.. Create 58 Fur Quick 20
.. Extrude 14 Game Controls 10
.. Snap 138 Generate Rig 232
Face On Image 163 Geometry Fluid Flow Behavior
Faces Beautify 50 210
.. Instancing 77 .. Node 68, 192, 201, 275
.. Triangulate 128 Geometry Node Distribute Points
Fade Effect 246 On Faces 280
Falloff 101 .. Group Input 283
Favorites Quick 140 .. Input Variable 283
FBX 115 .. Instances On Points 280
.. Autodesk 113 .. Join Geometry 281
Ffmpeg Video 230 .. Modifier 276
Figure Stick 35 .. Normal 279
File Browser 5 .. Random Value 278
Fill Grid 146 .. Rotate 281
Fine Control 13 .. Set Position 277
Finger Curl 233 .. Set Shade Smooth 279
.. Pose 233 .. Subdivision Surface 278
Flat Shade 17 .. Texture 282
.. Transform 278 Hatched Shading 258
Getting Assets From Other Files HDRI 55
87 Header 8
GIMP 158 .. Scroll 54
Gizmo 9 Hide Mesh Edit 92
Glass BSDF 20 Highlight 8
Grab Brush Sculpting 94 .. Orange 17
Graph Editor 6, 237 .. Yellow 17
.. Modifier 237 Holdout 270
Grass 79 HSV 254
Grease Pencil 275, 284 Human Meta-rig 231
.. Brush 285 Ico Sphere 14
.. Color Attribute 286 Ico Sphere Mesh 14
.. Edit Mode 291 Image 22
.. Erase 289 .. Editor 5, 101
.. Flood Fill 287 .. Face On 163
.. Layer 291 .. Flip 161
.. Line Art 261 .. Profile 163
.. Material 286 .. Reference 163
.. Modifier 293 .. Resize 155
.. Plane 285 .. Texture 101
.. Sculpting 293 Image Sequence 229
.. Stroke 284 Image Texture Node 67
.. Strokes On Back 289 Import 112
.. Tint 290 Importing Assets 87
Grease Pencil Line Art Collection In Front Viewport Display 138
265 Index Refractive 20
Grid Fill 146 Inflow Fluid Flow Behavior 210
.. Snap 35, 57 Input Variable Geometry Node
Group Node 256 283
Group Input Geometry Node 283 Insert Keyframe 197
Groups Vertex 102 Installation 2
Hair 21 Instances On Points Geometry
.. Particle 21 Node 280
Handle 47 Instancing Faces 77
.. Type 59 Intancing Vertices 77
Hard Surface Rigging 224 Interface User 1
Interpolation Mode 198 Loop Cut And Slide 136
Intersect 27 .. Edge 58, 129
Inverse Kinematics Bone Loop Cut 165
Constraint 227 .. Tool 72
.. Chain 239 Loops Edge 16, 72
Invert Select 61, 102 Low Poly 116
Island Selection Mode 168 Mac OS 2
Islands UV 167 Make Group Node 256
Join 17, 29 Manual Online 6
.. Meshes 149 Manually Assign Bone Weight 227
.. Object 17, 29, 149 Map Normal 174
Join Geometry Geometry Node Map Range Node 255
281 Mark Seam 156
Joining Mismatched Quads 143 Mask Modifier 137
Jump To Endpoint 197 .. Texture Paint 102
Keep Corners Subdivision 190 Mask Brush Sculpting 96
Keyboard Short-cut Quick Masking 97
Favourites 140 Matcap Viewport Shading 93
Keyframe 196 Material Duplicate 105
.. Duplicate 234 .. Grease Pencil 286
.. Insert 197 .. Nodes 18
Keys Short-cut 9 .. Offset 41
Knife Tool 134 .. Preview 18
Knots Celtic 121 .. Properties 18
Layer Grease Pencil 291 .. Slots 41, 60
Levels Subdivision 30 .. Surface 41
Light 7, 39 .. Texture Paint 101
.. Point 39 .. Transparent 20
.. Spot 194 .. Volume 41
.. Sun 39 Materials 18
Limit Rotation Bone Constraint Math Node 70
227 Menu Pie 11
Line Art Grease Pencil 261 Merge 51
Linked Select 54, 99 .. Auto 138
Linux 2 .. By Distance 51, 99
Liquid 210 .. Vertex 149
Location Camera 10 Mesh 11, 45, 49
.. Bisect 189 .. Mirror 35
.. Circle 77 .. Multiresolution 179
.. Clean Up 99 .. On Cage 132
.. Extra Objects 120 .. Order 28, 30, 213
.. Ico Sphere 14 .. Realtime 132
.. Shrink/fatten 244 .. Remesh 32
.. Torus 14 .. Render 132
.. Transform Move 12 .. Rigid Body 217
.. UV 156 .. Shrinkwrap 131, 148
Mesh Extras Wall Factory 244 .. Skin 35
Meshes Join 149 .. Soft Body 214
Meta-rig Human 231 .. Solidify 33
Metallic 18 .. Stack 28
Mirror 31 .. Subdivision 35, 148
.. Modifier 35 .. Wireframe 38
Mirror Modifier Symmetry 87 Modifiers 25
Mix RGB Node 255 .. Non-destructive 25
Mix Shader Node 67 More/less Select 16
Mixamo 115 Mouse Buttons 8
Mode Edit 11 .. Wheel 10
.. Object 11 Mouse-over 6
.. Selection 12 Move 12
Modifier Apply 28 Movement Constrain 13
.. Array 71 Multiple Object Selection 17
.. Boolean 26, 38 Multiresolution Bake 178
.. Build 33 .. Modifier 179
.. Cloth 212 .. Sculpting 179
.. Collision 212 Musgrave Texture 19
.. Curve 49 Naming 40
.. Decimate 32 Navigation Walk 10
.. Displace 36, 182 Node Bright Contrast 255
.. Edit Mode 132 .. Bump 67
.. Fluid 208 .. Color Ramp 69
.. Geometry Node 276 .. Colorramp 254
.. Graph 237 .. Combine Color 254
.. Grease Pencil 293 .. Combine XYZ 203
.. Mask 137 .. Converter 19
.. Geometry 68, 192, 201, 275 Nurbs Surface 57
.. Group 256 Object Data Properties 46
.. Image Texture 67 .. Empty 74
.. Make Group 256 .. Interaction Mode 11
.. Map Range 255 .. Join 17, 29, 149
.. Math 70 .. Mode 11
.. Mix RGB 255 .. Offset 73
.. Mix Shader 67 .. Properties 9
.. Noise Texture 70 .. Select 8
.. Normal Map 177 Object Info Node 202
.. Object Info 202 Objects 7
.. Select 176 .. Duplicate 26
.. Separate Color 254 OBJ Wavefront 113, 114
.. Separate HSV 67 Occlusion Ambient 62
.. Separate XYZ 201 Odd Sided Polygon 136
.. Shader To RGB 254 Offset Material 41
.. Tree 277 .. Object 73
.. UV Map 170 .. Relative 76
.. Vector Mapping 171 Ojbect Info Random 123
.. Vector Math Add 180 On Cage Modifier 132
.. Vector Math Normalize 180 Online Manual 6
Nodes 18 Orange Highlight 17
.. Material 18 Order Modifier 28, 30, 213
Noise Texture Node 70 .. Selection 52
Non-destructive Modifiers 25 Origin 35
Non-photorealistic 252 .. Set 219
Nonlinear Animation 6 Orthographic View 35, 36
Nonlinear Animation Editor 241 Outflow Fluid Flow Behavior 210
Normal Bake 175 Outline 8
.. Geometry Node 279 Outliner 5
.. Map 174 .. Columns 54
Normal Map Node 177 Output Properties 42
Normal With Bump Combine 180 Overlays Viewport 32
Normals 33, 50, 173 Pack External Data 104
.. Correcting 50 Paint Texture 100
.. Flip 141 .. Vertex 290
Numpad Emulate 16 Paint Slot Clone 161
Painting Weight 98 .. Image 163
Pan 10 Projection UV 160
Parent 52 Properties 4
.. Automatic Weights 98 .. Custom 184
Particle Density Texture 109 .. Material 18
.. Hair 21 .. Object 9
.. Properties 207 .. Output 42
Particles 79, 205 .. Particle 207
Paste 15, 39 .. Scene 221
.. Driver 185 .. Texture 36
Photoshop 158 .. Vparticle 21
Physics 205 .. World 43, 54
Physics Properties Collision 207 Proportional Connected Only 150
Pie Menu 11 .. Editing 13
Pin Group Cloth 213 .. Editing Size 16
Plane 23 Proportional Editing Connected
.. Grease Pencil 285 Only 160
Play Head 199 Push Down NLA Stack 249
Point Control 47, 59 Quads 127, 135
.. Light 39 .. Tris To 129
.. Vector Mapping 171 Quick Effect 20
Pointiness 68 .. Favorites 140
Poisson Disk Distibution 280 .. Fur 20
Poly Build Tool 144 Quick Effects Explode 206
Polygon Even Sided 136 .. Smoke 208
.. Odd Sided 136 Radius Curve 53
Polygon Count 117 .. Tool 53
Polyhedron 14 Random Ojbect Info 123
Pose Clear Transform 89 Random Value Geometry Node
.. Finger 233 278
Positioning Camera 22 Randomize Transform 108
Preferences 120 Re-topology 126, 135, 138
Pressure Cloth 213 Realtime Modifier 132
Preview 18 Record Video 245
.. Material 18 Reference Image 163
Principled BSDF 18 Reference Image Depth 164
Profile 13, 48 Reflection Window 23
Refractive Index 20 .. Geometry Node 281
Region Render 42 Rotation Track-ball 13, 231
Regions 8 Roughness 18
Relative Offset 76 Sampling 55
Remesh Modifier 32 Save All Images 104
Remove Vertex Group 100 Scale 9
Remove From Selection 16 Scaling UV Map 169
Render Animation 229 Scatter Volume 41, 61, 190
.. Engine 39 Scattering 195
.. Modifier 132 .. Ojbects Over A Surface 79
.. Region 42 Scene Properties 221
Rendered 18 Screen Space Reflections 62, 122
Rendering 22 .. Splash 3
Replay 209 Scroll 10
Resize Image 155 .. Header 54
.. Skin 35, 88 Sculpting 86, 90
Resolution 47 .. Blob Brush 94
.. Divisions 208 .. Brush Size 92
Reverse Sense Of Brush Sculpting .. Clay Strips Brush 94
92 .. Crease Brush 94
Rig Generate 232 .. Draw Brush 94
Rigging 223 .. Dyntopo 91
.. An Existing Mesh 230 .. Grab Brush 94
.. Armature 98 .. Grease Pencil 293
.. Hard Surface 224 .. Mask Brush 96
Rigid Bodies 205 .. Mesh Subdivision 90
Rigid Body Bake To Keyframes .. Multiresolution 179
263 .. Reverse Sense Of Brush 92
.. Connect 220 .. Slide Relax 147
.. Constraint 220 .. Smooth 92
.. Modifier 217 .. Strength 92
.. World 221 Seam Mark 156
Rigify 231 .. UV 156, 167
Rip Vertices 143 Segments Subdivide 56
Root Bone 89 Select All 12
Rope Simulation 243 .. Box 15
Rotate 10, 12 .. Circle 16
.. Edge Loop 58 .. Viewport 18
.. Edge Loop 165 Shape Keys 150
.. Edge Loops 16 .. Basis 150
.. Invert 61, 102 Sharp Edges 50
.. Linked 54, 99 Sheet Dope 6
.. More/less 16 Short-cut Keys 9
.. Multiple Objects 17 .. Variations 13
.. Node 176 Shortcut 6
.. Object 8 Show Mesh Edit 92
.. Tweak 139, 232 Shrink/fatten Mesh 244
.. Vertex 12 Shrinkwrap Modifier 131, 148
.. Vertices By Material 60 .. Snap Mode 131
Selected 17 .. Target Normal Project 133
Selection Add To 16 .. Wrap Method 131
.. Mode 12 Sidebar 8, 31
.. Order 52 Simulation Fluid 205, 208
.. Remove From 16 .. Rope 243
.. Separate 130 Single Bone Armature 225
Selection Mode Island 168 Single Image Texture Paint 107
Separate Selection 130 Size Voxel 33
Separate Color Node 254 Skin Modifier 35
Separate HSV Node 67 .. Resize 35, 88
Separate XYZ Node 201 Skin Modifier Armature 88
Sequencer Video 229 Skin Resize 244
Set Origin 219 Sky Texture 43
Set Position Geometry Node 277 Slide Edge 148
Set Shade Smooth Geometry .. Edge Loop 165
Node 279 .. Vertices 137
Shade Flat 17 Slide Relax Sculpting 147
.. Smooth 17 Slot Texture Paint 101
Shader 18 Slots Material 41, 60
.. Cell 253 Smart UV Project 100
.. Editor 5, 19, 42 Smoke Quick Effects 208
.. Volume 187 Smooth Auto 33
Shader To RGB Node 254 .. Edit Tool 91
Shading Hatched 258 .. Sculpting 92
.. Smooth 51 .. Shade 17
.. Shading 51 .. Keep Corners 190
Snap 35 .. Levels 30
.. 3D Cursor 233 .. Modifier 35, 148
.. Absolute Grid 35 .. Surface Modifier 30
.. Face 138 Subdivision Surface Geometry
.. Grid 35, 57 Node 278
.. Grid Absolute 57 Subsurface Scattering 105
.. Vertex 233 Sun Light 39
.. Volume 232 Sun Beams Compositor Node 272
Snap Mode Shrinkwrap 131 Surface 18, 45
Snapping 13 .. Deform 217
Soft Bodies 205 .. Depth 53
Soft Body Modifier 214 .. Draw 53
Solid 18 .. Draw Curve 266
.. Display As 212 .. Material 41
Solidify Modifier 33 .. Nurbs 57
Specular 116 Surfaces Curved 50
Sphere Ico 14 .. Flat 50
Spin Tool 78 Symmetrize Armature 248
Spinning Vertices 78 Symmetry Mirror Modifier 87
Splash Screen 3 TAB 11
Spline Active 57 Target Normal Project
Split Edges 50 Shrinkwrap 133
Spot Light 194 Text 46, 49
Stack Modifier 28 .. 3D 49
Start Cube 3 .. Extrude 49
Statistics 32 Texture Brush 103
Stick Figure 35 .. Clouds 36
Strength Sculpting 92 .. Displacement 103
Strip Speed 246 .. Environment 54
Stroke Grease Pencil 284 .. Geometry Node 282
Stroke Method Curve 93 .. Image 101
Stuck Zoom 10 .. Musgrave 19
Subdivide 12 .. Paint 100
.. Edge 12 .. Properties 36
.. Segments 56 .. Sky 43
Subdivision 30 .. Vector Mapping 171
.. Voronoi 21, 103 Triangles Avoiding 136
.. Wood 103 Triangulate Faces 128
Texture Mapping Brush 183 Tris To Quads 129
Texture Paint Bump Texture 171 Tunneling 221
.. Clone Tool 161 Tweak Select 139, 232
.. Mask 102 Type Handle 59
.. Material 101 Undo 12
.. Single Image 107 Union 27
.. Slot 101 Unwrap 153
Textures 3D 196 .. UV 156
.. 4D 198 Upgrade 2
Tilt Curve 55, 268 User Interface 1
Timeline 5 Users 259
.. Editor 33, 117, 196 UV Editor 5, 156
Tint Grease Pencil 290 .. Grid Image Texture 156
Tips Tool 6 .. Islands 167
Toggle Cyclic 49 .. Mesh 156
Tool Knife 134 .. Project From View 160
.. Loop Cut 72 .. Projection 160
.. Poly Build 144 .. Seam 156, 167
.. Radius 53 .. Unwrap 156
.. Spin 78 UV Map 100, 153, 283
.. Tips 6 .. Create 159
Toolbar 8 .. Node 170
Topology 126 .. Scaling 169
.. Clean 118 Variations Short-cut 13
Torus 14, 27 Vector 47
.. Mesh 14 Vector Mapping Node 171
Track-ball Rotation 13, 231 .. Point 171
Transform 9 .. Texture 171
.. Geometry Node 278 Vector Math Add Node 180
.. Randomize 108 Vector Math Normalize Node 180
Transformations Apply 56 Velvet BSDF 20
Transforms Apply 87, 95 Vertex 11
Transparant BSDF 190 .. Groups 102
Transparent Material 20 .. Merge 149
Tree Node 277 .. Paint 290
.. Select 12 Wall Factory Mesh Extras 244
.. Snap 233 WASD 10
Vertex Color 290 Wavefront OBJ 113, 114
Vertex Group Remove 100 Weight Painting 98
Vertex Weights Edit 99 Wheel Mouse 10
Vertices 11 Window Reflection 23
.. Intancing 77 Windows 2
.. Rip 143 Wire Display As 211
.. Slide 137 Wireframe 18, 116
.. Spinning 78 .. Modifier 38
Video Record 245 Wood Texture 103
.. Sequencer 229 Workspace 19
Video Sequencer 246 .. Active 19
View Camera 10, 160 World Properties 43, 54
.. Current 10 .. Properties Shader 61
.. Orthographic 35, 36 .. Rigid Body 221
View Layers 268 Wrap Method Shrinkwrap 131
Viewer Compositor Node 272 X-ray 12, 35
Viewport 3D 4 Yellow Highlight 17
.. Overlays 32 Zoom 10
.. Shading 18 .. Stuck 10
Viewport Display In Front 138
Viewport Shading Matcap 93
Visibility Eye 54
Volume Absorption 193
.. Material 41
.. Principled BSDF 193
.. Scatter 41, 61, 190
.. Shader 187
.. Snap 232
Volume Shader 82
Volumes 187
Voronoi Texture 21, 103
Voxel Size 33
Vparticle Properties 21
Walk Navigation 10
Walk Navigation Camera 160