FOUNDATIONS

FOR FILM, TV & GAMEDEV

VERSION 19.5 EDITION

FOUNDATIONS
FOR FILM, TV & GAMEDEV

VERSION 19.5 EDITION

ROBERT MAGEE
HOUDINI FOUNDATIONS

Author: Robert Magee

Research: Michael Buckley, Marcia Utama
Cover Art: Attila Torok

Special thanks to Kim Davidson, Richard Hamel, Chris Hebert, Cristin Barghiel
and everyone at SideFX who brings Houdini to life.

ISBN: 978-1-7753338-2-1

First Published in 2022

by SideFX Software
123 Front Street West, Suite 1401, Toronto, Ontario M5J 2M2

© SideFX Software

Written for the features found in Houdini 19.5

Document Version 2.0 | December 2022

© 2022 | All rights reserved. SideFX, Houdini, Houdini Engine and the Houdini logo are trademarks of SideFX Software Inc. registered in the USA
and other countries. Autodesk, Maya and 3DS Max are registered trademarks or trademarks of Autodesk, Inc., in the USA and other countries.
Unreal Engine and its logo are Epic Games’ trademarks or registered trademarks in the US and elsewhere. Unity is a registered trademark of
Unity Technologies. Other product and company names mentioned may be trademarks
of their respective companies.

DISCLAIMER: Every reasonable effort has been made to obtain permissions for all articles and data used in this book, and to make it as
complete as possible. This book should be considered “as is” and neither SideFX, nor its employees, officers or directors shall be responsible
or incidental for consequential damages resulting from the use of this material or liable for technical or editorial omissions made herein.
 CONTENTS:

HOUDINI FOUNDATIONS 1 | OVERVIEW

Learning Houdini
1
2
The Houdini Workspace 4
Panes & Desktops 6
Nodes & Networks 8
Parameters, Channels & Attributes 10
Selecting Geometry 12
Transform & Edit 14
Modeling Tools 16
UVs & Textures 18
LookDev: Shaders & Materials 20
Solaris: Layout 22
Solaris: Cameras and Lights 24
Rendering 26
Time & Motion 28
Character Rigging & FX 30
Dynamic Simulations 32
Cloud FX & Volumes 34
Terrain & Heightfields 35
SideFX Labs 36
File Management 38
Expressions & Scripting 40
Tasks 42
HOUDINI DIGITAL ASSETS: Procedural Tool Building 44
HOUDINI ENGINE 45
Sharing with other Apps 45
FILM & TV PIPELINE 46
Animation and VFX 46
GAMEDEV & VR PIPELINE 47
Interactive Experiences 47
Products & Licensing 48
Comparison Chart 49

2 | MODEL, RENDER, ANIMATE 51

1. Explore the Houdini UI 52
2. Create a Soccerball 54
3. The For-Each Node 56
4. Setting up UVs 58
5. Layout: Cameras and Lights 60
6. Lookdev: Materials 62
7. Rig the Soccerball 64
8. Animate a Bouncing Ball 66
9. Lights, Camera, Action! 69
10. Set up a Rigid Body Simulation 71
3 | NODES, NETWORKS & DIGITAL ASSETS 75
1. Create a Single Brick 76
2. Copy Bricks to a Point Cloud 78
3. Add Color and Switch to a Teapot 80
4. Color the Points using a Texture 82
5. Create a Brickify Digital Asset 84
6. Test the Digital Asset 87
7. Animating the Bricks 88
8. Loading HDAs into other Applications 90

4 | SMASHING WINE GLASS 91

1. Model the Wine Glass 92
2. Model the Bullet 94
3. Fracture the Wine Glass 95
4. Set up the RBD Simulation 97
5. Add Fluids to the Simulation 99
6. Cache and Re-time the Simulations 100
7. Set up and Render the Shot 103
8. Assign Materials and Render a Sequence 105

5 | DESTRUCTION FX 107
1. Model the Bomb 108
2. Model the Fuse 110
3. Animate the Fuse 112
4. Create an Animated Camera 114
5. Create a Soot Trail 116
6. Create Particle Sparks 118
7. Blow up the Bomb 120
8. Create the PyroFX Explosion 122
9. Export the Geometry to USD 124
10. Set up Shot in Solaris 127
11. Render the PyroFX 130

6 | TERRAIN GENERATION 133

1. Shape the Terrain using Heightfields 134
2. Add and Visualize Mask Layers 136
3. Remap and Erode the Terrain 138
4. Scatter points on the Terrain 139
5. Open the Terrain in Unreal 140
7 | KINEFX RIGGING | FUR DUDE 141
1. Draw the Skeleton 142
2. Capture the Geometry 144
3. Add More Bones 145
4. Joint Orientations 147
5. Attach Capture Geometry 148
6. Paint Capture Weights 150
7. Capturing the Rigid Geometry 152
8. Create Capture Rig Digital Asset 154
9. Create the Animation Rig Asset 156
10. Add More Control Joints 157
11. The Main Controls 159
12. Inverse Kinematics for the Legs 162
13. Reverse Foot Setup 164
14. Promote the Leg and Spine Controls 167
15. Eye Controls 169
16. Animate the Rig 172
17. Add & Groom the Fur 175
18. Set up an Render the Shot 177

8 | PROCEDURAL ASSETS FOR UNREAL 181

1. Create a Simple Building 182
2. Import the Asset into Unreal 184
3. Copy to Points 187
4. Create another Houdini Digital Asset 189
5. Set up Instancing 191
6. Use Geometry to Drive Asset 193
7. Import RBD Simulation into Unreal 195

9 | BUILD A CITY WITH PDG 197

1. Create a City Grid 198
2. Generate and Display Work Items 199
3. Add Attributes 201
4. Create Buildings for the City Grid 202
5. Combine the Buildings 204
6. Isolate a Building 207
7. Wedge the City Core Location 208
8. Create Geometry for the Streets 210
9. Wedge Four City Maps 212
10. Render a Mosaic 213
11. Scale Up to Create More Content 215
HOUDINI FOUNDATIONS

OVERVIEW
To create 3D animation and VFX for Film, TV, Video Games and VR, you need
a combination of technical and creative skills. Houdini is the perfect tool for
bringing these worlds together as you explore, create and refine your projects
from concept to final sign off.
While Houdini has a wide variety of tools designed for generating CG content,
its node-based procedural workflow is what sets it apart. This approach makes
it easier for you to create directable shots, explore multiple iterations and hit
deadlines. As you learn Houdini, understanding how to work with these nodes
and networks will be important to your success.

WHAT YOU WILL LEARN

This overview chapter contains general information about Houdini that will
help you become familiar with important concepts and ideas. While you might
not understand what it all means, this chapter will be a valuable reference point
as you work through the foundation tutorials and build up your knowledge.

ƒ If you already work with 3D software then learning Houdini will be a

transfer of existing skills. You will learn how to interactively build-up shots
using the scene view and shelf tools, then how to work with the nodes and
networks to take advantage of Houdini’s procedural nature.

ƒ If you are new to 3D and Computer Graphics then Houdini is a great package to start with. The
Foundations material assumes some general knowledge, therefore you may want to read up on CG
concepts that you are unfamiliar with. In the end, Houdini will help you achieve a deeper understanding
of what goes on under the hood of not only Houdini but other 3D apps as well.

Once you have finished the foundation lessons, visit SideFX.com for more tutorials. From the main menu,
go to Learn > Learning Paths for a comprehensive list of lessons created by SideFX and members of the
Houdini community. There are lots of lessons for you to explore as you build up your Houdini skillset.

DOWNLOAD INDIE
HOUDINI FOR FREE ANIMATORS AND GAMERS
SideFX has a free learning edition for you to use If you want to go beyond the free learning edi-
as you work through the lessons. The Houdini tion, Houdini Indie removes the watermarks
Apprentice edition gives you FREE access to found in Apprentice and offers higher render
all of Houdini’s features with a few restrictions resolutions up to 4K x 4K and limited commer-
such as limited render size and user interface cial use [less than $100K USD] of Houdini.
and render watermarks. The Indie program makes Houdini a great tool
You can download Houdini Apprentice from for developing personal projects and indie
the SideFX website where you can also get the games. To learn more you can go to:
latest versions which are updated regularly: SideFX.com/indie
SideFX.com/download

Document Version 4.0 | July 2022 | Houdini 19.5+

© SideFX Software 1
 Learning Houdini
Houdini is a computer graphics, or CG, application which you can use to model, animate, render and
simulate. In the process of learning Houdini, you will explore new ways of managing the creative
process that involves the interactive manipulation of nodes, networks and assets.

Everything is procedural in Houdini, which means that PROCEDURAL CONSTRUCTS

modeling, character rigging, lighting, rendering and visual
For motion graphics projects, a procedural approach offers lots
effects all benefit from a node-based workflow where artists
of visually-stunning eye candy. These special effects are often
can build up networks of nodes to manage all the steps
the result of animating parameters on nodes and adding noise
needed to complete a creative task. Networks can then “talk”
in interesting ways that you wouldn’t expect in real life.
to other networks to create even more sophisticated systems.

GOING PROCEDURAL
In Houdini, every action you take is stored as a node. These
nodes are then “wired” into networks which define a “recipe”
that can be tweaked to define a repeatable outcome where
each iteration can generate unique results. The ability for nodes
to pass important information down the chain, in the form of
attributes, helps give Houdini its procedural nature.

mnfst.studio

THE WIDER CG PIPELINE

Beyond VFX and motion graphics, Houdini has bread-and-
butter tools for all parts of the pipeline from modeling to
rendering to character work and gamedev. Its procedural
workflow supports you as you create all of your CG content.
Along the way, you will benefit from the ability to explore
multiple iterations and make changes deep into production.
Andriy Belichenko

KNOWN FOR VFX

Visual effects artists have traditionally gravitated to Houdini
because this procedural workflow is ideal for working with
particles and dynamics. Often visual effects are designed to
react to actions that are taking place in a shot and a procedural
solutions “automate” these reactions. Therefore, Houdini
provides studios with higher levels of productivity and more
control over the creative process.

While the nodes are what makes Houdini unique and give it its
power, there are lots of viewport and shelf tools that are used
HOUDINI FOUNDATIONS

to work interactively while Houdini builds the networks.

Igor Zanic

Houdini is also capable of working with large data sets which is

critical as visual effects become more sophisticated with many
layers such as rigid body destruction, fluids, and particles all
interacting to achieve the final result.

2
DIRECTABLE RESULTS FULL ACCESS TO ALL YOUR DATA
The reason you are able to make edits deep into production is As objects move through a typical animation or visual effects
because changes made to parameters on Houdini nodes will pipeline, they accumulate information which is often stored as
cascade right through the network to update the results. This point or primitive attributes such as velocity, capture weights
directability is retained throughout the creative process and or UV texture coordinates. While other 3D applications hide
can be used to make last minute decisions that would be too this information and attempt to control it for you behind the
costly in a traditional CG pipeline. scenes, Houdini gives you tools for working with and managing
this data. This results in a much more powerful and flexible
approach that makes a huge difference down in production.

TOOL BUILDING
Another benefit of the node-based approach is that it is easy
to encapsulate node networks to create custom nodes that
are shared with colleagues without writing any code. Houdini’s
A NEW WAY OF THINKING
re-usable networks can be wrapped up quickly and easily into
special nodes called Houdini Digital Assets. As you become more proficient with Houdini, you will find
new ways to approach a shot or a game level that will make
you and your team more productive. Houdini gives you the
flexibility to build tools that will support you throughout a
project’s life cycle and instead of simply reacting to issues and
problems, you will be able to anticipate the pain points and use
a procedural solution to work much more efficiently.

These assets open in Houdini, or in other applications such as

Autodesk Maya, 3DS Max, Unreal and Unity using Houdini
Engine plug-ins, with the asset’s procedural nature left intact.

Andriy Bilichenko

Now that you have chosen to learn Houdini, you will find
yourself exploring a versatile application that will redefine how
you approach future projects. The key is to embrace this new
way of working and be ready to explore CG at a level deeper
Marina Bade than you ever imagined.

DO I NEED TO WRITE CODE TO USE HOUDINI?

DEFINITELY NOT! In fact, because of Houdini’s node-based workflow, you will often be able to create results
LEARNING HOUDINI

interactively that would require writing code in other 3D applications. Houdini is very much an artist’s tool and while
it has a technical side that uses scripts and expressions, the out-of-box tools will let you accomplish amazing things.
And the nodes let you easily go back and make changes which mimics how the creative process works.
If you do want to work with code then Houdini supports a number of languages inside the Houdini interface. There
are Wrangle nodes for working with VEX and Python and PyQT is supported as well. You can also use Houdini’s
expression language hscript or you can mix all of them together to meet your specific needs.

3
 The Houdini Workspace
Houdini offers a user interface experience that will be familiar for artists coming from other
CG applications with the biggest difference being the panes used to manage nodes and networks.
The workspace is highly configurable and can be set up to support different ways of working.

Houdini gives artists many different ways to view the bits and SHELF TOOLS
pieces that make up a 3D scene. From the Scene View where
At the top of the workspace, you will find multiple shelves
you look through a camera at your geometry to the Network
filled with tools for creating and manipulating objects,
view where you manage the procedural nodes and networks,
geometry, cameras, lights and effects.
you will find many different ways to make creative decisions
while making sure each shot works at a technical level.

RADIAL MENUS
One way to access tools in the Scene View is the radial menus
which you can access using the X, C and V hotkeys. Each of
These tools work in the scene view and often involve some
these brings up a radial menu with lots of options for you to
sort of scene view interaction. Once you have used one of
choose from. The main focus of each menu is as follows:
these tools, one or more nodes will be created which you can
ƒ Snapping X then refine in the Parameter and Network panes.
ƒ Main C The shelves provide a very important resource to new Houdini
ƒ Views V artists because the shelves reduce clicks and often put down
Once you learn how a radial menu works, you can access the networks of nodes that you can learn from.
tool with a quick sweep gesture without dwelling on the widget.
TAB MENU
Another way to access tools in either the Scene view or the
Network view is to press the tab key. This brings up a menu of
available tools and nodes you can use in your work.

Start Typing and the menu will

show you available tools

You can change the custom menu at the top of the menu bar
which says Main by default. On OS X this is the Radial menu.

Tool Shelf - The shelf tools let you work with

objects and geometry in the scene view.
TOOL BAR:
Selection Modes - Focus on scene,
geometry, or dynamic objects.
Select Tools - Select, Secure Selection
Transform Tools - Move, Rotate,
HOUDINI FOUNDATIONS

Scale or Pose or the Handle tool for

node-specific controls.
Snapping Tools - Turn on Grid,
Primitive, Point or Multi-snapping.
Viewing - Use the View tool to Tumble,
Pan and Dolly or Render Region to render
in the Scene View. Click and hold to change to
2D Pan and zoom.
Output Tools - You can render or
flipbook your scene with these tools.

4
 3D VIEW TOOLS FIRST PERSON CAMERA
Here are some of the hotkey combinations available while While in the View tool, you can turn on a first person fly
viewing. You can skip spacebar/alt if you are actually in the through mode similar to those used in videos games.
View tool:
ƒ Toggle First person On/Off M
ƒ Tumble Spacebar or Alt[Opt] - Left Mouse Button [LMB] ƒ Dolly In/Out W/S
ƒ Pan Spacebar or Alt[Opt] - Middle Mouse Button [MMB] ƒ Pan Right/Left A/D
ƒ Dolly Spacebar or Alt[Opt] - Right Mouse Button [RMB] ƒ Look Around LMB

You can find the View tool in the toolbar. When you use the
spacebar or alt keys, you temporarily evoke the view tool. This VIEWPORT DISPLAY MENUS
can be quite useful when you are selecting or manipulating in a Change how objects appear and views are organized using the
view and need to quickly change your point of view. menus in the top right of the Scene view or the V radial menu.
If you want to focus on viewing then you can press Escape to
go to the View tool. Here are some other hotkeys, you can use
to get your bearings:
Shading Menu
ƒ Home Construction Plane Spacebar + H Object Display Menu
ƒ Frame All Spacebar + A Views Menu
ƒ Home Selected Spacebar + G Shading Menu - Choose from options such as wireframe, flat
shaded, smooth shaded or smooth wire shaded.
2D PAN AND ZOOM
Object Display Menu - As you dive into networks, this menu
You can click on the 2D Pan and Zoom tool in the Operation sets whether you hide, view or see geometry ghosted.
Control bar to change your view in 2D without altering your 3D
Views Menu - This menu lets you split your scene view into
camera position. The widget at the top left lets you click to pan
various views such as perspective or orthographic views.
and zoom or to reset the view with Ctrl - LMB click. This is a
great tool for working with locked off cameras.
DISPLAY OPTIONS BAR
2D Pan and Zoom Widget 2D Pan and Zoom Button At the right side of the scene view, the display bar gives you
access to options for viewport display. Here are a few examples.
Reference Plane/Ortho Grid - Turn on and off a grid that
LMB-Drag to Pan can be used for reference and for grid snapping.
Ctrl-LMB to Reset
MMB-Scroll or Construction Plane - Turn on and off a construction plane
RMB-Drag to Zoom
which is used to define where you place objects or points.
Lock Camera- Lock the current camera to the view so that
view changes modify the camera transform values.
High Quality Lighting with Shadows- Set the best quality of
viewport rendering.
2D Pan and Zoom
Display Primitive Normals - Show the normals belonging to
all primitives in the scene to determine their direction.
Viewport Display Menus - These tabs let you create and
organize multiple panes at the same time. DISPLAY OPTIONS
Pane Tabs - These tabs let you create and organize multiple The Scene and Network views each have Display option panels
panes at the same time. that you can access by clicking on the icon at the bottom of the
Operation Controls - Use the Handle tool with this bar to
Display options bar or using the following hotkey:
access parameters from your selected node. Display Options D
THE HOUDINI wORkSPACE

Parameter Pane - This pane lets you set values, add expressions
and keyframe selected nodes.
Display Options Bar - These toggles let you control scene
display options such as normals, point numbers or lighting.
Scene View - Visualize your work and use handles to
manipulate objects interactively in your scene.
Network Pane - View and manage networks of nodes to work
with the underlying structure of your scene.
Playbar - Set the current time and edit keyframes on selected
nodes. You can also use the Playbar to copy and paste
keyframes.

5
 Panes & Desktops
The Houdini workspace is broken up into Panes which offer unique ways of organizing your scene
data. You can work interactively in a 3D view or analyze attribute values in a spreadsheet. It is
important to learn how these different UI elements can be used to get your work done.

PANES AND PANE TABS

The Houdini workspace is divided into panes so that you Viewers > Scene View [Alt-1] - Work interactively in 3D space. This
type of view can be set up with one or more viewports. You can have
can set up and explore your scene. Pane tabs let you overlap
more than one scene view panels open at the same time to look at
several panes within the same zone to keep them handy but your scene from different points of view.
not visible by default
Composite View [Alt-0] - View images and composites created using
You can access a pane tab by clicking on it in the workspace. Compositing [COP] nodes.
You can close it by clicking on the x. The + menu can be used Viewers > Motion FX View [Alt-^] - This lets you view motion created
to change the pane you are looking at or to add new panes. using Houdini’s channel operator [CHOP] nodes.

Solaris > Scene Graph- This pane shows you the USD scene graph when
working with Solaris [LOP] nodes.
Solaris > Render Gallery - This pane lets you save test renders then
revert back to the settings of each image after reviewing them all.
Solaris > Light Linker - Connect lights and objects.
PANE TYPES
RMB-click on the pane tab to change its type. There are many Render Scheduler - This panel shows you Renderings that are
pane types to choose from. Here are some of them which have completed and in progress. You can pause and kill renderings here.
hotkeys. Some of the other ones are listed here. Refer to the Material Palette [Alt-7] - This palette lets you see all the materials in
documentation to find learn more about all of the others. your scene and select and assign them to objects and geometry.

Animation > Animation Editor [Alt-6]- Manage keyframes and

animation curves. The editor also has Table and Dope sheet views.

Network View [Alt-2] - This view lets you see the nodes and networks
and connect, rewire, and reorganize them to suit your needs.
Parameters [Alt-3] - Set values on parameters, add expressions and
control the properties of your nodes.
Tree View [Alt-4] - This is a hierarchical view of the nodes. This can be
a great way to understand how scene hierarchies work.
HOUDINI FOUNDATIONS

Animation > Channel List - Create channel groups and manage the
scoped channels as you animate in Houdini.
Animation > Autorigs- This pane gives you access to tools for building
your own rig out of modules for biped, quadruped and facial rigs.
Animation > Character Picker - You can use this pane to make it easier
to select parts of a character rig.

6
 DESKTOPS
Inspectors > Geometry Spreadsheet [Alt-8]- A view of the attribute As you open up tabs, add dividers and organize your pane tabs,
values you have on your geometry. This could include UVs, normals, or
you start to set up your own workspace. To save any layout, go
custom values you have set yourself.
to the Desktop menu (Windows > Desktop on OSX) where you
Inspectors > Data Tree- This view gives you access to a light bank,
can access saved desktops, save your own and manage them
Material Stylesheet and Object appearance list.
as you work. When you save a desktop, it will save the Pane
layout, Radial menus and visible Shelf sets.
Mantra Rendering > Render View [Alt-9] - Start interactive Mantra
renderings, that will update when you change something in your
scene.
Mantra Rendering > Take List- This list lets you explore different
“takes” by making changes to specific parameters. You can then
manage the takes to focus on your preferred creative choices.

TOPS > Task Graph Table - The Task Graph Table shows the metadata of
all work items in the graph, or all work items in a particular node.

Misc > Orbolt Asset Browser - This browser lets you access assets
from Orbolt.com. To use this pane you need to log in with an orbolt.com
account.
Misc > Textport - You can use this pane to type commands.
Misc > Python Shell -You can type Python commands using this pane.

ORGANIZING AND COLLAPSING PANES

When you save your scene, it remembers which Desktop you
Both panes and tool bars can be collapsed and expanded by are looking at but not any changes to the pane layout while
clicking on the arrows found in their UI. Whole panes can be working. These changes will go away unless you explicitly save
collapsed to the left or right and you can flip the contents them to the desktop or create a new desktop.
using the center grip. These options let you focus on certain
panes by hiding others using a single click of the mouse. SHELVES AND SHELF SETS
Drag here to position divider To manage the shelves at the top of the workspace, access the
menu found under the arrow icon. You can use this to work
LMB-click here to collapse up
with Shelf sets. You can also bring up shelf sets that might be
LMB-click here to flip panes hidden in your desktop.
Alt LMB-click to flip orientation

LMB-click here to collapse down

PANE MENU
At the top left of each pane is a button for maximizing and
minimizing the pane and an arrow which gives you access to
COLOR SETTINGS
the Pane menu. This menu lets you tear off the pane or a copy
of the pane, close, or split your pane. You also have options for You can customize the look of the Houdini UI by choosing a color
determining the UI of each pane. scheme for your workspace. Select Edit > Color Settings to bring up
the option window then you can choose from the default Houdini
Dark or Houdini Light. Click on the Download button to choose
from a list of color schemes created by the Houdini community.
PANES & DESkTOPS

7
 Nodes & Networks
With Houdini’s node-based workflow at the heart of its procedural architecture, the ability to work
directly with these nodes and networks becomes very important to using it effectively. While the
idea of nodes might sound technical, they are actually quite artist friendly and easy to work with.

As you use tools in Houdini, nodes are created and wired with CONNECTING AND DISCONNECTING NODES
other nodes. The resulting networks offer a history of your
When you work in the Viewport, nodes are often placed and
actions while providing a simple way to make changes and
wired together automatically. When you want to reconfigure
refine your work. Learning how to work effectively with the
how a network is set up, you will need to connect and
node networks is an important part of working with Houdini.
disconnect nodes by hand.
NODE FLAGS Here are some ways to interact with nodes and connections in
Each of the nodes have various flags which determine if it is the Network pane:
displayed, locked or bypassed. You can evoke these by either ƒ Connect Node LMB-drag from output to input
clicking on the flag itself or using the radial node menu. ƒ Connect Nodes J drag across the nodes
ƒ Insert New Node RMB on an output or connector
Toggle Bypass Toggle Display
ƒ Insert Node LMB drag and drop onto connector wire
ƒ Disconnect from Wires LMB select then Jiggle node(s)
ƒ Cut Wire Y drag across connector wire
Info Box ƒ Move node LMB-drag
ƒ Copy selected nodes Alt + LMB-drag
ƒ Reference Copy Alt + Shift + Ctrl + LMB-drag

Dot nodes can be used to organize your networks:

Toggle Freeze
Toggle Template ƒ Add Dot Alt + LMB wire
ƒ Pin/Unpin Dot Alt + LMB dot
Display Flag [R] - This flag lets you choose the display output
node for the network and is highlighted with a hollow ring. The NODE GALLERIES
Render flag [T] sets which node will be output for rendering and
The galleries offer quick access to nodes that you want to add
is highlighted with a solid circle. You can set this separately from
to your network directly. The galleries contain those nodes
the Display flag by Ctrl-clicking on the Display flag.
used the most in day-to-day work while the tab key gives you
Template Flag [E] - This flag displays the node in grey and can access to all the available nodes.
be used for reference or snapping. You can create your own galleries using the Windows >
Gallery Manager and you can add items to your galleries by
Freeze Flag - This caches at the locked node and all nodes
RMB-clicking on a node then choosing Save to Gallery...
earlier in the chain are ignored when the network is cooked.
Nodes saved in the Mat network will also be available in the
Bypass Flag [B] - This flag lets you ignore the node when Material Palette as long as they are given the proper keywords
the network is cooked. such as Mantra for Mantra materials.

NETWORK VIEW
Network Path - The path leading to the
current network level. You can also use
this bar to navigate to other networks.
Pane menu - These menus and icons
are for organizing your network.
Network Background - Use the Pane
HOUDINI FOUNDATIONS

menu to add an image or set up a grid

to help you organize your nodes.
Network Box - Group related nodes
then quickly collapse and expand them.
Sticky - Add notes to help other artists
read your network or to offer ideas for
their networks.
Node Gallery - Drag nodes from here
to your network. Use the filter at the
bottom to find the node you need.

8
NETWORK TYPES NAVIGATING NETWORKS
Houdini includes different kinds of nodes which each work To jump between network types there are a number of
in their own context. The network type is labelled in the top different approaches you can take. Some of these happen
right corner of the network view. Nodes from each type can naturally as you work with objects in the scene view and
connect to other networks. While the different types of nodes others offer shortcuts which allow you to work more quickly.
are similar in how they are wired together, they each have
unique capabilities: Selection Modes - As you select in the scene view, the
network editor jumps to the location of the selection.
ƒ Scene Objects OBJ Different selection modes will in turn take you to different
ƒ Geometry Surface Operators SOP network types as you make a selection.
ƒ Solaris Lighting/Layout Operators LOP
ƒ Materials VEX Builder MAT Network path - You can LMB-click on a parent node to
ƒ Motion FX Channel Operators CHOP navigate back up the path or LMB-click on the container node
ƒ VEX VEX Builder VOP to access parallel nodes or dive into the contents of other
ƒ Outputs Render Operators ROP container nodes.
ƒ Tasks Task Operators TOP
Radial menu - Press n to get a radial menu that lets you
ƒ Dynamics Dynamic Operators DOP
navigate up, down and to different network types.
ƒ Compositing Compositing Operators COP/IMG
Hotkeys - These help you navigate up and down as you work
As you work with Houdini, you will begin to learn how to use
with a selected object.
this “secret” language to talk about the node types and how
ƒ Dive in I
they apply to working procedurally.
ƒ Jump up U
ƒ Toggle Objects/Geometry F8
NETWORK PATHS
ƒ Previous or Next Network Alt + f or Alt + 
Nodes are organized hierarchically with some nodes nested
in other nodes known as network managers or subnetworks. Quick Marks- These let you quickly set and return to network
To help you manage these hierarchies, a browser-like path is locations. You can use them as-needed and then override them
available at the top of most panes. or forget about them. They are not saved with the scene file.
ƒ Set a Quickmark Ctrl + 1, 2, 3, 4 or 5
Back/Forward Button ƒ Return to a Quickmark 1, 2, 3, 4 or 5
Path Hierarchy ƒ Go Back to Previous View `

SELECT AND VIEW HOTKEYS

Parent Node
Pin Path In the network pane, you will need to pan and zoom around
Container Node
Drag Target to another Pane to work with the complete network. Here are the key
combinations for these actions.
Use this path to navigate up and down the hierarchy or to ƒ Pan MMB
other networks. By default, the path changes as you make ƒ Zoom RMB
selections in the scene view although you can Pin down a ƒ Select Nodes LMB
path to keep it focused. You can also drag the Target icon to a ƒ Add to Selection Shift + LMB
pinned pane to sync up their paths. ƒ Remove from Selection Ctrl + LMB

Node - This represents an operation that

contributes to the final output of the network.
LEARN ABOUT YOUR NODE
Network Type - Shows which network type you
Bring up the Info Box using either
are working in. the Radial Menu or MMB-press on
the node. This panel gives you info
Connector - The connecting lines show how your
nodes are linked together and how the data is about the node’s contents, groups,
moving through the network. attributes and other important
Dot - You can add dot connectors to make it facts. This panel also highlights any
easier to organize your nodes. errors that are interfering with your
Nodes & Networks

workflow.
Display Ring - This small circle shows which node
is displayed in the Scene view. This panel will close automatically but
Render Ring - This large circle indicates a render you can click on the Pin icon to keep
node even if another node is displayed. it visible as you work. You can add
Comments - Node comments can be displayed comments and display them in the
to help other artists understand your thinking. Network view using this panel.
Palettes - Buttons on the menu display Palettes
that let you set the color and shape of nodes.

9
 Parameters, Channels & Attributes
All of the nodes in Houdini are driven by parameters, channels and attributes to help you achieve
the results that you want. The terminology used in Houdini may differ from other 3D applications
therefore it is a good idea to take a moment to understand them in a Houdini context.

PARAMETERS CHANNELS | KEYFRAMES

Parameters refer to the values, sliders, buttons and checkboxes You can set a keyframe on a parameter by pressing the Alt
found on Houdini nodes. These are sometimes referred to as key and LMB-clicking on the name or value field. Once you
attributes in other applications but Houdini uses attributes in a have set a keyframe, the parameter’s field changes color
different way. and you have created an animated channel. There will now
You can change parameter values in the Parameter pane or be keyframes associated with the parameter which you can
using handles in the viewport. There is a RMB menu on each access in the Animation editor.
parameter that gives you a number of important options such
as copying and pasting and reverting to defaults.
CHANNELS | EXPRESSIONS
Instead of a raw value, you can add expressions into the
parameter using either hScript or Python. There is a menu in the
top right of the Parameter pane for choosing which language
you want to use. You can press Ctrl-E to bring up an expression
editor with a number of scripting tools to make it easier to work.

SEARCHING PARAMETERS
A node might have a large number of parameters and sorting
through them all can take time. If you click on the magnifying
REFERENCE SCENE DATA
glass in the top right, you get a search bar that lets you filter You can also RMB-click on a parameter and choose Reference
the parameters based on name and content. You can find > Scene Data to bring up a window for choosing specifically
parameters using expressions, overrides and even a raw value. what you want to link to. Once you have made a choice from
any node in your scene, to create a channel reference. This
method lets you create references without worrying about the
exact syntax needed to write the proper expression.

PARAMETER PANE
Navigation Bar- This bar lets you see where the
node is located in the scene hierarchy.
Node Type and Name - Here you can see the
node type and set its name. Clicking on the icon
gives you a menu for working with the node.
Search Bar - Click on the magnifying glass icon to
search parameters by name or by content.
HOUDINI FOUNDATIONS

Changed Parameter - when a parameter has been

changed from its defaults then its value is bolded.
The folder tab name is also bolded.
Animated Parameter - when you have keyframed
a parameter it is highlighted in green.
Locked Parameter - You can RMB-click on a
parameter to lock and unlock it. It will be
highlighted in grey
Select to Match - These icons let you match
these parameter values to other objects.

10
CUSTOM PARAMETERS ATTRIBUTE TRANSFER
If you click on the Gear icon in the top right of the Parameter Within a node chain, attributes are attached to geometry then
pane, you can choose Edit Parameter Interface. Here you can used by other nodes. You can also pass attributes to other
add custom parameters which can then be linked to other pieces of geometry using Attribute transfer. Here the sphere
parts of your node network. is passing color attributes to the boxes based on a defined
threshold value.
ATTRIBUTES
Attributes let you attach data to your geometry that can be
used by nodes down the chain to complete an operation. A
fuel attribute can drive a Pyro FX simulation or a UV attribute
sets up texturing. Some attributes are created by Houdini
nodes or you can create custom attributes.

ATTRIBUTE WRANGLE
Houdini has a wide variety of nodes that let you create and work
with attributes. You can also use the Attribute Wrangle node to
Class - Attributes can belong to Points, Primitives, Details and use a script-based approach to this work. For a lot of Technical
Vertices. This will affect how they get used down the chain. Directors this may be the most comfortable way to work.
Type - You can set up float, integer or string attribute types
amongst others.

ATTRIBUTE RANDOMIZE
Attribute Randomize lets you create an attribute and immediately
randomize its values. For instance here you can see the Color,
rotation and Scale of these boxes being randomized.

For artists, working with nodes will make it easier to deal with
this kind of information. A lot of Houdini’s power is found in the
proper use of Attributes and you will eventually need to learn
about them.

GEOMETRY SPREADSHEET

Parameters, Channels & Attributes

You can view all attribute values, even invisible
ones using the geometry spreadsheet.
Navigation Bar- This bar lets you see where
the node is located in the scene hierarchy.
Node Name - This shows you which node is
currently selected and which node is generating
these attribute values.
Attribute Class Buttons - Use these buttons to
filter which kind of attribute you are looking at.
Point Number - Here are the geometry point
numbers to help you determine where the
attribute is on your model.
Attribute Values - These are the values at this
point in the node network chain.
Filter - You can type parameter names here
to filter the list when you are working with
lots of parameters.

11
 Selecting Geometry
Working in Houdini involves the selection and manipulation of many different elements. There are
a number of tools and options available to help you work efficiently with objects and geometry
components such as points, edges and primitives.

SELECT TOOL There are also some selection filters that let you focus on
visible geometry or select groups. You have a wide range of
The Select tool lets you focus on making selections therefore
selection options to help make this easier as you work.
it doesn’t have any manipulation handles.
ƒ Select Visible Geometry Only Shift + V
ƒ Select Tool Tap S
ƒ Select Fully Contained Geometry Only Shift + C
When working with tools such as Move or Rotate and Secure ƒ Select Groups or Connected Geometry 9
Selection is on, you need to invoke the Select tool to make a ƒ Select Whole Geometry Choose in Operation Control bar
selection. Toggle Secure Selection to Off to select freely. ƒ Select by Normals Choose in Operation Control bar
ƒ Evoke the Select tool while in other Tool Press and Hold S
ƒ Toggle Secure Selection ~ SELECTION MODES
Selection modes, give you access to objects and components.
SELECTION TYPES They also let you easily jump from object level to geometry
There are different shortkeys for adding, subtracting or level using a the buttons in the toolbar or the hotkey.
toggling your selection as well as for selecting all or none. Objects - The object network level is where you work with an
These techniques play an important part in this workflow. object’s transforms. In any tool other than the View tool, the
following hotkey will bring you back to the Object level:
ƒ Select LMB
ƒ Add to Selection Shift + LMB ƒ Objects 1
ƒ Remove from Selection Ctrl [Cmd] + LMB
Geometry - You can use any of the following hotkeys, when
ƒ Toggle Selection Ctrl [Cmd] + Shift + LMB
not in the View tool, to jump into the geometry level with the
ƒ Select All A [Object Level] / N [Geometry Level] chosen components available for selection.
ƒ Select None N [Object Level] / Shift + N [Geometry Level] ƒ Points 2
ƒ Edges 3
SELECTION TECHNIQUES ƒ Primitives [Faces] 4
In the viewport, you can choose from four different selection ƒ Vertices 5
types that offer different ways of accessing geometry.
ƒ Box Select F2 TWEAK MODE
ƒ Lasso Select F3 Only one geometry selection modes can be active at a time. If
ƒ Brush Select F4 you are working with an Edit node then Tweak Mode lets you
ƒ Laser Select F5 choose any combination of points, edges and primitives.

SELECTION OPTIONS
SELECT MODE MENUS Edit | Components You can choose which
Each of the selection modes comes with components you want to work with from this
options which let you alter how you collection of buttons. Here edge selection has
been chosen.
interact with your scene. You can access
these options by either LMB or RMB- Select Tool - The Select tool lets you make the
selection. To access it press the S hotkey.
clicking on each mode’s icon.
Secure Selection - This locks your selection
While working with components, this when using other tools. To invoke the Select
menu lets you choose to show Display or
HOUDINI FOUNDATIONS

tool with it on, press and hold on the S hotkey.

Current Operators. These same options
Selection Types - You can use this top bar to
are available at the top of the Scene view change your type of selection. You can choose
when working with the Edit node. from Box, Lasso, Brush or Laser. There are also
This menu is different at the object level some filter options.
where it includes some filters for differ- Edge Loop - To select an edge loop you can
ent kind of objects as well as options for double-click while selecting edges. To select
partial loops, select one edge then press A and
more easily selecting materials, con-
then an end edge. This works with points and
straints and Digital Assets.
primitives. You can also select point loops or
primitive loops using the same technique.

12
HOW SELECTIONS ARE USED BY TOOLS SELECT ALL AND THE GROUP FIELD
When you make a selection in the viewport then use a tool, To select all of the primitives on incoming geometry, leave
a node is created and the selected points, primitives or edges the Group parameter blank. If the topology of the incoming
are listed in the node’s Group parameter. geometry changes, everything will be operated on by the node.
Using Select All [N] in the viewport will usually ensure that
this field remains blank when a tool is used. With some tools,
the Group field will show all the selected parts and you would
have to clear the field manually to make it blank.
For example here we see primitives 5, 6, 9 and 10 are being
used by a polyextrude node. You can see them listed in the
THE GROUP NODES
Group node and then used to extrude the faces. Group nodes let you refer to a defined selection of points,
vertices, polygons, or edges by name. You can define a
group interactively, by selecting components in the viewer,
or mathematically, using ranges or an expression. The group
name can then be assigned to the Group parameter instead of
using point or primitive numbers.

If you were to change the topology of the incoming geometry

node then there might be more or less faces and the
extrusions will have moved to a different location. This may
not be what you want and you may need to reselect faces.

Here are some Group nodes for you to choose from:

ƒ Group Create - Use interactive selection, a bounding box, face
normal direction or edge angles to populate the group.
ƒ Group by Range - This lets you choose a range and a simple pattern
To do this, you can select polyextrude, press Enter to go to the to populate the group.
Handle tool and press ` to go into re-select mode. Select new ƒ Group Expression - With this node you can use a vex expression to
primitives then press Enter and your new selection will be define the membership of the group.
used in the Group parameter. ƒ Group Paint - This node lets you use an interactive paint interface
to choose the geometry for the group.

Shading Options - The shading

options determine what you see in
the Scene view. In this case, we are
using Smooth Wired Shading.
RMB Menu - While in the Select
tool, this menu gives you access to
selection options such as inverted
selections, boundaries or growing
and shrinking your selection.
Display Filter - This filter lets you
turn off things you don’t need
Selecting Geometry

such as bones, null objects, lights

or cameras to let you focus on the
work at hand.
Display Options - While selection
modes will show you edges or
points to help you select, they will
not be visible when using other
tools. Use these options to keep
them visible even when not using a
specific modeling tool.

13
 Transform & Edit
From basic transformation tools for objects, to the pose tool for animation rigs and the edit node for
reshaping geometry, there are a number of different tools that let you use interactive handles in the
viewport. In Houdini, these handles are tied closely to the node you are working with.

TRANSFORM TOOLS MMB TRANSLATION

The transform tools give you handles that you can use to If you don’t want to click directly on handles, you can use
manipulate objects or reshape geometry. When you transform the Middle Mouse Button in open space combined with a
objects, the parameters at the object level are updated to drag to move it along the construction plane. To change this
reflect your changes. to translate along nearest axes go to Edit > Preferences >
Handles and set Translate Handle to Map Drag to Axis.
ƒ Move T
ƒ Rotate R
ƒ Scale E
ƒ Pose Ctrl-R MMB-Drag up/dwn
ƒ Handle Enter Moves along Y axis if
The Handle tool gives you access to handles that are specific the preference is set to
Map Drag to Axis
to your selected node. While using these tools, you can
re-select by pressing and holding S, making a new selection
then releasing S and continuing to transform.

TRANSFORM HANDLES POSE TOOL

When you are working with a Move handle you can work with When you are animating, you can use the Pose tool to work
a single axis, two axes or move along the camera plane using with bones and to display motion path handles that reveal the
the center. Rotate and Scale handles offer similar controls. motion of an object. You can then use tangent handles and
keyframe points to modify the motion in the viewport.
Single Axis

Two Axes IK Handle

Camera Plane
[Three Axes] Motion Path
Handle

EDITING GEOMETRY
Edit | Components You can choose
the components you want to edit
using these buttons. The Points
option has been chosen here.
Move Tool - The Move tool lets
you translate the selection using
the Scene View handle.
Move Handle - This handle lets
HOUDINI FOUNDATIONS

you move along one axis using the

lines or two axes using the square
dots. RMB-click on the handle to
access the handle options.
Soft Edit Radius - when moving
points on a surface, you can use
this radius value to create a soft
falloff as you move the points. The
Soft Edit Radius doesn’t work with
primitives or edges.

14
 EDIT NODE HANDLE TOOL
If you try to move geometry components then an Edit node After using a shelf tool, you often find yourself in the Handle
is placed down to accept your transformations. In addition tool. Or you can select a node in the network and press the
to transforming the geometry, you can slide on surface, work Enter key in the Scene view to go into the Handle tool. This
perpendicular to the normals or sculpt the surface. brings up a handle which focuses on the specific parameters
for the selected node such as the distance parameter on a
ƒ Edit T/R/E
polyextrude node.
ƒ Slide on Surface L
ƒ Peak H Show Current Operator - By default when you select a node
ƒ Sculpt B other than the display node, it becomes the current node and
you get a wireframe display of the geometry. You can then
SOFT FALLOFF use the handle to manipulate this intermediate node while
evaluating the results on the shaded surface.
When you are transforming points, you can use the Soft Edit
Radius to create a falloff. There is a visualizer that is evoked to
show you where the falloff is occurring on the surface.
Peak Handle

Soft Edit
Visualizer

Show Display Operator - Another option is to always show the

Display Operator. In this case, selecting a node in the chain
Soft Edit HUD
does not show the wireframe and the handles stay focused on
the display node.

EDIT OPTIONS
If you RMB-click while in the Edit node, you can access
options for transforming your selection. You can make a circle
or straighten the selection. These options work with points
and edges but not always with primitives.
ƒ Make Circle Shift-C
ƒ Evenly Space Selection Shift-E
You can change parameters in the parameter pane for the
ƒ Relax Selection Shift-R
current node but the handles will continue to work with the
ƒ Straighten Selection Shift-S
parameters on the display node.

Shading Options - The shading options

determine what you see in the Scene view. In HANDLE OPTIONS
this case, we are using Smooth wired Shading. All handles have a menu that can
Sloppy Selection - Three of the component be accessed by RMB-clicking on
buttons can be selected at the same time if the any part of the handle.
Edit node is using Sloppy selection which makes
all of them available for a more fluid selection This menu gives you options for
process. aligning the handle, detaching it
RMB Menu - This menu gives you access to Edit from the parameters of the node,
tool options such as which type of edit you want pivot mode and more. You can
to focus on. This information is also available on use these options to customize
the top bar of the Scene view. how the handle works.
TRANSFORM & EDIT

Component Selection - You can also select the You can also keyframe param-
component type using this menu. This offers
eters on a handle and promote
you the same options you would find on the
main toolbar.
all parts of the handle to a Digital
Asset. By promoting the param-
Edit Options - You can use this menu to edit
eter the handle will be accessible
the components using operations such as Make
circles and Straighten the selection. at the asset level.

15
 Modeling Tools
Houdini has a lot of tools for creating, shaping and deforming geometry to achieve a desired look.
Here are just a few of the many tools you will use on a regular basis when building models in the
geometry, or SOP, context of Houdini.

CREATION UTILITY NODES

To start creating geometry, you can start with some basic Because of Houdini’s procedural nature, modeling actions like
shapes or draw a curve. In each case you get an object with copy, clip and mirror create nodes in your network. This can
a geometry/SOP node inside with the tool’s name. You can makes it easier to go back and make changes later on.
access these on the Create shelf or in a radial menu. Clip - Cut your model based on a clipping plane. You can set the
direction of the clip and whether you keep one half, the other or both.
Primitives - Houdini includes Box, Sphere, Tube and Torus primitive
Mirror - This tool flips the geometry based on a clipping plane.
shapes along with a variety of Platonic solids.
There is an option to fusing the points after mirroring.
Grid - The grid tool offers a great starting point for a wide range of
Copy and Transform- This node will let you create multiple copies
models. You can set its shape and size at the geometry level.
based on transformation values.
Curve - Draw a curve by laying down control points then create a
Blast - This node lets you delete polygons from your model. You
Bezier, NURBS or poly curve.
can choose to remove or keep the selected polygons. If you press the
Delete key when points or a polygons are selected they will be blasted.
POLYGON MODELING
Dissolve - This tool lets you remove edges without breaking up
Polygons are one of the most popular geometry types the surrounding geometry. Pressing the Delete key when an edge is
especially in video game projects where they are mandatory. selected will dissolve it.
Houdini has a comprehensive set of poly modeling tools which
you can use to develop your models. SUBDIVISION SURFACE MODELING
In Houdini, you can model with polygons then display and
PolyDraw - This tool lets you interactively draw a polygonal mesh
render them as subdivision surfaces using options found on
on the construction plane or by snapping to existing geometry.
the Render tab on the object’s parameter pane. You can also
PolyExtrude - Push or pull on one or more polygons to reshape the
create a Subdivide node at the geometry level to add polygons
geometry. Control the extrusion profile to get a wide variety of shapes.
to give you a more detailed topology to work with.

PolyBevel - Bevel selected edges to create chamfered or rounded

bevels. You can often use the output group from a previous node such
as Polyextrude or Boolean to automatically find the right edges.
PolyBridge - Connect two sets of polygons with control over the
SURFACING TOOLS
shape of the bridge. There are tools in Houdini that will take profile curves and
build surfaces. Those input curves can be either bezier,
polygonal or NURBS or a mixture of them.
Revolve- Create geometry by revolving a profile curve around an
axis. There is a handle available for tweaking the results.
Skin - Take a series of profile curves and turn them into a surface.
HOUDINI FOUNDATIONS

Rails - Copy one or more profile curves along two or more rail
curves, then Skin the results to get a surface.

PolySplit/Edge Loop/Knife - These tools let you split polygons to

add more detail to your model.
PolyExpand 2D - Take curves and edges that sit on a 2D plane and
creates geometry based on a desired offset value.
PolyReduce - Create different levels of detail by reducing the
number of polygons while preserving quads and UVs.
PointWeld- Interactively snaps groups of points to another target
point, and merges them.

16
 BOOLEANS COPY TO POINTS + SCATTER & ALIGN
Subtract, Union or Intersect geometry using the Boolean tool. A typical Houdini workflow is to Scatter & Align points on a
This node can handle very complex topologies and can be surface then Copy to Points. Attributes for scaling and rotating the
used to break up a surface for destruction using rigid body objects can then be applied to create a more organic result. This is
dynamics. This often creates more realistic results compared to often used to create landscapes with trees and rocks.
the Voronoi-based Shatter node.

The Boolean tool can create output groups that you can use TOPOBUILD
to feed other nodes such as the Polybevel node. This way any Houdini has a Topobuild node that lets you draw polygons
updates you make to the boolean will update properly when it directly onto high-resolution geometry that you either scanned
feeds into the second node. or created in an application such as Pixologic’s Z-Brush. You
can create a cleaner topology for animation then bake the
DEFORM TOOLS details from the original model into a normal map.
While you can shape your geometry by editing points directly,
there are times that you need a more generalized approach.
The following nodes provide options for shaping your
geometry procedurally.
Bend - This node lets you set a capture range and direction then
bend, twist, taper and squish the encompassed geometry.

VOLUMES
Volumes allow you to store values for voxel, or three
dimensional pixels, in a space. These are often used to support
collisions when using dynamic tools or to create clouds. They
can also be used for modeling to combine multiple shapes into
a single volume which you then convert back to a surface.

Lattice - This builds a lattice around your geometry then lets you
edit points on the cage to reshape it. You can also use a custom cage.
Mountain - Apply a noise function to deform the surface to create
a random result. The points are actually being moved with this node.
Ripple - This node creates a ripple shape in your geometry.
Waves - This node adds noise functions to create a wave-like
pattern that animates over time. Perfect for creating realistic oceans.

GEOMETRY TYPES
MODELING TOOLS

Houdini supports a number of different geometry types including Primitives,

Polygons, NURBS and Beziers. You can Convert back and forth between them and
you can have more than one geometry type merged together in a single object.
Polygons models can be set up to display and render as Subdivision Surfaces using
PIXAR’s OpenSubdiv standard. Both Subdivisions and NURBS will render in Karma
and Mantra as perfectly smooth without relying on any tessellation settings.

17
 UVs & Textures
To make 2D maps fit properly onto 3D objects, UV coordinates are needed to define a flattened
view of your geometry. When you first create geometry in Houdini it will not have any UVs. Even
primitive objects do not have any built-in UVs. This means that you will need to add them at the
geometry level using one or more SOP nodes.

UV TEXTURE DISPLAY UV EDIT & DISTORTION

Since geometry in Houdini does not have UV texture To edit individual vertices or groups of vertices you will use
attributes set up by default, you need to add them either the UVedit or UVtransform nodes. The UVedit node lets
using UV tools. Once you have UVs set up, you will see you perform many edits using a single node while UVtransform
a texture grid on your geometry because Show UV Texture is allows one edit per node which can provide a more procedural
on in the Display Options bar. You can toggle it off if you don’t result. You can use Display > UV Distortion from the UV
want to see the UV texture or change it to a color texture. Viewport menu to see whether you are editing too much.

UV PROJECT UV LAYOUT
This node let you assign UVs using one of several projection UV layout will let you create UV islands and pack them into
techniques. Once you choose your projection type you can UV space as efficiently as possible. This lets you maximize how
initialize the projection to match your object. This may invert much of a texture is being used on your geometry which is
the UVs and you will need to set a rotate x value of -90 instead important when optimizing for both rendering and gameplay.
of 90. Above you can see an Orthographic projection and
below you can see a Toroidal projection.

You can use the region handle to put a the UV layout into a
particular part of your UV space. Subsequent layouts can then
UV FLATTEN work around this layout using the Pack Islands in Cavities option.
UV Flatten unwraps your geometry based on predefined UDIMS
boundaries created using selected edges or edge groups. The
In addition to working with a Single UV tile, you can use
results can then be tweaked by pinning points in the UV view
UDIMs to spread your UVs over many tiles. This technique
and adjusting the islands to get the look you want.
lets you create more detailed texture maps because your UV
islands are not packed in too tight. Properly numbered texture
HOUDINI FOUNDATIONS

maps will then be assigned to the appropriate tile.

Statue scan from threedscans.com

18
UV ATTRIBUTES UV VIEWPORT MENU
Earlier, you learned how attributes can be assigned to The UV viewport menu lets you display UVs based on the
geometry that carry important information down the line. UVs UV attribute. You can also use this menu to figure out the
are vertex attributes that let you wrap texture maps around background image which can either be the default UV grid or a
your model and are also carried down your network. texture map pulled from an assigned material.
These attributes are visualized in the UV viewport and
analyzed in the geometry spreadsheet. These attributes work
with various SOP nodes including the attribute wrangle node
which lets advanced TDs manage their UVs using scripting.

UV SETS
You can create more than one UV set on the same geometry.
When you use the UV nodes, you can set the UV attribute. By
default this is uv but you could create a uv2 to create a second
set. These different UV sets are used when you assign textures in This menu also has display options for displaying UV Overlap,
VOPs so that different texture maps use different UV attributes. UV Backfaces and UV Distortion. These can be helpful when
evaluating your UVs to decide if more tweaking is required.

ATTRIBUTE TRANSFER
One of the SOP nodes that can be used to manage attributes is
Attribute Transfer which lets you take the UV attributes from
one piece of geometry and transfer them based on proximity
to another.
This can be useful when the topology of a model has changed
but you want to preserve some of the work you did creating
UVs for the original model.

In the two images shown here, the first uses a toroidal

projection and is assigned to a uv UV attribute while the
second uses a planar projection and is assigned to a uv2 UV
attribute. These UV attributes can have any name, for instance,
it could be bob instead of uv2.

SCENE VIEW | UVS UV menu - when you are in

Current Tool - At the top of the UV view, this menu gives
the Scene View, you will find you a number of different
the selected node when the options for working with UVs.
Handle tool is active. View menu - Use this menu
Background - The to choose the UV view for this
background of the main tile viewport.
can be set using options Layout Handle - This handle
found in the UV menu. is part of the UV Layout node
Outside Main Tile - Polygons and lets you focus the UVs to
a certain part of the tile.
UVS & TEXTURES

positioned in the area

outside the main tile would Pack Around - The UV layout
overlap the same area on the will pack your UVs around any
main tile because the texture existing geometry that has
repeats unless you are using already had its UVs set.
UDIMS which cover more
than one tile.

19
 LookDev: Shaders & Materials
To render objects in your scene, you must assign materials, also known as shaders, to your geometry.
In Houdini, these materials and shaders are created in the mat/vex builder networks. The ability to
build up materials using nodes is a powerful tool when defining the look of a shot.

Houdini organizes different types of nodes into network types PRINCIPLED SHADER
and for materials you will use the /mat network type. This is
In the Material Palette, you will find the principled shader,
where you can set up VEX operators for Karma and Mantra
a material based on the Disney “principled” BRDF by Brent
or Material X for Karma. Material X is an open standard, that
Burley. This shader is “principled” rather than physical in order
originated at Lucasfilm,® which is used or the transfer of look-
to make it easier for artists to work with.
development content between applications and renderers.
The Principled Shader has been pre-built to include the ability
to assign textures directly to parameters such as base color,
bumps, normals, displacements and more. The texture maps
you assign will be displayed in the viewport and you can
achieve a wide variety of looks right out of the box.
You can extend this material by feeding it with other VOPs but
that isn’t necessary in all cases. Many of the materials found in
the gallery are variations on this shader.

THE MATERIAL PALETTE

You can add VEX-based materials using the Material Palette
then assign them to objects by clicking and dragging. This
pane has a workspace for managing the materials in your
scene which is organized into tabs that represent different
subnetworks such as Material Library LOPS. Use the tab key
to add Material X shaders to material networks. Once they are
PRINCIPLED SHADER CORE
in place they will show up in the Material Palette.
The Principled Shader Core node sits inside the Principled
ASSIGNING MATERIALS IN LOPS Shader and contains the main features of the shading model
but doesn’t have all of the texturing features built-in.
To assign a material in the Solaris LOP context, first create a
Material Library LOP which contains a /mat network. You can To build a robust shader from scratch using this node, you
assign the material using this node or use an Assign Material would need to add VOP nodes using Houdini’s shader building
LOP later in the chain. The Material Palette will let you drag tools. You can accomplish this by either wiring nodes together
from the gallery to the library or in LOPS you can use the in the Node view or adding them using the Shader FX menus.
arrow button to access a list of materials in your scene graph.

MATERIAL PALETTE Update Material Icon - To

Materials in Gallery - The update all material icons
Materials listed here are saved you can click on this button
to disk in a gallery file. You or RMB-click on material to
can drag these into the scene do it one at a time.
area found on the right or onto Assign Material - If you
objects in the viewport. select an object in the
HOUDINI FOUNDATIONS

Material in Scene - This is where scene and the material in

you find materials that are part the palette then you can
of your scene file. They can be use this button to assign the
assigned to objects by dragging material.
from here into the viewport. Double Click to Edit - If you
Material Library LOP - Materials double-click on any of these
set up in the LOP context can be materials, you will jump to a
placed into Material Libraries - Node View where you can
you can drag materials into here make edits at the /mat level.
from the gallery.

20
LAYERING MATERIALS SHADER FX MENU
You can layer materials to create a unique look for your object. While working with material VOPs, you can add nodes in the
Using a Layer Mix node, you can combine two different Network view and wire them together or you can use the
materials into a single look. For instance, a shiny metal material Shader FX Menu by clicking on the icon at the far right of each
and a matte rust material can be layered using this technique. parameter. This menu lets you focus on the parameter that you
You can then texture an alpha channel and you can choose to want to work with and create the node in context.
mix the surface, the displacement or both.

MATERIAL BUILDER In the Parameter Pane, you can see what type of connection
If you want to turn your layered nodes into a new material each parameter has by looking at the icon on the far right:
then you can select them and choose Edit > Collapse Selected
ƒ No Connection
into Material. This puts the nodes inside a material builder
ƒ Parameter Node
where you can continue to tweak it. At this level, there are
ƒ Connected with other Nodes
output and collect nodes to make the network work efficiently.
ƒ Hidden Connection

PROCEDURAL MATERIAL ASSIGNMENT

When working in production with lots of data, it is often necessary
to take a procedural approach to assigning materials. With Solaris
and Karma, you can accomplish this using nodes such as the Assign
Material LOP or Material Variation LOP.
If you are working with Mantra, materials can be assigned to the
object using the Data Tree panel which gives you access to
Material stylesheets. Stylesheets let you use rules to assign
materials and textures to large groups of objects.
MATERIALS AS DIGITAL ASSETS
You can make materials even more efficient by saving them out
as Houdini Digital Assets. In the Asset Properties pane, you
can go to the Save tab and choose Save Cached Code so that
the material is pre-compiled when you render with Mantra. You
can also load texture maps into the digital asset then reference
them from inside the asset file. Turning materials into HDA’s
can make it easier to share with your team.

SHADER BUILDING Principled Material - This is a

Node Path - This will confirm your typical material which can be
assigned on its own or fed
LOOkDEV: SHADERS & MATERIALS
current path and help you navigate
into and out of the material network. into the Layer mix.

VOP Nodes - In the material context, Layer Mix - The Layer output
you can start with material nodes on the materials can be fed
then wire in VOP nodes to customize into this mix node. You can
the texturing of the material. Once assign this to your geometry.
you are finished you can choose Material Flag- If you want
to collapse all of this back into a the layer mix to show up
material builder node. in the material palette then
Node Connectors - You can add check this flag.
nodes to this area using the Shader Alpha - Here the layer mix
FX menu which can be accessed by node’s alpha is being fed by
MMB-clicking on the dot. RMB- VOP nodes to create the
clicking on the dot gives you a full alpha mask for the two
node menu. layered materials.

21
 Solaris: Layout
Solaris is the context in Houdini dedicated to Lookdev, Layout and Lighting that has USD at its core.
Objects and geometry brought into this context become USD and you can use specialized nodes for
positioning objects, instancing geometry and managing shot layout.

SOLARIS: LOPS SCENE IMPORT: OBJECTS TO LOPS

The Solaris environment can be found in the /stage network For artists used to working at Houdini’s Object level for layout
or by creating a LOP network. Here you will place nodes for and lighting, the Scene Import LOP node makes it easy to
bringing in geometry, assigning materials and adding lights bring your geometry, lights and cameras into LOPS to render.
and cameras. These nodes let you create sequences and shots If your goal is to create a controlled USD scene graph then you
using shared assets and procedural edit nodes to customize may want to take another approach but for quick access to
settings for each shot. Karma and other renderers the Scene Import node works fine.

PREPARING ASSETS FOR USD

Another approach is for you to configure props using either
a Component Builder LOP network or the USD Export SOP.
This would be geometry and materials set up on a prop-by-
prop basis then exported as USD for use in the layout stage.

At its core, the Solaris environment converts everything into

USD [Universal Scene Description] which is an open source
initiative created by PIXAR. The Solaris/LOPS context allows
you to work with USD natively using procedural nodes to
manage USD concepts such as references, payloads, layers,
collections, variants and level of detail.

For some of the props, you will use variants to create differ-
ent variations that can be chosen during layout. There are LOP
nodes that let you set up this USD concept.
Once you have the props set up as USD, the files can then
be brought into a larger layout scene file using Sublayer or
LOP networks can then be rendered using either Karma or Reference LOPS then Edit LOPs can be used to position the
other Hydra-compatible renderers. Hydra is the technology props. If the assets are prepped properly then they will come in
that lets you render your USD to the viewport for interactive with geometry, materials and variants set up and ready.
exploration or to disk for final renderings.

SOLARIS DESKTOP
Stage Manager - This
Stage - The Stage view provides
node lets you load USD
a place for viewing your layout
files to be positioned on
and lighting and to interact with
the stage.
primitives, lights and cameras to
set them up properly. LOP Nodes - All your
actions in LOPS are
Viewport Render - On the stage,
HOUDINI FOUNDATIONS

accomplished using
you can render using Houdini GL
procedural nodes that
or Storm the USD GL solution.
make it easy to step
You can also render interactively
back and make changes.
to karma and Hydra compliant
3rd party renderers. Scene Graph Details -
As you select items in
Scene Graph - The USD
the Scene Graph you
structure offers a renderable
can check for details
scene graph that can be
that inform you of its
inspected using this panel.
status in the pipeline.

22
STAGE MANAGER INSTANCING
The Stage Manager is designed to be a one-stop node for There is an instancing solution in USD that can be utilized in LOPS.
referencing USD assets from disk, transforming them in 3d The Instance LOP lets you input one or more objects that will be
space, and adjusting your scene hierarchy. This does involve distributed to points set up inside the LOP. These points can be
a flattening of the inputting layers which will block up-stream created by importing geometry from your scene then scattering
changes from coming through. This lack of flexibility is points on surfaces extracted from the model.
balanced by the rapid setup possibilities of this node. Materials can be assigned to the instances using a variety of
methods including the Material Variation LOP which also offers
per-geometry render properties.

PHYSICALLY-BASED EDITS
Once you have all of the props loaded onto the Stage, you
can use the Edit LOP to begin moving them around. The edits You can also use the Layout Asset Browser and the Layout LOP to
become a separate non-destructive layer which keeps the use a brush workflow to place, edit and transform instance points
referenced assets intact in case you need to step back. that reference the assets you choose in the browser.

EXPORTING USD
At various points in your LOP graph, you can inspect the USD
code by RMB-clicking on the node and selecting LOP Actions
> Inspect Active Layer. You can also inspect the Flattened
Stage. When you export to USD, you can break out all the
layers or save it as a single flattened graph.

In the Edit LOP, there is a Use Physics option that lets you
leverage the rigid body capabilities of Houdini to detect
collisions and make it easier to position objects in a realistic
manner. This lets the artist achieve a natural organic look while
working interactively in the 3D view.

SCENE GRAPH
Stage - This is the top level Draw Mode - Here you
network context for creating can change the display
LOPS. You can also do all of this of any element in the path
work in LOP Networks. to either Full Geometry,
Scene Graph Path - These Bounding Box or Textured
represent the USD layers and Cards.
sublayers that define the look of Display Options- Here
the stage. These are most likely you can set the visibility
different USD files on disk being or activation of each layer.
Solaris: Layout

referenced into a shot. It is not possible to delete

Primitive Type - Each primitive anything from the scene
comes with a type or schema graph therefore hiding or
that defines how it works. deactivating is necessary.
This can help you determine Variants- If a Layer has
the contribution each layer is variants then the chosen
making to the stage. one is displayed here.

23
 Solaris: Cameras and Lights
Before you render a shot, you need to look through a camera and light your scene. There are lights
and cameras in the Solaris/LOPS context which is especially designed for layout, lookdev, lighting and
rendering with Karma or at the object level for rendering with Mantra.

CAMERA CAMERA PROPERTIES

Cameras can be accessed from the LOP Lights and Cameras The Cameras in the LOP Context have key properties that
shelf. Press Alt-click on the shelf tool to convert your current determine how the camera will generate images.
view into a camera view. If you create the camera node in the
Network view, click on the No Cam menu in the top left of the View Tab
viewport to look through it.
ƒ Projection - Choose between perspective or orthogonal projection.
ƒ Focal Length - Determines the length of the lens. Smaller values
create wider shots and higher values create long shots.
ƒ Horizontal/Vertical Aperture - Aperture is a gate that controls how
much light goes into your camera.

Sampling Tab

To adjust the camera, you can use the Camera handles either ƒ Shutter Open/Close - Determines how long the shutter is open
from another view or while looking through the camera. There which has an effect on motion blur.
is a lock camera to view button in the Display Options bar ƒ Focus Distance - Distance from the camera to the focal plane.
that lets you use the View tools to reposition the camera. Determines which objects are in focus when using Depth of Field.
ƒ F-Stop - Lens aperture. Default is 0, which turns off focusing.

Press Shift + F to show a visualization of the focal plane and

how it intersects geometry. When looking through a camera,
you can Shift-click or drag on a surface to move the focal
plane to intersect that point. Otherwise, you can use a handle
on the focal plane to move it and set your Depth of Field.

Shift-click
on Surface

There is a lock camera to view button in the Display

Options bar that lets you use the View tools to reposition
the camera. Just make sure that you don’t leave this on later
otherwise your camera view may get changed by a simple view SHOW FOCAL PLANE
change. Once you like a camera view, you may want to lock its [Shift- F]
transform values to avoid losing it to a view change.

LIGHTING SETUP Primitive Path - This sets

Viewport Rendering - To where your light will be in
make lighting decisions, it is the USD scene graph.
important be able to render in Light Type - You can choose
the viewport using either karma your light type from this
or other renderers such as menu. The Light LOP gives
RenderMan or Arnold. you access to all the light
HOUDINI FOUNDATIONS

You can also use HoudiniGL but types and you can switch
this is not as effective as one of between them.
the main renderers. Light Parameters - There
Light Handles - You can step are a range of parameters
back and manipulate lights in for controlling light
the view in a similar manner to properties such as cone
cameras or you can use special angle and intensity.
controls to set lights right in the Light LOP Nodes - Every
camera view. light is added into the
network as a LOP node.

24
 LIGHTS LIGHT LINKER
Lights can also be accessed from the shelf and have similar Linking lights to specific objects is a great way to control the
handles to help you position them. There are a number of lighting of the shot and this can be accomplished in Solaris
different light types in Houdini to choose from. using a Light Linker LOP. This node includes an interface for
Point Light - Emits light from a point in all directions and is similar making the object/light connections that you need.
to a light bulb.
Spot Light - Radiates a cone shaped beam of light from a point in a
certain direction.
Area Light- Automatically distributes a number of light sources over
a specified area. There are five area light shapes to choose from: Line,
Tube, Grid, Disk an Sphere.
Geometry Light- This object emits light into the scene using a
geometry object’s surface shader for the colors of the emitted light.
Distant Light- Emits parallel rays of light, which are similar to the
rays of the sun.
Environment Light- Casts light into the scene as if from a
surrounding hemisphere or sphere.
You can use collections of lights to apply the linking more
efficiently using rules that define the interactions between
LIGHTING WHILE IN THE CAMERA VIEW
primitives and lights.
When you have a Light or Light Mixer LOP selected and
displayed, you can set many of its properties while looking INSTANCING LIGHTS
through the camera. You can use the Specular [Shift-S],
Within the Solaris/LOP context, you can use Houdini’s
Diffuse [Shift-D] or Shadow [Shift-F] options that let you click
procedural geometry nodes to create points then Instance
on surfaces in the scene to set up the light.
Lights to those points. You can then add attributes to the
SHADOW MODE points to create effects such as a rotating intensity you might
[Shift- F] find in an old school marquee. This approach makes it easier
LMB-click to place Pivot to set up the lights and even easier to add effects and make
changes to meet the needs of the shot.

Shift -LMB-click
to place Target

You can then use Ctrl-drag to change the distance of the light
from your shot and Ctrl-Shift-drag to change the brightness.
Doing this in the viewport lets you stay focused on the shot
you are working on instead of pulling away to drag on handles.

LIGHT MIXER
Light List - This list shows all the
lights feeding into the mixer.
SOLARIS: CAMERAS AND LIGHTS

Collection - You can organize lights

into collections so that they work as
a group in the mixer.
Solo Lights - Click on the star icon to
solor the light or the collection.
Intensity Slider - The first slider gives
you control over the intensity of your
lights or light collections.
Exposure Slider - The second slider
gives you control over exposure
which offers a tweak to the intensity.
Light Color - Click here to tint the
light or the collection of lights.

25
 Rendering
When you render a shot, you are digitally photographing or filming your 3D objects using cameras
and lights to generate either an image or a sequence of images. Game artists may also use rendering
to render game cinematics or to bake textures from a high resolution to a low resolution object.

KARMA 3RD PARTY RENDERERS

Karma is a physically-based ray-tracer built to work with USD With its support for USD, Solaris makes it possible to render to
files in the Solaris/LOPS context. It works on the CPU and other Hydra delegates such as RenderMan, Autodesk Arnold,
includes features such as adaptive tesselation for displacement V-Ray, Maxon RedShift and AMD ProRender.
and subdivision surfaces, multi-segment motion blur,
instancing, hair & fur and strong volume rendering support.
Karma works with Hydra,the USD imaging framework. This
makes it available in the Viewport for interactive updates or
you can render to disk using a Karma node.

Karma works with shaders created using VEX, USD preview

and Material X.
VIEWPORT RENDERING
KARMA XPU One of the key benefits of the Karma renderer is the ability
Houdini 19.5 includes a beta version of the Karma XPU render to use it in the perspective view. By choosing Karma, you get
engine. This hybrid GPU/CPU renderer is being released in interactive rendering that helps you make lighting and lookdev
Alpha. Many features are still under development and this decisions within the Solaris/LOPS context.
engine is for testing purposes only. You can choose XPU in the
Scene view’s Display Options or in the Karma node.

Karma XPU works with USD preview shaders and Material X

but does not work with VEX.

VIEWPORT | KARMA
Render Settings - To render to disk,
you can use the Karma LOP to define your
render settings. Here you can set a path
for rendering to disk, camera
settings and more.
HOUDINI FOUNDATIONS

Karma LOP - Add the Karma LOP to the

end of a Solaris node network. Different
versions of this LOP can be used to set up
unique results such as test renderings or
final shots.
MPlay - You can either render from both
Karma and Mantra to MPlay directly or
render to disk then open up with MPlay
to review the results.

26
RENDER SETTINGS OUTPUT NODES
When rendering the Stage, you use the viewport render To render out a shot, you will need to create a render output
settings. To create the final look of your rendering, the Karma node. You can do this by choosing Render > Create Render
LOP can be used to set a frame range, camera resolution, Node > Mantra - PBR.
denoiser and higher quality render settings.

You can also add a Karma ROP to the output network using
the tab key. This has a LOP network inside it that grabs all the
visible objects from the object level.
You can use ROP nodes to render to disk or to Mplay. These
nodes contain many of the parameters that you need to
RENDER GALLERY control the final image such as sampling, noise level and overall
render quality.
The Render Gallery lets you take Snapshots of your progress.
Each snapshot contains all the settings of the look and at It is possible to have a different ROP per object or group of
any time the scene can be reverted to match the snapshot. objects and you can use them to select Mattes and Phantom
Snapshots can then be labelled and filtered for easy access. objects. You can create ROP dependencies by wiring different
nodes together. If you press the render button on the last node
in the chain then all the other nodes will render first.

RENDER OUTPUTS | AOVS

On the ROP, you will find controls for setting up Image Planes to
create render layers for Direct lighting, Indirect lighting, Shadows,
Depth etc. Both Karma and Mantra can render out these passes
which can be loaded in Houdini’s compositing context, COPS, or
in an external compositor such as Nuke.

MANTRA
Mantra is the Houdini renderer developed before Solaris was The Background Plate LOP can be used to matte objects in
introduced. It is a physically-based rendering engine that is order to leave holes in the scene to make the background
deeply integrated with highly efficient rendering of geometry, visible. This geometry can take shadows and contribute to
instances and volumes but it does not work in Solaris/LOPS. reflections to create a more realistic fit for your objects.

MPLAY
Render Time - Since you are
MPlay lets you view images
sometimes rendering directly
rendered in Karma, Mantra or
to this view, render time info
other renderers.
will be displayed.
Main Menu - This menu lets
View Options - MMB-drag to
you load images or sequences
pan and RMB-drag to zoom in
of images to preview them.
and out of the image.
You can also save them out to
another format if needed. Channels - You can click on
these buttons to focus on red,
Render Layers - This menu lets
green, blue or alpha channels
you display different render
or to see them combined.
layers such as color, normal,
Rendering

diffuse_direct or reflect-direct. Gamma Settings - Set the

brightness, contrast and
Playbar - If you have loaded up
gamma for the viewport. By
a sequence of images then you
default a gamma of 2.2 is used
can use these controls to play
to support a linear workflow.
and scrub through the sequence.

27
 Time & Motion
Animation involves changes happening over time. Whether it is an object’s position, shape or color,
once changes occur over time, you are animating. Houdini has a variety of tools for keyframe-based
workflows in addition to Motion FX & CHOPs for more advanced manipulation of time and motion.

SETTING KEYFRAMES The Playbar can also be used to edit keyframes. You can RMB-
click on the frame range in the Playbar to access options such
Keyframes let you set specific parameter values at a specific
as Cut, Copy and Paste of keys along with special pasting such
points in time. As these values change, the objects in your
as Replace, Cycle, Repeat and Stretch. All these options also
scene are animated. You can then use animation curves to
have their own hotkeys which you can find on the menu. You
determine the in-between quality of the motion. Here are a
can accomplish a lot working in the Playbar before moving to
few main hotkeys used to set keyframes on your objects while
the Animation Editor.
working in the scene view:
ƒ Set keyframe k CHANNELS
ƒ Toggle Autokey Alt + k When you set a keyframe or display animation curves in the
ƒ key Handle Ctrl + k Animation Editor, you are working with channels.
ƒ Key Translate Shift + T
If you select an object that has keyframed channels then they
ƒ Key Rotate Shift +R
become active and the keyframes are loaded into the Playbar
ƒ Key Scale Shift + E
or Animation Editor. If you deselect the object then the
You can also set keyframes in the parameter pane by channels will no longer be selected unless you pin them.
pressing the Alt key and clicking on either a parameter You can pin channels using the Channel List which is available
name or parameter field or by RMB-clicking on a parameter on the right side of the Playbar, the left side of the Animation
and selecting Keyframes > Set Keyframe. This lets you set Editor or in the Channel List pane. In this list, you can select
keyframes one parameter at a time one or more channels to refine which of them you are
keyframing and editing.
THE PLAYBAR
The Playbar is at the bottom of the main workspace and lets you CHANNEL LIST PANE
playback and scrub through your animations. Time is measured in The Channel List pane, lets you work with Channel Groups,
frames with a default rate of 24 frames per second. Animation Layers and Active Channels. You can use the list
At the left are the playback controls. Here are some hotkeys to create groups of channels so that they can be more easily
for quickly setting up playback and moving through time. accessed. You can also use the groups to pin channels so that
they are available even if you select different objects. This is a
ƒ Play Forward 
useful pane if you are setting keyframes on characters.
ƒ Play Back 
ƒ Next Frame  FLIPBOOK
ƒ Previous Frame 
As you animate your scene, you will want to preview the
ƒ First Frame Ctrl +  results in motion. The Flipbook tool found on the toolbar on
ƒ Next keyframe Ctrl +  the left side of the Scene view lets you capture frames from
ƒ Previous keyframe Ctrl +  the viewport then playback the results as a movie in realtime.

PLAYBAR
The Playbar is where you scrub through time and create and edit keyframes. The Playbar can also be used to
make quick edits, while the Animation Editor is used for more comprehensive refinements.
HOUDINI FOUNDATIONS

Playback Controls Current Time Frame Range Edit Keys Set Key
These controls let you The current time is The overall range is defined As you set keys, they are The Set Key button will
quickly play, pause, and shown in the field and on by the Animation controls shown in the Playbar. Press set keyframes when you
move to next key. Below the black marker in the button at the far left. Then the shift key to select keys click here. Click on the
that are buttons for frame range. The marker the visible range can be with the LMB and then edit small arrow to bring up
Animation options and can be used to scrub reduced using the handles the keys by dragging with a menu of options such
Real Time playback. through the Playbar. at the bottom of the range. the MMB. as Auto key.

28
ONION SKINNING MOTION FX
Onion Skinning lets you display a ghosted version of your object While keyframes and animation curves are stored in the
on the frames before and after the current frame. Turn it on using parameters of your nodes, you can also use channel operators
the Misc tab on your object while the onion skinning options such (CHOPs) for a more procedural node-based approach to
as Frames Before, Frame After and Frame Increment can be found working with motion.
in the viewport’s Display panel [d] under the Scene tab. The easiest way to create a channel operator is to RMB-click
on a parameter and select from the Motion FX sub-menu. You
can also apply these effects to channel groups when working
in the Channel List.

MOTION PATH HANDLES

When using the Pose tool to animate, you can click on the
Motion Path option to bring up handles that let you see the
animation of the selected object over time. You can also use
the handle to manipulate the shape of the motion.

Motion Path Handle

Onion Skinning Motion FX can be applied to keyframed motion which is

extracted and stored in a Channel CHOP. You can then apply
effects such as cycle, noise, smooth, limit or lag to the existing
Tangent Handles motion. On the Constraints shelf, you have tools which let you
have one parameter either look at, lag or jiggle behind another.

ANIMATION EDITOR
Selected channels are loaded into the animation editor where
they are represented as keyframes and animation curves, or as
a spreadsheet or a dope sheet. The keyframes can be selected
and edited and the curves can be shaped using tangent handles.
The curves define the motion in-between the keyframes and
play a key role in defining the quality of the motion.
While working with channels, you can view the keyframes and
animation curves using these hotkeys:
ƒ View All/Home H
ƒ Pan MMB This non-linear approach to working with motion offers a
ƒ Zoom RMB unique way of working that can be very flexible.

ANIMATION EDITOR Key Handles - Move the key

Editor Options - This editor back and forth in time using
can be shown as a graph, a the vertical bar or edit its
dope sheet or a table. value using the box.

Channel Groups - This area Tangent Handles - They

of the graph shows channel define the tangents coming
groups which make it easier in and out of a keyframe to
to select and pin channels. help you shape the curves.

Animation Layers - This Curve - The animation

area lets you layer different curves determines the
Time & Motion

channels on top of each other motion in-between the

to create multiple iterations. keyframes, which defines
the quality of the motion.
Channel List - Channels on
selected objects show up in Curve Functions - These
this area. You can then select let you set various display
the names of channels you options for the animation
want to see in the graph. curves and handles.

29
 Character Rigging & FX
Houdini includes a wide range of rigging tools for creating characters and creatures that can then be
wrapped into Houdini Digital Assets to be handed off to animators. Houdini also has Character FX
tools such as hair, fur, muscles, cloth and crowds to enhance the look of your characters.

BONES DIGITAL ASSET CHARACTERS

In Houdini, you draw and edit bones using the Bones tool and To rig a Houdini character and share it with the animation
the Bones from Curve tool found on the Rigging shelf. Each team, you will need to wrap up the bones, geometry and
bone chain is made up of a chain root and bones. Whereas materials into a Houdini Digital Asset.
other 3D apps are joint-based, Houdini uses Bone nodes that This creates a file on disk that can be easily referenced by
have parameters for Length and Rest Angle. animators into multiple shots. Handles and key parameters
You can also use the Bones tool to add inverse kinematics to are made available at the top level so that Animators can set
the chain which adds an end effector and in some cases a twist keyframes without worrying about the inner workings of the
effector. The kinematics are driven by Channel Operator or rig. You can also save Pose Library and Character picker setups
CHOP nodes which exist in their own subnetwork. into the asset for quick access.

CAPTURING GEOMETRY
The character’s geometry can then be captured to the bones
to create the deformations needed to convey realistic motion.
Changes made to any part of that character are saved into
Houdini bones include Capture Regions that you can set up
the asset and all shots will be updated. This creates a robust
to encompass your geometry while creating the right overlap
character pipeline that is easy to manage.
at the joints. This process results in weighted attributes being
assigned to the points that get fed into a Deform node that KINEFX
controls the geometry when the bones are moved and rotated.
KineFX is a character toolset designed to provide a procedural
foundation for retargeting, motion editing, and in future
releases, rigging and animation. Set in the geometry context,
these new workflows make rigging a fast, plug-and-play
experience with unlimited flexibility and caching capabilities.
Implemented in the geometry [SOP] context, KineFX treats
joints as regular point geometry with edge connections
defining the rig hierarchy. You can bring rigs in from the object
level of Houdini or import FBX characters.

At first, captured geometry might not get you exactly the look
HOUDINI FOUNDATIONS

you want, therefore you would use various tools to Edit and
Paint Capture Weights. This lets you smooth out the weighting
at the joints to create more realistic bending. You can also
smooth out the effect of point deformations using a Delta
Mush node that wires into the Deform node.
A new technique called Bone Capture Biharmonic allows you
to capture geometry without requiring extensive point weights
to get a desirable look at the joints. This method sets up
biharmonic functions on a tetrahedral mesh to create a much
more holistic solution.

30
CHANNEL GROUPS CHARACTER FX | MUSCLES & SKIN
When you animate in Houdini, channels that are scoped can Houdini lets you add muscles to an animated creature and
be keyframed and displayed in the Animation Editor. Generally then apply them as a skin deformer without requiring any
these channels are based on your current selection. You can simulation. Start by creating simple muscle forms using the
also put together Channel Groups that let you scope and pin Muscle nodes in the geometry context.
down channels to assist with keyframing. You then adjust the muscle’s shape and placement, attach it to
When you have a character set up as a Digital Asset, you can your character rig then enable automatic secondary animation
click on its icon at the top left of the parameter pane and or jiggle. Houdini’s muscle system has been designed to serve
choose Parameters and Channels > Create Nested Channel both FEM (dynamics simulations) and non-FEM (skin deformer)
Groups to create groups using the asset’s folder hierarchy as a workflows while using a unified set of Digital Assets.
guide. A well designed character asset makes this easy.

CROWD SIMULATIONS
CHARACTER FX | HAIR & FUR A crowd simulation begins with agents that are made up of a
Houdini has a hair and fur toolset that you can use to setup character skeleton, skin geometry, and animation clips. These
and groom your character starting with the Add Hair tool. are assigned to points, where simple rules can combine to
These tools also let you work with guide hairs and then create complex behaviors and the agents can interact with
animate them using wire sims to create added realism. other dynamic elements. For example, an agent might be
struck by a passing car and become a rag-doll or a crowd might
be triggered to react to an action on the field.
Andriy Bilichenko

CHARACTER PICKER POSE LIBRARY

This pane lets you create an The Pose Library pane lets
interface for choosing parts of you capture poses and clips
your rig. This can then be saved from your character for future
into a file that you add to your reference. You can go to a frame
Digital Asset file on disk. in your Playbar and apply the
Tabs - Set up multiple tabs for pose by clicking on it here.
different areas of the body such Pose - These save out a pose
as hands, feet or face. taken from a single frame. All of
Character Rigging & FX

Controls - Place markers for the the parameter settings at that

different handles on your rig pose would be applied to the
then add text or color to help character in your current scene.
you distinguish between them. You would then use this to
interpolate to another pose.
Background Image - Add a
visual representation of your Clip - A clip contains a set of
character to properly associate keyframes for a longer time
the markers with the parts of period. These might be a walk
the body. cycle or a particular movement
like a back clip.

31
 Dynamic Simulations
Whether you are creating Bullet Rigid Body destruction, Pyro FX fire and smoke, Vellum
Soft Bodies or FLIP fluids, Houdini lets you work in an integrated dynamics environment.
Different solvers know how to talk to each other to allow for more directable results.

SHELF TOOLS OPENCL

Setting up dynamic simulations involve a network of nodes You can use the GPU for faster sim times using OpenCL on
in the Dynamic or DOPS context, as well as nodes at the solvers such as the POP Grain node, the Pyro solver (Advanced
Geometry or SOPS context. It is always a good idea to use the tab) and the FLIP solver (Volume Motion > Solver tab).
shelf tools because they will add all of these nodes for you and
reduce the number of clicks needed to set up a sim. You can FORCES
then dive into the networks to explore all of the nodes. To create dynamic motion, forces are needed to “get the ball
rolling.” The most basic of forces is gravity although other
external forces such as fans, fluids and magnets can also play a
role in initiating motion in your simulation.
Gravity Force A downward force on objects which works as if they
were inside a gravity field.
Drag Force - Applies force and torque to oppose an object’s existing
motion to slow it down or dampen its momentum.
Uniform Force- A precise amount of force and torque that can be
augmented by a noise DOP to add turbulence.
Fan Force- Applies a cone-shaped force on objects.
The shelf tools are great for setting up groups of nodes Fluid Force- Deform soft bodies such as cloth or wires with fluids.
automatically. Seeing how these shelf-built networks are put Wind Force- A pushing force that will increase the velocity of
together can be very useful later if you choose to set up DOP objects up to but not beyond its own speed.
networks from scratch. Magnet Force- Attracts or repels objects using metaballs to
represent a force field.
DYNAMIC SOLVERS Vortex Force- Creates a vortex-like force that causes objects to
At the center of any simulation is the dynamic solver. It is the orbit around a curve much like objects around a tornado.
brain of the simulation and takes all of the dynamic objects,
forces and collision objects and integrates them to create the DYNAMIC OBJECTS
final result. The shelf tools put these solvers into a Dynamic When you select an object and use a shelf tool to add it to
Network and wire up the nodes for you. your simulation, Houdini creates a Dynamic Object that uses
Rigid Body Solver- Simulate rigid objects falling and colliding using the object’s geometry and adds dynamic properties such as
the efficient Bullet solver or Houdini’s built-in solver. density, friction, and bounciness.
Static Solver- For situations where you want objects to work as
collision geometry but not be affected by the simulation.
Flip Solver - This solver creates FLIP Fluid simulations to create
splashing and wave effects.
Whitewater Solver - After completing a FLIP solve, you can run this
solver to create foam, spray and bubbles.
Vellum Solver- A type of POP Solver that includes integrated
support for cloth, hair, grains, fluids and soft bodies such as balloons.
POP Solver- Used for particles and grains, this solver simulates a
wide range of different particle-based scenarios. Grain simulations can
ACTIVE VS STATIC
HOUDINI FOUNDATIONS

also be used for soft body and cloth-like simulations. Active dynamic objects are affected by forces and collisions
Wire Solver- You can use this solver for hair and fur or other wiry while Static objects are not. If you want to use animated or
objects such as the rigging of a ship or the branches of a tree. deforming geometry then you must define this on the dynamic
Finite Element Solver- Simulates the physics of continuous materials object using either the Initial Object Type menu or the Use
or solids as determined by tetrahedrons. This solver is used for muscles, Deforming Geometry checkbox.
soft body sims and destructions shots such as breaking wood.
Cloth Solver- Create cloth simulations that can collide with
deforming geometry such as a character.
SOP Solver- Use a SOP network to evolve an object’s shape over
time such as a wall being dented as it gets hit by objects.

32
COLLISIONS PLAYBAR FEEDBACK
Collision objects are also a big part of any simulation. You To start a simulation, you press Play in the Playbar. As the
can set up a Ground Plane to create a continuous surface for simulation progresses, the Playbar is highlighted to show how
collisions or use either static or deforming objects. much of the sim has been cached to memory. You can then
scrub through that area without re-simming.

On each Dynamic Object, there are also settings for displaying CACHING TO DISK
and optimizing the collision volume. While you often want Once you have completed a simulation, you can either lock
collisions to be as accurate as possible, you need to balance it down by saving out a sim file from within DOPs or more
that with the need for faster simulation times. commonly, write out the simulated geometry to a bgeo
sequence using the File Cache node. This will make it easier to
work with the results of a sim during the lighting and rendering
stages of production.

RIGID BODY CONSTRAINTS

On the Rigid Bodies shelf, you will find a number of constraints
that can also be used to influence a simulation. These include
Pin, Spring and Slider Constraints. You can also use Glue Objects
REALTIME FX FOR GAMES
when you set up a rigid body sim to hold objects together until In games, you need effects, such as explosions, to be optimized for
you either “loosen up” the glue or a collision occurs. real time in the game engine. Check out the SideFX Labs Tools to
learn more about converting different kinds of Houdini sims such
as rigid bodies, Pyro FX and Fluids into game ready art.

AUTODOPNETWORK
When you use a shelf tool to create
Dynamic Object - This node
a dynamic object, collision object
brings geometry into the
or force, the AutoDopNetwork is
DOPs environment and
created to combine all of the parts.
assigns basic properties.
Static Objects - These nodes set up
Rigid Body Solver - The solver
the properties of the ground plane
Dynamic Simulations

that generates the simulation

and a static collision object.
of the participating objects.
Static Solver - This solver keeps the
Forces - The nodes that
incoming objects still while dynamic
influence the dynamic objects
objects interact with them.
using forces such as gravity
Merge Node - Brings together or wind.
parts of a dynamic system. During
Output node - You can use
simulation nodes are evaluated
this node to output .sim files
up and down the chain so that
if you want to cache out the
everything interacts.
simulation.

33
 Cloud FX & Volumes
A big part of visual effects in Houdini is the use of volumetric data. In Houdini, volumes often
sit under the hood to help tools get the job done but it is a good idea to learn what they are
and, in time, learn how to work with them directly.

With volumes, you describe objects using voxels rather than The resulting network imports the cloud source then applies
points and polygons. A voxel is a 3D pixel, a cubic grid where these other nodes to create the cloud-like effect using Houdini
each voxel contains information that informs how the volume Volumes and VDBs. Houdini also comes with a Sky Rig tool
will be displayed which makes it ideal for wispy cloud-like that fills the sky with volumetric clouds.
shapes. The visual quality of a volume-based object is defined
by the resolution of that 3D grid. With more resolution,
the results are of a higher quality but performance may be
affected.

To create a cloudscape for a game engine such as Unreal you

can use a Sky rig in Houdini, then convert it to a mesh and use
it as a spawn surface. There is a tutorial on the SideFX website
by Andreas Glad that teaches this approach.

ISO OFFSET OPEN VDB

The Isooffset node found in the geometry context lets you “OpenVDB is an Academy Award-winning open-source C++
take any manifold Polygon geometry and construct a Houdini library comprising a suite of tools for the efficient storage and
Volume for Houdini to use. You can choose from a variety of manipulation of sparse volumetric data discretized on three-
different Output types to see the shape as fog or a tetra mesh. dimensional grids. It is developed and maintained by DreamWorks
Animation for use in volumetric applications typically encountered
in feature film production.” - openvdb.org

Original Geo Fog Tetra mesh: Cubes

CLOUD FX Houdini has a variety of OpenVDB Volume nodes available in

geometry [SOP] networks that convert geometry into volumes.
This toolset converts geometry into a cloud-like VDB volume
complete with lighting. The Cloud Rig tool can be useful VOLUMES UNDER THE HOOD
for shaping an individual cloud, or simply as a way to better
Many of the tools in Houdini make use of volumes under the
understand lower-level tools such as Cloud, Cloud Noise and
hood where you can’t see them. Here are some of the areas
Cloud light that all contribute to the final look.
where volumes are making contributions to your work.
HOUDINI FOUNDATIONS

ƒ Colliders - By default, volumes convert geometry into colliders for

dynamic simulations.
ƒ Simulation Fields - Volumes define fields such as density or velocity
that contribute to dynamic simulations.
ƒ Hair and Fur tools - These tools use volume data to assist with
grooming calculations.
ƒ Terrain - Heightfield tools use 2D volumes where each voxel
contains the height of the terrain at each grid point.
ƒ Rendering - Volumes create water depth and fog effects in Mantra.

34
 Terrain & Heightfields
Procedural terrain generation in Houdini is achieved using a collection of heightfield nodes that
let you layer shapes, add noise and run erosion simulations to define the look for your digital
landscapes. This a workflow that is similar to compositing, but you do all your work with 3D shapes.

Houdini provides a variety of geometry nodes for generating

and shaping terrain. These tools represent terrain using 2D
volumes where each voxel contains the height of the terrain at
that grid point, called heightfields. The data passing through a
geometry network can contain multiple heightfields. You can
access these tools using the Terrain desktop.
The Houdini viewport lets you visualize the 2D heightfield
as a 3D surface, and the mask field is displayed as a red tint
on the 3D surface. There is a dedicated Mantra procedural
for rendering heightfields and they can be used as collision
surfaces for dynamic simulations.

MASKS
The heightfield tools also use a secondary type of 2D volume,
where each voxel contains a mask layer. Most terrain nodes
take a second input that can contain a mask layer to control
which parts of the terrain the node will modify.
You can use a variety of different methods to create masks and
then use them to assist you as you add detail and shape the
ARON KAMOLZ terrain. You can also draw or paint masks onto the heightfield.

PATTERNS
After laying down a Heightfield node to define your base
resolution, the Heightfield Pattern node gives you access to a
number of starter shapes. You can set up linear, concentric, or
radial ramps, linear steps, radially symmetrical shapes such as
stars and voronoi cells.
These shapes can then be blurred and distorted to get shapes
that you can use to begin your terrain. You can also combine
and layer elements to achieve even more sophisticated results.
EROSION
The Heightfield Erode node uses rainfall, the erodibility of the
soil, and entrainment rates as variables to simulate erosion and
deposit buildup. This node works iteratively during playback.
It will appear to have no effect on the first frame. You need to
press play in the Playbar to watch as it sims the erosion.

EXPORT OPTIONS
TERRAIN & HEIGHTFIELDS

There are a couple of different ways to export your terrain for

NOISE use in another application such as a game engine. You can use
As you build up your terrain, you can then add noise to layer the Heightfield Output node to export height and/or mask
in a natural look. You can choose from a variety of different layers to disk as an image then bring these in as textures.
types of noise including, Perlin, Sinusoid, Worley and more. This You can also create Houdini Digital Assets that will open up in
adds realism to your terrain and by combining different shapes applications such as Unreal and Unity using the Houdini Engine
with different kinds of noise, you can achieve a wide variety of plug-ins. In the game engine, these Digital Assets will interact
hyper realistic results. with built-in Terrain tools.

35
 SideFX Labs
SideFX Labs is a collection of high level tools aimed to speed up artist and gamedev related
workflows in Houdini. There are a growing number of tools being developed that range
from Mesh Processing, Realtime FX, UVing, to creating Motion Vectors from simulations.

While all of Houdini can be used to generate content for film, Make Loop - Takes a mesh, points or volume that is animated and
TV and games, the Labs toolset addresses artist-specific tasks loops them.
that may not be available in Houdini today. These tools are Volume to Texture - The Volume Texture tool lets you write out a
developed separately from the regular Houdini development texture that can be used with Ryan Brucks’ volume plugin in UE4.
cycle and become available the moment they are ready for Flowmap - This utility tool sets up a flowmap template on your
testing. They can be downloaded directly from within Houdini input geometry.
or accessed through the SideFX Labs github page. Flowmap Visualize - A high quality realtime preview of a flowmap
texture in the Houdini viewport.
DOWNLOADING THE TOOLS Flowmap Obstacle - The flowmap obstacle SOP allows for easy
The Labs tools can be installed with Houdini or accessed from modifications on the flowmap based on geometry.
the SideFX Labs shelf tab. It not visible on most desktops so Niagara ROP - All-in-one HDA to extract and write out impacts,
you will need to add it. Once it is visible, click on the Update split data and interpolation data from a bullet sim to be used with the
Toolset button. This will pop up a dialog prompting you to UE4 Niagara data interface.
install. Since many tools are in beta, you may want to turn Off Gamedev Procedural Smoke - The procedural smoke SOP will
the Production Builds Only option. generate an animated volume to represent smoke.
ROP Vector Field - Generate UE4 compatible vector fields from
volumes or point clouds.

MESH PROCESSING
There are a lot of steps to get quality meshes into your games -
The shelf will be populated with many of the tools but some
photogrammetry, topology cleanup, mesh reduction, uv layout,
are only available when you press tab in the Viewport or
baking maps. Lab tools let you simplify that process with
Network view. These nodes are prefaced with Labs to make
wrapped up workflows and integrations with popular software.
them easy to find.
AliceVision Photogrammetry - AliceVision is a Photogrammetric
Computer Vision Framework which provides a 3D Reconstruction and
FX TOOLS Camera Tracking algorithms.
Houdini is known for its strong FX tools and SideFX Labs has
tools to process the results for use in a realtime environment
for games or virtual production. Labs has a range of tools to
help optimize and export your sims to textures, fbx, csv, etc.
Vertex Animation Textures - The Vertex Animation Textures ROP
will export a mesh and textures to be used with a realtime material
that can playback complex animations for cloth, rigid body destruction,
fluids and particles.

ZBrush Bridge - GoZ is Zbrush’s fast file transfer feature that lets
you send meshes between Houdini and Zbrush seamlessly without
having to deal with file paths or extensions.
HOUDINI FOUNDATIONS

Delete Small Parts - Removes parts based on connectivity and size.

Delight - Removes ambient lighting information found in high-res
photogrammetry scans.
Flipbooks Texture - Fast GL or Karma based creation and preview of
GameRes - Full Pipeline Node to Take High Res Models to Low Res.
texture atlases for Pyro FX.
Maps Baker - Generates textures bakes from a high resolution to
Destruction Cleanup - Prepare rigid body simulation results for
low resolution model at near interactive speeds.
export, reducing redundant geometry, cleaning normals, cleaning
attributes. LOD Hierarchy - Create and export an LOD Hierarchy as FBX.
Skinning Converter- Skinning Converter is a SOP that can convert Mesh Sharpen - This tool sharpens the geometry based on the
any non-changing topology deforming mesh sequence into a bone curvature found on the mesh.
based animation. Edge Damage - This tool will add edge wear to your geometry.

36
WORLD BUILDING UV MAPPING
Digital worlds are becoming bigger and more complex and Texture UVs are a big part of creating game art and these tools
it is important to have an efficient world building workflow. augment Houdini’s existing UV toolset to make you faster and
Whether you want to recreate New York City, grow a dense more efficient.
forest or add interior details to your sci-fi adventure there are Auto UV - Automatically generates the seams for an object and
Lab tools for you. immediately runs a UV Flatten after the fact.
Physics Painter - Physics Painter is a SOP that allows users to paint Inside Face UVs - Create UV’s for inside faces of fractured geometry.
physics objects onto any other object.
UV Transfer - Transfer uv’s between a source and target geometry.
Building Generator - Convert low-resolution blockout geometry
UV Visualize - Helper script to visualize UVs. Including features
into detailed buildings using a library of user defined modules.
such as: Visualize Seams, Warp between UV space and Model Space,
Modify the tiling of the grid texture and Visualize Islands.
Texel Density - This tool calculates the current texel density of an
asset per primitive based on the asset and project resolution.

INTEGRATIONS
These tools make it easier to import and export data into
Houdini and out to game engines.
Substance COP - The Substance Plugin for Houdini lets you load
Substance Archive files into Houdini in COPs.

OSM Import - Open Street Map is a great database for city street
data. This node efficiently loads the OSM files into Houdini as well as
all of the different tagged attributes on the buildings and streets.
OSM Buildings - Generate Buildings from OSM Data.
Tree Tools - The Tree Tools in Labs consist out of several tools that
can be used together to create intricate branching structures such as
trees, bushes, plants and much more.
Cable Generator - Given a curve that represents the high ‘pin’
points and low ‘sag’ points of a cable, this sop will generate sagging
cables, with user definable cable count, shape, color.
Rizom UV- The RizomUV Bridge is a set of 4 different SOPs that
Curve Branches- Scatters curves over curves, with many intuitive
facilitate the communication between Houdini and RizomUV.
controls to go from clean geometric branches to organic vines.
Duplicates of this sop can be chained together for recursive growth, Quad Remesher - The QuadRemesher node is a wrapper around
approximating the look of L-Systems but much more controllable. Exoside’s QuadRemesher command line interface.
Dirt Skirt - Create a geometry ‘skirt’ where an object and ground Instant Meshes - Reads in DDS (DirectDraw Surface) Files.
plane intersect, to be used as a soft blend in a game engine. Sketchfab - Uploads geometry to Sketchfab.
Lot Subdivision - Divide polygons into panels. Useful for city blocks 3D Facebook Image - Quickly render a 3D scene to a 2.5D image
or greeble. that can be uploaded to Facebook.
MapBox - Generate color, height and Open Street Map (OSM) Marmoset ROP- The Marmoset ROP allows you to quickly generate
curves using data provided by mapbox.com. an mview inside Houdini.
SciFi Panels - Example HDA to generate Sci Fi Paneling. Gaea Tor Processor - The Gaea Tor Processor allows you to load up
Snow Buildup - Adds geometry to an input mesh to mimic the build build .TOR files made in Gaea.
up of snow.
Terrain Texture Output - The Terrain Texture Rop SOP renders UX
image data from a heightfield. Some Labs tools are specifically designed to enhance the user
MODELING experience for artists working with Houdini.
Crash Recovery - Found under the File> menu, this functionality
The Labs Tools includes a variety of modeling tools designed to
allows for quick recovery from an unfortunate crashed file.
make it easier to create game-ready geometry.
Network Paint - Allows for colorful annotations of the network by
Decal Projector - Project a decal (localized piece of geometry and a simply drawing in the network editor.
texture) onto geometry.
Sticker Placer - Helpful for annotating networks by placing
Calculate Slope - Calculate the slope of a surface by comparing to a numbers, icons and user-created graphics.
direction, and optionally blur and remap the result.
External Script Editor - Sets up a live connection with an external
SIDEFX LABS

Curve Sweep - Sweep a profile along input curves, with simple IDE when working with VEX, Python, OpenCL and expressions.
controls for profile type, width, and twist behavior.
Extract Silhouette - Create an outline of an object projected from AND MUCH MORE...
one of the axes, xyz.
More Tools are being added on a regular basis. Go to SideFX.com/labs

37
 File Management
Understanding how to manage all of the files you create while working with Houdini is very
important to your success as an artist. A typical scene file can have outside dependencies on disk
and managing these is important especially if you are moving your files to a different computer.

PROJECT DIRECTORIES To create and load assets you can use the Asset menu. You
can also manage assets loaded into your scene using the
While Houdini can work with files scattered all over your hard
Asset Manager found on that menu. If you have two HDA
drive, this will make it harder to share your work and manage
files loaded into your scene that have the same name, Houdini
file dependencies. It is better to set up project directories
will choose one of them based on rules set up in the manager.
using File > New Project or use File > Set Project to choose an
Changes made to an asset definition in an HDA file will be
existing project directory as the home base for your work. This
automatically pulled into scenes that reference that file. Note
will make it easier to set up local dependencies with respect to
that older Digital Asset files may have a .otl extension which
all of the required project files.
will work exactly the same as .hda files.

SCENE FILES | .HIP APPRENTICE AND INDIE FILES

The main file type when working with Houdini is the .hip file. Houdini Apprentice and Houdini Indie use different file types
This file contains all your nodes and networks and is the file that cannot be opened in a commercial version of Houdini.
type used when you save your work. Apprentice uses .hipnc (non-commercial) for scenes and .hdanc
files for assets and Indie uses .hiplc (limited commercial) for
scenes and .hdalc files for assets.

UNIVERSAL SCENE DESCRIPTION | USD

In Houdini, the Solaris lighting and lookdev environment works
with USD [Universal Scene Description] which is an open
BACKING UP YOUR WORK
source initiative created by PIXAR. In Solaris, USD is native
and you can use procedural nodes to manage references, By default, Houdini creates a numbered backup of your scene
payloads, layers, collections, variants and level of detail. files and Digital Asset files every time you save. This gives you
a file to go back to if you want to review an early iteration or
HOUDINI DIGITAL ASSETS | .HDA if something happens to your working file. You can also set up
Houdini to AutoSave in the Edit > Preferences > Save and Load
You can also encapsulate then save out Houdini networks into
Options. Just remember that all those backup files take up disk
Houdini Digital Asset or .hda files. Parameters from inside the
space and you will want to clear them from time to time.
asset can then be promoted to the top level to create a custom
HOUDINI FOUNDATIONS

UI for the asset. These files can be easily shared with other
artists and provide a robust referencing architecture as your
assets evolve through the life cycle of a project.

38
FILE SOP INTEROPERABILITY
When you import geometry into Houdini using File > Import To import and export from Houdini, there are a wide variety of
> Geometry it puts down a File node at the geometry (SOP) file formats that you can use. Here is a list of some of the main
level. This file maintains a connection to the file on disk and formats you will work with in a typical Houdini pipeline.
changes made to that file will also update in your Houdini
scene. If you want to break this connection then you would
need to lock the File node. Houdini Files - Here are some file formats, other than .hip and
.hda that work exclusively in Houdini.
.bgeo - This format saves geometry along with related attributes such
as UVs, velocity and normals. Animations and simulations can be saved
out as numbered bgeo files to save out motion. A bgeo.gz file is a
compressed version of this format.
.sim - These files let you save out simulation data to cache the sim to
disk. Some people use these files while others use .bgeos to cache sims.
.ifd - This is a scene description format that is created when rendering
FILE DEPENDENCIES [$HIP/$JOB] to Mantra. Typically these are created while rendering in Houdini but
When you work with nodes that reference files on disk such sometimes they are saved to disk to be rendered by Mantra directly.
as geometry or texture files, the path you use will determine .pic - This is an image file format that was used by Houdini in the past.
what happens if you move the project directory to another It was replaced as the default format by the open source EXR.
computer or to the cloud. A direct path will break if you move .rat - This image format is ideal for texture maps being rendered in
the files therefore you should either use $HIP which uses the Mantra. All textures get converted to this format anyway so it speeds
scene file as the “home base” for the path or $JOB which uses up rendering to convert to this format using Mplay.
the project directory. You can use Render > Preflight Scene to
check to make sure that your scene file is set up properly.
Image Formats - These industry-standard formats are used to
render out shots and for texture maps.
.exr - OpenEXR is a high dynamic-range (HDR) image file format
developed by Industrial Light & Magic that is now the default format
for saving out renderings from Houdini.
.jpg/.png - These formats are used to publish images to the web
.tga/.tif - These formats are often used to texture map video games.

Geometry Formats - When importing and exporting geometry,

these formats are the most popular.
DISK SPACE MANAGEMENT .usd - This is the format used in Solaris/LOPS and provides an open
Large scene files, backup files and large simulations can take source interchange format for sharing with other applications.
up lots of disk space. Be sure that you are not filling up too .abc - Alembic is an open computer graphics interchange framework.
much space on your computer which could lead to instability .fbx - This format owned by Autodesk is popular when exchanging data
issues. Try to use external drives to save out the largest files with game engines and other 3D applications. It can hold geometry,
and leave the main disk on your computer with enough space rigging, motion and shader information.
to accomplish your day-to-day work. .obj - This is an simple geometry format originally created by Wavefront.

PREFLIGHT PANEL
From the Render menu, select Preflight Scene
to evaluate your scene setup.
Referenced Files - The preflight panel can
either reference $HIP or $JOB when verifying
file references for your scene file.
Greenlit Reference - If the reference is
relative to either $HIP or $JOB then it will be
displayed in green to indicate it is working.
File Management

Incorrect Reference - If a file reference is a

direct path and not relative to $HIP or $JOB
then it will be displayed in red and will need
to get fixed before sharing your project with
other artists or on the cloud.
Edit Expression - Click on any file name then
on the right click on the expression to open
up the edit expression window.

39
 Expressions & Scripting
Houdini is a production-level solution which means that scripting will play an important role in your
work. Artists can usually get by with simply writing expressions while technical directors will spend
more time using these techniques. Houdini includes support for Hscript, Python and VEX.

HSCRIPT EXPRESSIONS Python, the hou package is the top of a hierarchy of modules,
functions, and classes that define the HOM. The hou module
HScript is designed to be a fast and concise way to retrieve and
is automatically imported when you are writing expressions in
manipulate information that can be used to write expressions.
the parameter editor and in the hython command-line shell.
An expression is typically any value that is not either a simple
string or number. This can be something as simple as a variable, You can also use it to write expressions in Houdini. To do
a math equation or an expression function. this, change the expression language option at the top of the
parameter pane for the node.

You can enter expressions directly into a parameter by simply There is also a Python Shell Panel which you can use to enter
typing into a field. When you press Enter the field highlights in Python commands. You can also import the hou module into
green. You can click on the parameter name to toggle back and a regular Python shell to integrate Houdini into your existing
forth between the expression and the result of the expression. Python-based scripts.
If you are creating channel references then you can RMB-click
on a parameter and choose Copy Parameter then go to the
TOOL SHELVES
parameter you want to link it to and choose Paste Relative The shelf tools are also set up using Python. You can see this
References. code by RMB-clicking on any shelf tool and choosing Edit Tool.
You can also achieve this by RMB clicking on the second
parameter then choosing Reference > Scene Data. This opens
up a panel where you can choose data from other objects and
nodes and an expression will be built for you. This method can
even set expressions on multiple parameters.
EXPRESSION EDITOR
Depending on the complexity of your function, or the type of
parameter, you may instead choose to use the Expression PYSIDE/PYQT
Editor. The expression editor can be opened by RMB-clicking The Python Panel Editor pane lets you create, edit and delete
on a parameter and selecting Expression > Edit Expression, or PySide2 or PyQt5 interfaces. The editor also lets you manage
by placing the mouse over the parameter and pressing Alt - E. the entries in the Python Panel interfaces menu as well as the
entries in the Houdini pane tab menu. The panel comes with
some sample code that you can try out for yourself.
HOUDINI FOUNDATIONS

PYTHON
Python is a popular scripting language that is well known in the PYTHON STATES
CG industry that supports integration and standardization. This You can also write viewer states in Python that let you
makes it perfect for tool development. customize user interaction in the viewport for your node. You
Python in Houdini is built on the Houdini Object Model can use these to build more artist friendly interfaces for tools
(HOM) which is an API that lets you get information from and you can refer to the documentation for more detailed info.
and control Houdini using the Python scripting language. In

40
VEX WRANGLE NODES
VEX is a high-performance expression language used in many You can also use a wrangle node such as the Attribute Wrangle
places in Houdini, such as writing shaders. VEX evaluation is which provides a low-level node that lets coders who are
typically very efficient, giving performance close to compiled familiar with VEX tweak attributes. There are also wrangle
C/C++ code. nodes for working with channels, volumes and deformations.
VEX is not an alternative to scripting, but rather a smaller,
more efficient general purpose language for writing shaders
and custom nodes. VEX is loosely based on the C language,
but takes ideas from C++ as well as the RenderMan shading
language.
VEX is used in several places in Houdini:
Modeling – The VEX SOP allows you to write a custom surface node
that manipulates point attributes. This can move points around, adjust
velocities, change colors. As well, you can group points or do many
other useful tasks.
Rendering – Karma and Mantra use VEX for shading computation. This
includes light, surface, displacement and fog shaders.
Compositing – The VEX Generator and VEX Filter COPs allows you
to write complex custom COPs in VEX. The expressions evaluate very
close to C/C++ speeds and run 1000’s of times faster than the Pixel
If you are interested in learning how to work with wrangle
Expression COP. nodes, you should take a look at Entagma.com where you will
find lots of great tutorials that generally take a more technical
CHOPs – The VEX CHOP lets you create custom CHOPs. The CHOP
functions can manipulate arbitrary numbers of input channels and approach to creating content but with an artist’s mindset.
process channel data in arbitrary ways. In some cases, the VEX code
can run faster than compiled C++ code. COMPILE BLOCKS
Fur – Procedural fur behavior is implemented with VEX. In geometry networks [SOPs], you can put a part of the
network inside a compiled block that makes it function as
VOPS efficiently as if you had written code. This imposes a number of
If you want to use VEX but don’t want to write the code then restrictions on how the network can work, but can potentially
you can use the VOP context to use a node-based interface. deliver big benefits in the right circumstances.
You can do this in the SOP context using an Attribute VOP
node that lets you dive in and use VOPs to create VEX code.
You can take input geometry and manipulate it.

HOUDINI DEVELOPERS KIT | HDK

An even deeper way to work with Houdini is to use the
HDK which is the same comprehensive set of C++ libraries
You can use parameter vops to build interface element such as
that SideFX programmers use to develop the Houdini family
float sliders that are then available at the SOP level. This way
of products. With the HDK, you can create plug-ins which
Expressions & Scripting

you can execute the VEX code without diving back down to
customize different areas in the Houdini interface. Here are
the VOP level.
some examples of what you can do with the development kit:
ƒ Add custom expression functions
ƒ Add custom commands (hscript or HOM)
ƒ Add custom operators (SOPs, COPs, DOPs, VOPs, ROPs, CHOPs,
The VOPs context is designed to give artists an interactive way
and even Objects)
of creating VEX code. For people with a scripting background,
ƒ Add output nodes to support a non-standard Renderer
it might make more sense to write the code directly into a
ƒ Add custom lighting or atmospheric effects to the renderer
Wrangle node.
To learn more about working with the HDK, go to the SideFX
website and choose Support > Documentation > HDK.

41
 Tasks
With the Task Operators or TOPs, you can organize and schedule tasks then distribute them
intelligently to your compute farm. This allows for parallel processing of data while maintaining a
dependency graph that shows how each task relates to proceeding tasks.

PROCEDURAL DEPENDENCY GRAPH COOKING TOP NODES

TOPs is a network type in Houdini built using the Procedural Once you wire together a task graph, you will want to cook the
Dependency Graph, a technology which makes it possible nodes. You can either cook a node in the middle of the graph
to describe complex dependencies visually with nodes, then or cook the output node at the end of the chain.
generate a set of actionable tasks which can be distributed
ƒ Cook Selected Node Shift-G
to a compute farm with the help of a scheduler. Once you
ƒ Dirty and Cook Selected Node Shift-V
evaluate the results, it is possible to make changes to parts of
the graph without re-cooking the whole network.

TOP NODES Use the Task Bar to monitor your progress. On the TOP nodes,
Task or TOP nodes let you manage pipeline tasks with the you can RMB-click on a task dot and choose to Cook or Dirty
ultimate goal of parallelizing the processing and distribution the task. When you dirty a task, it means that if you recook the
of each task. As a TOP node generates a task, the task is network those tasks will be recomputed. Clean tasks will not
displayed as a dot. Once the task is cooked then new tasks can be recooked which is one of the benefits of TOPs because you
be executed on this node and on children TOP nodes. don’t have to redo work that is already completed.

DEPENDENCIES
When you click on a task dot in the graph, you will see a thin
line that connects to upstream tasks it is dependent on and
downstream tasks that depend on it.
If there are changes upstream then tasks may be automatically
dirtied and this will in turn dirty tasks downstream where there
are dependencies. This process is an important part of how
SCHEDULERS PDG graphs work as an effective pipeline tool.

Scheduler nodes take tasks that have met required

dependencies and assigns compute resources. As each task is
completed, the scheduler informs the task graph, which in turn
informs the PDG graph to move on to the next available task.
PDG supports industry-standard schedulers such as HQueue,
Deadline, Tractor, or any scheduler plugged in with Python.

TOP NODE
Input - The node takes the information feeding
into the Input and breaks it into one task for
every piece of data.
Progress Wheel - The progress wheel shows
you how many tasks are completed, how many RMB on Node....
are in progress and how many are in the queue.
TOP Node - This is the node that is currently
HOUDINI FOUNDATIONS

being cooked. It contains the instructions for

the tasks being executed. You can RMB-click on
the node for a menu of supporting actions.
Tasks - Each task is represented by a small dot.
The coloring indicates their current status and
you can RMB-click on a task dot to learn more
about that one part of the graph.
Output - Once tasks are completed, the output
passes the results on to the next node even if
other tasks on this node are still active.

42
TASK GRAPH TABLE OUTPUT IMAGE MOSAICS AND MOVIES
If you RMB-click on a node, you can choose to Open Task In TOPS, you can interface with ImageMagik to create a
Graph Table. This gives you an itemized list of tasks along contact sheet that can be used to evaluate design iterations to
with information such as index, state, cook time and priority. make the best choice or to generate prop variations to richly
Clicking on items in this window will highlight the task dot on populate your scene. You can use an overlay to pull info from
the node in the Network view. the network to help you make the best decision.

IMPORT/EXPORT DATA
To get data into the TOP graph, there are a number of different
options giving you access to geometry, images, scripts and INTEGRATIONS WITH OTHER APPS
other kinds of data. Houdini Digital Assets can be used to apply
TOPs includes nodes for working with other applications such
procedural networks or you can connect with other parts of
as Shotgun or Autodesk Maya. This allows your network to
Houdini to import and export data.
extend beyond Houdini to help with all parts of the pipeline.
WEDGE NODE
A key workflow in PDG is wedging that lets you quickly create
multiple iterations of a design. You can then process all of the
different options through the TOP graph then collect them at
the end for final output.

PILOT PDG APPLICATION

While TOP networks can be set up and executed from within
Houdini, wranglers who are managing the farm or pipeline TDs
who are exclusively creating TOP networks can use PilotPDG.
Tasks that are Houdini-related will call on Houdini Engine to
work non-graphically to complete the task.

TOP NETWORK
Network Path- This shows you
This network type lets you the path to the TOP network
manage and view the network where the graph is set up.
being processed.
TOP Menu- This menu
Task Bar- The task bar lets you includes a range of options for
start and stop a network and organizing and processing a
monitor its progress. TOP network.
Scheduler- The scheduler node Progress Bar - This bar lets you
determines where your data is see the progress of the overall
being processed and how many network tasks.
nodes are participating.
TOP nodes - These nodes are
Completed Tasks - When a node where specific commands are
is finished processing all its tasks, turned into tasks and sent by
a check mark appears. the scheduler to be completed.
In Progress Tasks - While in Dependency Line - You can
progress, you can see which click on a task to see how it
tasks still need to be completed.
Tasks

connects to other tasks in the

network.

43
 HOUDINI DIGITAL ASSETS
Procedural Tool Building
Networks of nodes give Houdini its procedural nature and define a recipe that can be applied over
and over. Houdini Digital Assets let you wrap up these networks to create custom tools and smart
assets. These artist-built tools can be used repeatedly to increase productivity across your studio.

One of the things that Houdini’s node-based workflow is great PIPELINE FRIENDLY
at is allowing artists to avoid repetitive steps and to generate
When a Houdini Digital Asset is loaded into a scene file, it
multiple iterations by making changes to an existing network
references the .hda file on disk. This means that changes made
of nodes. This lets you achieve results that are unique without
to the asset will be picked up automatically by everyone who
starting the whole process from scratch.
is referencing that file.
Houdini Digital Assets take this one step further by letting you
This makes it very easy to deploy updates throughout your
encapsulate a network or collection of networks into a single
pipeline. Now artists can point to a single asset on disk
node with parameters that have been promoted to the top
knowing that once it gets updated with the most current
level. This node is then saved to disk which creates a shareable
iteration, they will immediately have access.
file that other artists can load into their scenes.
Houdini Digital Asset files can also hold more than just the asset
ARTIST BUILT TOOLS definition. You can store images, geometry files and scripts that
are used by the asset. This ensures that all the relevant parts are
The process of creating a Houdini Digital Asset works with the
available when other people work with the asset.
interactive tools in Houdini. You build a high level interface by
dragging parameters from nodes to an asset properties panel
CONTENT LIBRARY & ORBOLT
which allows you to create custom tools without writing any
code. This means that technical artists can build custom tools The Content Library is an online asset repository which hosts
then deploy them quickly to colleagues. 2D and 3D assets, from complete scene files, to fully-rigged
props, to render-ready visual effects, animatable characters,
A Houdini Digital Asset might be a procedural prop such as
game assets and more. Go to Get > Content Library on the
a staircase or a piece of furniture, a visual effect such as an
SideFX website to access it.
explosion, or a more generalized tool such as a populate tool
for scattering objects over a surface. Whether you are creating Orbolt an online asset marketplace that offers a wide variety
content specifically for your current project or building a larger of digital assets. There is a panel in Houdini where Orbolt
toolset for all your projects, your artists can build a collection assets you download or purchase can be stored and made
of Houdini Digital Assets to meet your production needs. available as you work.

CREATING DIGITAL ASSETS

1 Create nodes and

networks in Houdini
Houdini
4 Load the .hda file back into
Houdini to use the asset.
Only those parameters promoted
to the asset level can be used. All

2 Package up the networks to be saved others are locked.

out as a Houdini Digital Asset [.hda] You can use the asset in any num-
file that can be shared with other artists. ber of Houdini scenes. If you make
HOUDINI FOUNDATIONS

changes to the HDA file then all

SAVE AS other assets can be easily synced to

3
Houdini
Build an interface for your asset by Digital the changes.
Asset
promoting parameters and handles
to the top level of the asset.

.hda file

44
 HOUDINI ENGINE
Sharing with other Apps
Houdini Engine brings a procedural node-based approach to your favorite app. This technology lets
you share Houdini Digital Assets with colleagues who can load them directly into 3D apps such as
Autodesk® Maya® or 3DS MAX® or into game editors such as Unity® or Unreal..®

The benefits of Houdini Digital Assets can be

experienced by artists using other applications
thanks to the Houdini Engine plug-ins. Created with
the Houdini Engine API, these plug-ins allow host
applications to load .hda files and all of the handles
and controls. When parameters are set on the asset,
Houdini works “under-the-hood” to cook the nodes
and networks then deliver the results back to the host.

HOUDINI ENGINE API

This is made possible by the Houdini API which is
used to create plug-ins for host applications. HAPI
is a flat and small API that is easy to learn and is
available on github for developers that want to
create their own proprietary plug-in. A Houdini Digital Asset loaded into Autodesk Maya using the Houdini Engine

HOUDINI ENGINE PLUG-INS Plug-ins that work with FREE Houdini Engine for Unity/Unreal
There are a number of Houdini Engine plug-ins that artists can or FREE Houdini Engine Indie licenses:
access either through the Houdini installer or online. These ƒ Unreal
have been production tested and can be used confidently by ƒ Unity
artists and studios.
Plug-ins that work with Houdini Engine licenses or FREE
Each of the plug-ins are designed to create a bridge between
Houdini Engine Indie licenses:
the features in a typical Houdini asset and the nature of the
host application. For instance a cloud asset that use volumes ƒ Autodesk Maya
would work fine in Maya but would not make sense in Unity or ƒ Autodesk 3DS Max
Unreal where volumes are not supported. ƒ Proprietary Plug-ins

HOUDINI ENGINE PIPELINE

1 Load the .hda file into a host

application using the Houdini
Engine plug-in.
.hda file
Houdini
Libraries

2 The Host Application

accepts the asset and interfaces
with the Houdini Engine.
LOAD
Houdini
Digital
Asset

3 The Houdini Engine calls

on the Houdini library files
to “cook” the nodes and
HOUDINI ENGINE

network inside the asset.

Host App
Houdini
Engine

4 When an asset is loaded or a

parameter is changed then the
Engine grabs the Houdini libraries,
cooks the nodes, then delivers the Host applications with
results back to the Host. Houdini Engine plug-ins
45
 FILM & TV PIPELINE
Animation and VFX
Whether you are creating live action plates enhanced with visual effects or full CG shots, the
ultimate goal of Film and TV projects is moving pictures. These pictures are created using assets
such as characters, sets and effects which all come together in a final composition.

Houdini is a full featured package that contributes to all stages of INTEROPERABILITY

a Film & TV pipeline. From modeling to rendering to animation
Most studios are equipped with a variety of 3D applications to
and final compositing, Houdini has procedural tools that support
each handle a different part of the pipeline. Houdini has a lot
your creative process. Over the years, VFX is one area where
of strong interoperability tools to allow for this interchange of
Houdini is known as an industry standard. SideFX has been
data. Whether they are using USD, Alembic, FBX or EXR, your
honored with several Scientific and Technical Achievement
artists can easily work back and forth with a wide variety of DCC
awards including an ACADEMY AWARD OF MERIT Oscar.
applications. They can also use the Houdini Engine plug-ins to
Other areas such as procedural modeling, lighting or character bring the Houdini Digital Assets into other applications, such as
work continue to get stronger to the point where a growing Autodesk® Maya® or 3DS MAX,® while maintaining the asset’s
request from studios is more skilled Houdini artists. procedural controls.
Smaller studios may want to avoid extensive file exchange,
HOUDINI CORE / HOUDINI FX
especially with tight deadlines, therefore Houdini provides a
There are two commercial versions of Houdini that you use full featured procedural “pipeline-in-a-box” that can take you
in your pipeline. Houdini Core covers all of Houdini’s tools through all of the stages under one roof.
except for DOPS, and Houdini FX has a full toolset. Scenes
and VFX created in Houdini FX can be staged, animated, lit DISTRIBUTED RENDERS AND SIMS
and rendered in Houdini Core. This gives you a robust pipeline
Rendering images and simulating VFX can be time consuming,
with Houdini FX licenses for your FX artists and Houdini Core
especially as you aspire towards photorealistic results. For this
licenses for everyone else.
reason, Houdini lets you distribute both rendering and simulation
One solution is for senior technical directors to use Houdini tasks to a compute farm using Houdini Engine in Batch mode.
FX to solve a particular production challenge then wrap up
Distributed simulations allow you to work faster or to handle
the resulting nodes and networks into Houdini Digital Assets.
effects that would max out the memory on any one computer.
An artist-friendly UI is then built to support the animators and
By slicing the sim and distributing it, memory is managed
VFX artists who can then use the more cost effective Houdini
without compromising the final result. Studios should definitely
Core to execute shots.
consider using Houdini Engine to simulate on the farm.

FILM & TV PIPELINE

Houdini FX
Distribute
Houdini CORE Render
Distribute
Karma Render/Mantra Sims

Houdini Engine Particles and Lighting Render

Dynamics VFX VFX
File Transfer Simulate
VFX
USD, FBX, OBJ, EXR etc...

Animate
HOUDINI FOUNDATIONS

VFX

Distribute
Render

Model Rig Layout Animate Light Render Composite

46
 GAMEDEV & VR PIPELINE
Interactive Experiences
In Video Game and Virtual Reality projects, the main focus is creating interactive 3D worlds built
using content that is highly optimized for a smooth gameplay experience. This creates a different
kind of pipeline compared to rendered out game cinematics which are more like film.

At the core of a games pipeline is a game engine

like Unreal or Unity. The engine is where the game
art and the game interactions are put together to
create a playable experience. Houdini can be used by
game artists to create terrain, design and populate
levels, build procedural models, build and animate
characters and create Realtime FX such as fire, fluids
and destruction.

EXPORTING TO GAME ENGINES

There are two ways of getting content from Houdini
to a game engine. The traditional approach is to
export out to a format like FBX or OBJ and import
this into the engine. You would create procedural
systems in Houdini then flatten out the results. Houdini Digital Assets loaded into Unreal using the Houdini Engine
The second approach is to create Houdini Digital
Assets and load these into the game engines using the Houdini REALTIME FX
Engine plug-ins for Unreal and Unity. These assets import into Houdini is known for VFX and it is a great tool for creating
the game editor with their parameters and controls intact. You FX for games. But these FX need to be optimized using
can therefore make changes inside the game editor and the techniques such as texture sheets, flowmaps and vertex
Houdini Engine works in the background to update the artwork. animation textures. This way the footprint for the effect is as
This proceduralism is available to game artists inside the editor light as possible and does not take away from the frames per
then when the game is compiled the artwork is baked down. second of the game. The SideFX Labs Tools mentioned earlier
The Houdini Engine is not a runtime solution and you cannot in this document have been designed to support these kinds
access it as part of the gameplay. of workflows.

GAMEDEV & VR PIPELINE Houdini FX

Houdini CORE
Game Engines/Other Apps

Houdini Engine Plug-ins

Maya/
3DS Max
Procedural Level File Transfer
Modeling Design FBX, OBJ, USD, EXR etc...

Save as
HDA
GAMEDEV & VR PIPELINE

Terrain Populate
Generation Environment Game Compile/
Engine Bake Assets

Realtime FX

Play Game

Model Rig Animate Export 47

 Products & Licensing
As you begin working with Houdini, it is useful to understand LICENSE TYPES
what Houdini products are available for you to work with.
Workstation [Node-Locked] – This license type can be used
Whether you are a large studio, small studio or a team of indies
on a single computer and can only be accessed either from a
just getting started, there are different Houdini products to
local server or from SideFX.com.
suit your needs. There are also versions of Houdini for school
labs and for students who want to learn for free. Local/Global Access [Floating] – These licenses can be set up
on a server and shared with a team of artists. When an artist
COMMERCIAL LICENSES starts Houdini a license is checked out of the server as long as
there is one available. Local licenses are designed for a single
Houdini Core – Designed for modelers, lighters, character
studio and global licenses are for sharing between studios in
riggers, animators and game artists, Houdini Core also includes
different locations.
features such as compositing and motion editing. Scenes
created in Houdini FX can be opened and rendered in Houdini
INSTALLING LICENSES
Core which makes it an ideal lighting tool for your VFX.
Once you have acquired a license, you will install it by opening
Houdini FX – Houdini FX has all the tools found in Houdini
up the Houdini License Administrator [hkey] application. From
Core but adds particle and dynamic simulation tools. With its
there you can choose File > Install Licenses. You will be asked
procedural workflow, Houdini FX lets you create fluids, Pyro
for a log in and a password that will match the ones you set up
FX, grains, cloth, hair & fur, crowds and soft body effects.
on the SideFX.com website.
Houdini Engine – Houdini Engine gives you command-line
You can now use sidefx.com as the license server instead
access to run in batch mode to batch process renderings and
of installing locally. To use Login licensing, you need to be
distributed dynamic simulations. Houdini Engine also lets you
constantly logged in with your sidefx account. You can login
load Houdini Digital Assets into other digital content creation
using different computers but only one of them can be used at
applications, such as Autodesk® Maya,® Autodesk® 3DS MAX,®
a time. This approach is ideal for Indie and Education users.
Unity® or Unreal.®
Local and Global Access licenses can be installed using this
INDIE LICENSES method on a central server. You will then need to make that
Houdini Indie – Houdini Indie makes all of Houdini’s animation server available to anyone who needs access to the licenses.
and VFX tools available under a limited commercial [less than You can also view your licenses on the SideFX.com website by
$100K USD] license to animators and game makers who want clicking on your avatar in the top right and choosing Services.
to use Houdini during the incubation stage of their business. You can then click on the Manage Licenses link.
Houdini Engine Indie – Your Houdini Engine Indie license can
be used to run Houdini Indie in batch mode or to load Houdini
ANNUAL UPGRADE PLAN
Digital Assets into other content creation apps. For visual effects studios, games studios and 3D artists who
want to maximize their investment in Houdini, the Annual
LEARNING LICENSES Upgrade Plan provides key advantages such as production-
Houdini Education – Houdini Education is a full-featured version level technical support and access to full and dot releases
of Houdini FX designed for use by schools and training centers as containing the latest software enhancements as well as daily
well as students. Houdini Education will also open files created by builds containing bug fixes.
using Houdini Apprentice.
SIDEFX SUPPORT
Houdini Apprentice – Houdini Apprentice is a free version
All customers including Apprentice customers can contact
of Houdini FX which can be used by students, artists and
SideFX using our email support system to discuss installation
hobbyists to create personal non-commercial projects. With
and licensing issues. After that, only Annual Upgrade Plan and
Houdini Apprentice, you have access to virtually all of the
Commercial Rental Customers may contact our support team
features of the award-winning Houdini FX to develop your
to discuss more in depth production issues.
skills and work on personal projects. Apprentice lets you save
HOUDINI FOUNDATIONS

to disk and render out with a word mark. Our Support Specialists can be contacted directly via
[email protected] . Be sure to include the following
Note: Scene files and Assets created in Indie, Apprentice or
information in your email:
Education cannot be used in commercial Houdini. The file formats
are different and the EULA [End User License Agreement] prevents ƒ Your Operating system [Windows XP, etc.]
you from sharing files between different license types. ƒ Version and Build Number of Houdini
ƒ Summary of the installation issue and a diagnostic file if you are
having a licensing issue.
To learn about support programs visit SideFX.com/support.

48
 Comparison Chart
COMMERCIAL INDIE LEARNING

PRODUCT HOUDINI FX HOUDINI CORE HOUDINI INDIE EDUCATION APPRENTICE

INTENDED USER Studios | Commercial Artists Indies | Freelancers Schools | Students Hobbyists
PRICING Visit SideFX.com $269 USD per year $75 USD per year FREE
Operating System Windows, LINUX, Mac OSX
Modeling     
Character     
Animation     
Solaris: Layout Tools     
Solaris: Lookdev and Lighting     
Karma/Mantra Rendering     
Terrain     
Compositing     
Volumes     
Pyro FX  Simple Fireball   
Fluids  Simple Flip   
FEATURES

Rigid Bodies  Simple Fracture   

Particles  -   
Vellum Cloth  Simple Cloth   
Wire Dynamics  -   
Crowds  -   
LICENSING Commercial Limited Commercial Non-Commercial
Workstation [Node-Locked]    - 
Local/Global Access [Floating]   -  -
USER INTERFACE
Houdini GUI Access     
Command Line Access     

GUI Watermark - - Unobtrusive Unobtrusive Unobtrusive

Plug-in Support     
HOUDINI ENGINE
Houdini Engine Plug-ins     No
Create Assets for Engine     For Education Licenses
Create Assets for Orbolt     
RENDERING
Karma Tokens 5 / 10* 5 / 10* 1 10 1
Mantra Tokens Unlimited Unlimited 1 10 1
3rd Party Rendering     No
Render Watermark - - - - 
Resolution Unlimited Unlimited Unlimited Unlimited 1280x720
SCENE
.hip   .hipalc .hipanc .hipanc

.hda   .hdalc .hdanc .hdanc

GEOMETRY
USD     .usdnc
FBX     IMPORT
Comparison Chart

Alembic     IMPORT
.bgeo     
FORMATS

IMAGES
.pic   .piclc  .picnc
.exr     watermarked
.tif     watermarked
.png/,jpg     watermarked

* Commercial Workstation Licenses come with 5 Karma Tokens and Local and Global Access Licenses come with 10 Karma tokens

49
50
HOUDINI FOUNDATIONS

MODEL, RENDER, ANIMATE

Welcome to Houdini. In this lesson you will start from scratch to model, render, animate, and simulate
a soccer ball (also known as a football in many parts of the world). You will create a classic bouncing ball
animation using the principles of squash and stretch, apply textures and materials, add lights and cameras, and
explore the use of dynamics to simulate a group of soccer balls.
These tasks will introduce you to many different parts of Houdini as you create your first Houdini scene,
explore the interface and discover some of its most important tools. You will learn how to work interactively
in the Scene View and how to use the Network View to manage your nodes as you refine your model
and build your animation rig. You will also set up materials and textures on the Solaris Stage then you will
render using Houdini’s built-in renderer Karma, and finally create a Rigid Body Simulation.

LESSON GOAL
Model, Render, Animate and Simulate a soccer ball using Houdini’s procedural node-based workflow

WHAT YOU WILL LEARN

ƒ How to work with the View Tools
ƒ How to use Shelves, Radial Menus and the Tab key
ƒ How to create Geometry
ƒ How to work with Nodes and Networks
ƒ How to set up Custom Attributes and a For-Each Loop
ƒ How to set up Materials and Texture UVs
ƒ How to Layout a shot and render with Karma
ƒ How to Set Keyframes and add Motion FX
ƒ How to use Rigid Body Dynamics

LESSON
COMPATIBILITY
Written for the features in Houdini 19.5+

The steps in this lesson can be

completed using the following
Houdini Products:

Houdini Core ✔
Houdini FX ✔
Houdini Indie ✔
Houdini Apprentice ✔
Houdini Education ✔

Document Version 4.0 | July 2022

© SideFX Software
51
 PART ONE PROJECT FILES
Explore the Houdini UI Go to the soccerball tutorial page on
SideFX.com, where you likely got this docu-
To get started, it is important to learn how to work with the Houdini work- ment, to download the intro_lesson directory.
space and the three panes you will use the most. The Viewport lets you cre- Put it into the Houdini Projects directory
ate objects interactively, the Parameter Pane lets you edit node properties which you can find in either the home direc-
and the Network Editor lets you work directly with the node networks.
tory or the documents directory.

01 Select File > Set Project. Find the intro_lesson directory

that you downloaded earlier and press Accept. This
makes this project directory and its sub folders the place for all the
files associated with this shot.
Select File > Save As... You should be looking into the new
intro_lesson directory. Set the file name to soccerball_01.hip (or
football_01.hip if you would prefer) and click Accept to save.

02 In the viewport, press c to bring up a radial menu. From

this menu, choose Create > Geometry > Box. Your cursor
now shows the outline of a box waiting to be placed in the scene.
Press Enter to place it at the origin.
This creates a box in the Scene view, adds a node in the Network
editor and shows the object parameters in the Parameter pane.
As you work through this project, you will touch on all of these
interface elements.

03 You can now explore the View tool in Houdini. Press the
following hotkeys:
ƒ Tumble Spacebar or Alt[Opt] - LMB click-drag
ƒ Pan Spacebar or Alt[Opt] - MMB click-drag
ƒ Dolly Spacebar or Alt[Opt] - RMB click-drag
In some cases, you will want to home in to get your bearings.
There are some hotkeys for that as well:
ƒ Home Grid Spacebar -H
ƒ Home All Spacebar - A
ƒ Home Selected Spacebar - G
HOUDINI FOUNDATIONS

RADIAL MENUS
One way to access tools in Houdini is radial menus which you can
access using the X, C and V hotkeys. Each of these brings up a radial
menu with lots of options for you to choose from. The main focus
of each menu is as follows:
ƒ Snapping X
ƒ Main (or Custom) C
ƒ View V

52
SELECTION HOTKEYS
RMB-click to access menu
If you are using the Select, Move, Rotate, Scale or Handles tools,
the following hotkeys will determine your selection mode as well as
which level you will be working at.
ƒ Objects Object Level 1
ƒ Points Geometry Level 2
ƒ Edges Geometry Level 3
ƒ Primitives/Faces Geometry Level 4
ƒ Vertices Geometry Level 5

04 With the object selected, press i to go to its geometry

level. Use the Shift key to drag on handles to make it
longer along z axis around the origin.
When an object is created in Houdini, there is an Object level
which is where you manage the transformations of the object and a
Geometry level where you define its shape. Pressing i brought you
down into the geometry level of this object. You can also get there
by double-clicking on the object node in the Network editor. Later,
to get back to the Object level, you will press U.

05 Press S to go to the Select tool then 4 to access Primitive

selection. Press n to select all then press c to bring up the
radial menu and choose Model > Polygons > Poly Extrude.
In the either the Operation Control Bar at the top of the Scene
view or the Parameters pane, set Divide Into to Individual
Elements and use the handle to set the Distance to around 0.4.
This extrudes the faces along the normals of each primitive.
You can see that there are now two nodes in the Network view.
Each step you take in Houdini creates a node that you can work
with to refine your scene.

06 Press n to select all of the new faces and press Tab and
begin typing sub… then select Subdivide from the list.
The Tab key is another way to access tools in Houdini. Typing the
tool name lets you focus the list making it easier to find what you
want without navigating the submenus.
In the Parameter pane, set Depth to 2. This subdivides the
geometry to create more polygons. Houdini also has a subdivision
display option at the Object level which you can use to see
subdivisions without actually adding any geometry, but in this case
you do want to create more polygons.

07 Select the different nodes in the chain. The handles for

EXPLORE THE HOUDINI UI

each of the nodes appear as you select them but the

display remains on the final shape. Set the Display Flag on each of
the nodes to change which node is the display node. You can also
try some of the other flags such as Bypass or Template. Wiggle the
polyextrude node out of the network then drop it back in.
At the end, return everything to normal and set the Display flag
on the subdivide node. This is very important. The Display flag
determines what you will see at the object level. Always check to
make sure you have the right display flag set!

53
 PART TWO
Create a Soccerball
You are now going to replace the box with a soccerball shaped platonic shape. Using Houdini’s procedural approach, you
can replace the box node with a platonic solid node. From there you will adjust the other nodes to make it look like a soccerball.
This ability to swap out input nodes lets you prototype networks with simple geometry for added flexibility.

01 In the Network editor, use the Tab key to add a Platonic

Solids node to the network. Click to place it down near
the top of the chain. Wire the platonic node into the polyextrude
node. In the parameter pane, set Solid Type to Soccer Ball. Select
and delete the box node.
Because of Houdini’s procedural nature, it is often possible to
replace an input node and have the whole network function
properly. This gives you flexibility as you work and if you don’t like
the results after the change then you can always wire back the
original shape.

02 Select the polyextrude node. Make sure the Handle

tool is active then use the handle in the viewport to
set a smaller Distance. You can also set the parameter value in
the Parameter pane. This creates a better look for the soccerball.
Remember that even though you are viewing the subdivide node,
selecting the polyextrude node gives you access to its handles and
parameters.
You might think that with this primitive type you are all set but it is
really just a truncated icosahedron with flat faces. You need a round
soccerball so you will have to put a little more work into it.

03 Press V in the viewport and from the Radial Menu, select

Shading > Smooth Shaded. You can also use the menu in
the top right of the viewport to change your shading.
This soccerball looks like a cheap plastic ball rather than a proper
leathery soccerball. You are now going to branch off and add more
nodes to get a better look.
After analyzing it, set the shading back to Smooth Wire Shaded.
HOUDINI FOUNDATIONS

SHADING OPTIONS
Disable Lighting
There are a number of Shading Options available from either the
Headlight Only
View radial menu or the Shading menu in the top right of the Viewport.
Normal Lighting
For the shading of your objects, the lighting is determined by the
Display Options on the right edge of the Viewport. You can choose High Quality Lighting

from a headlight, normal lighting or high quality lighting with shadows. High Quality with Shadows

To quickly toggle from your shaded view to wireframe press the W key.

54
 04 In the Network view, press Y and drag across the line
connecting the subdivide node and the polyextrude node
to break the connection. You are now going to move the subdivide
in between the other two nodes so that you get a rounder
soccerball.

05 Drag the subdivide node in between the platonic solid

node and the polyextrude node. You can drop it on the
connecting wire and it will insert itself in automatically. If not then
jiggle it a little until it finds the connection. This will give more detail
to the sphere before it is extruded.

06 Use the tab key in the network editor to add a Ray node
and wire it in after the subdivide. Now add a sphere node
to the network and set its Radius to 1, 1, 1 and the Primitive Type
to Primitive. Now wire the sphere into the second input on the ray
node. This will project the subdivided ball onto a perfect sphere.
This is a very powerful node in Houdini that lets you project points
from one piece of geometry onto another. It is the perfect solution
to our problem of a subdivided soccer ball that wasn’t truly round.

07 Set the Display flag back on the polyextrude node. With

Divide Into set to Individual Elements all the small
polygons are extruded but you don’t want that. Set it to Connected
Components then all the polygons are extruded.
You need a way for this network to extrude the original patches of
the soccer ball but after the ball has been subdivided. You can do
this using the primitive numbers on the original geometry.

THE RAY NODE

CREATE A SOCCERBALL

The ray node is a tool that projects points out to another piece of geometry. This
is similar to the pinboard toy you played with as a kid. In fact, this is the node you
would use to set up a pinboard in Houdini.
GETTING HELP | To learn more about each node, you can click on the help button
in the top right of the Parameter pane to open up the node’s online documentation.
You can also hover over the tool in the shelf and press F1. In many cases, there are
sample files that you can open in Houdini to learn what the node can do.

55
 PART THREE
The For-Each Node
Now you get to see the magic, as the attributes you just created in the last part are fed into a for-each loop where the original
patches are extruded even though each contains many polygons. This will provide a more leathery look for the soccerball once
you subdivide it once more time after the poly extrudes.

01 Add an AttributeCreate node after the platonic node. –

Set its Name to patches and its Class to Primitive. Now
set the first of the Value fields to @primnum.
This expression takes the primitive number attribute and turns it
into a new attribute called patches.

02 With the attributecreate node selected, click on the

Geometry Spreadsheet tab next to the main viewport.
Click on the Primitive button and you can see the primitive
numbers on the left, three color attributes which show the color of
the patches and the patches attribute which matches the primitive
PATCHES = 0 numbers.
Click on the ray node. This attribute will be carried forward when
the shape is subdivided. You can now see there are a lot more
primitives but the patches attribute only goes as high as 31 and
then it goes back to 0.

03 Go back to the Scene View tab, RMB-click on the

Visualizer display button on the Display Options bar and
click on the + Plus sign next to Scene and choose Marker. In the
Edit Visualizer panel, set Name and Label to Patch_Numbers, set
Type to Marker, Class to Primitive and Attribute to patches.
HOUDINI FOUNDATIONS

WORKING WITH ATTRIBUTES

Attributes can be assigned to Points, Primitives or Vertices. Some
typical types of attributes include color (Cd) or UVs. You can see
the attributes at any point in your chain by mousing over a node and
choosing the i from the radial menu. You can also review the attribute
values in the Geometry Spreadsheet panel.
In this lesson, you will be creating a custom attribute called patches
which will help you in the for-each loop.

56
 04 Now make sure the Visualizer display button is on
and you will see the patch values on the soccerball in
the viewport. You can see that the prim number from the original
platonic solid have been transferred to the subdivided faces.
Change the display flag to different nodes to see the relationship.
This information will be used to polyextrude the patches properly
using a for-each loop.
Turn OFF the Visualizer display and save your work. These steps
have been a bit abstract and probably feel a bit too technical, but
don’t worry the payoff is coming.
attributecreate node ray node

05 In the Network editor, press tab and start typing For-

each Named Primitive to access two nodes that you can
then place into the scene. Wire the foreach_begin between the ray
node and the polyextrude node and then the foreach_end after the
polyextrude node. Select the foreach_end node and in the parameter
pane leave Piece Elements set to Primitives and set Piece Attribute
to patches. Set display on the foreach_end node.
You should now see the original patches being extruded together
based on the patches attribute. If they are not make sure Divide to
is set to Connected Components on the polyextrude node.

06 Click the checkbox next to Single Pass to explore what

is happening. Drag on the slider to watch as each of the
patches is polyextruded individually. You can also set values higher
than 10 to see more of the patches.
Turn off Single Pass to see the full shape. The for-each nodes create
all of the patches then return the final geometry. The for-each loop
is a powerful set of nodes that you will use often with Houdini.

07 Add a Fuse node after the foreach_end node and set

its Display flag. This connects the pieces into a single
topology. The for-each nodes broke them into the different patches
but didn’t fuse them back together.
Add a Subdivide node after the Fuse. Set the depth to 2. This will
give you more detail in the viewport that you can use to evaluate
your model. This adds more polygons but will not yet render as
a true subdivision surface. Houdini also lets you set Subdivision
display in the viewport without adding geometry but this Subdivide
node is needed for later in the lesson.

08 Play with polyextrude values to get a nice leathery soccer

ball. Here you set a Distance of 0.1 and an Inset of -0.02.
This gives nice rounded patches that look much better.
THE FOR-EACH NODE

Go to the Object level and rename the object soccerball_geo in the

Parameter pane. Either select the node and press F2 or double click
on the name. Click on the Render tab and turn on Render Polygons
as Subdivisions (Mantra) to set up true subdivisions at render time.
Select the Render Region tool, then draw a box around the
soccer ball in the viewport create a preview rendering. To cancel,
click on the x button in the top right of the region.

57
 PART FOUR
Setting up UVs
In order to set up materials and textures, it is important to make sure that there are proper UVs set up on your object.
Geometry in Houdini does not come with UVs, therefore you must create them yourself. This means adding extra nodes
to the network, which in this case means adding UV Quickshade and UV Flatten nodes.

01 Dive into the soccerball_geo object. In the Network view.

press tab > UV Quickshade and place the new node right
after the foreach_end node. Set its Display Flag.
Click on the File Selector button next to Texture Map. Navigate
into the /tex folder and select soccerball_color.rat. Now you can see
this texture map on your geometry but everything seems stretched
because the UVs were created using a projection method.
Note: Setting up UVs to match an existing texture is not the normal
order of operations. We are taking this opposite approach so that you
don’t need to paint a texture yourself.

02 In the Scene view, press n to select all then press tab >
UV Flatten. In the Group field, enter the expression
@patches>19. This will flatten the dark patches on the soccerball
geometry using the patch boundaries to lay out the UVs.
Using this tool brings up UV view. Click on the UV (vertex) menu
in the top right and choose Background > soccerball_color.rat. Now
you can see this texture in the background of the UV panel. This
texture has a team insignia at the center of the image and a darker
area at the top for the dark patches.

03 RMB-click on the Handle tool and turn on the Min:

handle. Use the arrow handle at the bottom and pull it
up until all the patches are inside the dark area at the top of the
texture map.
If you had placed this node after the fuse or subdivide then there
Drag up
wouldn’t have been boundaries to work with. Houdini’s ability to let
you set up UVs in the middle of an existing network offers a lot of
flexibility.

04 Add another UV Flatten node after the first uvflatten

HOUDINI FOUNDATIONS

Drag Down node and before the Fuse node. In the Group field, enter
the expression @patches<20. This will flatten the light patches.
RMB-click on the Handle tool and turn on the Min: handle. Use the
arrow handle at the top and pull it down until all the patches are
inside the light area at the bottom of the texture map.
When you finish you can RMB-click on the Handle tool and turn off
the Min: handle.
RMB-click in the Scene view and from the menu, select Texture
Visualization > Off. This will remove the grid.

58
 05 Select the uvflatten node and click on the Handle tool.
Mouse over the geometry you can see the patches
highlight. Click on the Pin Vertices button in the Operation
Controls bar at the top of the viewport.
Mouse over the patch shown in this image and in the 3D view,
Shift-click on the center vertex of the desired patch to select it. This
adds a pin in the UV view and a handle to work with.

Shift-Click

06 Go to the UV view and move the pin and the patch so

that it is centered on the logo. Press Y to get a rotate
handle and rotate the patch until you line up the logo.
By pinning vertices on patches in the UV view, you can lock down
their position. At first there is some overlap of the neighboring
patches but you can easily fix this by repacking.

07 Click on the Repack button in the Operation Controls

bar to reorganize the other patches around the new one.
You can continue moving the patch around using the pin but will
need another Repack to avoid overlapping UVs.

08 Set the quickshade node to Bypass to hide the

assignment of the texture map. This node was only
needed when you were setting up the UVs. You now see a UV grid
on the soccerball.
To hide the UV display in the perspective view, got to the Display
Options bar and turn Off the Show UV Texture when UV’s Present
button.
Add a Null node to the end of the chain and call it GEOMETRY_
OUT. It is a good idea to have this kind of node to define the end of
Show UV Texture
the network chain. Save your work.

UV FLATTEN
The UV Flatten node works in two steps. It takes individual texture pieces, defined by
seams, and flattens them into 2D texture space, trying to equalize polygon size.
SETTING UP UVS

This node lets you add constraints for the flattening algorithm. Constraints force the
layout algorithm to satisfy certain conditions, giving you extra control over the final
UV layout. You can use this node interactively, using the tools in the node’s state to
specify constraints, or you can turn off Manual layout and use the node procedurally.

59
 PART FIVE
Layout: Cameras and Lights
To create a scene for rendering, you are going to bring the geometry into the Solaris or LOPS context of Houdini. This is an
environment dedicated to lookdev, layout and lighting and is built on the foundation of USD (Universal Scene Description).
This will allow you to render to the Karma renderer which works right in the Scene View as part of the Solaris workflow.

01 In the Network View, press tab > Match Size then add
the node between the subdivide and GEOMETRY_OUT
null. Set the Display flag on the GEOMETRY_OUT node.
Select the matchsize node and set Justify Y to Min to raise the
ball up to sit on the ground. This will put it in the right position for
rendering.

02 Change the desktop to Solaris. Make sure you are

looking at the Stage in the path bar.
In the Network view, press tab > Scene Import and click to place
the node down. Next to the Force Objects field, click on the node
selector button and from the pop up window, select the soccerball_
geo object then click Accept Pattern.
In the Scene View, use your view tools such as spacebar-h for
homing the view to get a better look at the soccerball.
Note: Force Objects is used instead of the Objects field because it will
bring the object into LOPS even if the object’s display flag is off.

03 In the Network view, press tab and type out Grid. Click
to place down the node and rename it backdrop. Double-
click on the backdrop node to dive down to the geometry level.
Select the grid node and set the size to 80, 80 and the Center to
0, 0, -20. RMB-click on the grid node’s output and type Bend. Click
to place the node then set its Display Flag and set: Bend to 75.
In the Capture section, set Capture Origin to 0, 0, -30, Capture
Direction to 0, 0, -1, and Capture Length to 5.
RMB-click on the grid node’s output and type Subdivide. Set its
Display Flag then set Depth to 2.

04 Go back to the Stage level. Wire the backdrop node

HOUDINI FOUNDATIONS

into the sceneimport node. RMB-click on the output of

sceneimport and type out Camera. Press Enter to place the node
then set its Display Flag.
In the Scene View, you will see camera handles at the origin. Zoom
out and look down at the scene then adjust the handles so the
camera is looking at the soccerball from the left. You may want
to activate the Construction plane so that you move the handles
along the ground. You can also use the axis handles to control the
direction and lift the camera up from the ground.

60
 05 In the top right of the Scene View, click on the No cam
menu and choose camera1. Now you are looking through
the camera and can adjust how it looks.
This is probably not the view you are looking for therefore some
view changes are needed. On the right side of the Scene View,
click on the Lock camera to view button. Now use the View
tools [Spacebar-LMB/MMB/RMB] to reposition the camera.
IMPORTANT: When you finish, toggle off the Lock Camera button.

06 From the LOP Lights and Camera shelf, click on the

Environment Light tool then press Enter to place it at the
origin. Set the Intensity to 0.5 to tone it down a bit.
Now click on the menu just to the left of the camera menu, and set
Denoise it to Karma. Now you are using the Karma renderer in the viewport.
If you have an Nvidia graphics card and have the latest drivers
installed, you can turn on the Optix Denoiser to resolve the image
more quickly. Turn it on in the Display Options bar or press d and
setting Enable Denoising in the Render Display Options.

07 From the LOP Lights and Camera shelf, click on the

Point Light tool then press Enter to place it at the origin.
You are looking through the light. Set the camera back to camera1.
With the node active, press Shift-F to turn on the Shadow mode.
You can also click on it in the Operation Control bar. Now click
on the top of the soccerball to set a pivot point then Shift-click to
place a target on the ground. Ctrl-Drag to set the light distance.
Now you can use Ctrl-Shift-drag to change the intensity of the
light. You may need to set it quite high go see some impact on the
look of the soccerball.

08 In the Network view, RMB-click on the pointlights node’s

output and type Light then press Enter to place the node
and set its Display Flag.
With the node active, press Shift-S to turn on the Specular mode.
Now click on the right side of the soccerball to define a specular
area of focus.
Now you can use Ctrl-drag to move the light away from the soccer
ball and Ctrl-Shift-drag to change the intensity of the light.
Layout: Cameras and Lights

09 In the Network view, RMB-click on the lights node’s

output and type Light Mixer then press Enter to place
the node and set its Display Flag. This will create a special panel in
Solo
the Parameter pane which has a list of lights on the left side.
Drag the three lights from the list to the area on the right. Click on
the Star icon to Solo each light to determine its contribution then
tweak Exposure to adjust the lighting. Since the intensities are so
high you can click on the icon above the intensity bar and from the
pop-up set a Max value that works for your shot. When you are
finished be sure to turn off the Solo button to see all the lights.

61
 PART SIX
Lookdev: Materials
Materials and shaders can also be created within the LOPS/Solaris context. This involves adding the materials to the Scene
Graph then assigning them to the geometry. The Materials are created inside a Material Library node then assigned at the
LOPS/Solaris level. To add textures to the backdrop, UVs will have to be created to position the maps properly.

01 In the Network view, press tab > Material Library and

place the node just above existing network of nodes
then wire it into the backdrop node. Double click on the node to go
down to the VEX Builder level.
Press tab > Principled Shader and place the node down. Rename it
soccerball_mat. Change its Base Color to white (1, 1, 1).
Alt-drag on this node to create a second principled Shader and
rename this one backdrop_mat. Change its Base Color to a dark
green.

02 Go back to the Stage level then press tab > Assign

Material and place the node in between the sceneimport
and the camera nodes.
In the Parameter pane, next to Primitives, click on the arrow button
and in the viewport select the soccerball geometry. Press Enter to
add its path to the Primitives field. Now click on the arrow next
to Material Path and from the pop-up window, select materials >
soccerball_mat and click OK.
Click on the + button then repeat these steps for the backdrop and
the backdrop_mat.

03 Double-click on the materiallibrary node to dive into it

and select the soccerball_mat node. Click on the Textures
tab and under Base Color click on Use Texture then use the button
next to Texture to call up the file window. Click on $HIP in the side
list then click on the tex folder to open it and then click once on
soccerball_color.rat to select it. Click Accept to assign the texture to
the material.
The $HIP reference makes sure that the reference is relative to the
location of your scene file. That way if you were to move your project
directories to another computer the reference will still work.

04 Under the Textures tab and use the technique you

HOUDINI FOUNDATIONS

learned in the last step to assign textures to Roughness

and Reflectivity. You will find the appropriate textures in the tex
folder.
Go to the Bumps & Normals tab on the material and click the
Enable button. Click on the arrow next to Texture Path and from
the tex directory choose the soccerball_normal.rat file. Set the Effect
Scale to around 0.5 and see how it looks.

62
MATERIALS IN HOUDINI
Materials in Houdini live in the VEX Builder context which in this
case is nested inside the Material Library node. A material is made up
of VOP nodes or Material X nodes that define the material qualities.
The Principled Shader is an uber material that can be used on its own
to assign texture maps and achieve a large variety of looks. You can
also build your own shaders and materials for more advanced looks.

05 In the Scene view, look through camera1 to see your

shot. The position of the soccerball’s logo isn’t quite right
in relation to this camera therefore you need to rotate the ball.
In the Scene view, select the soccerball. Press tab > transform. This
adds a transform node to the end of the chain. With the Handle
tool, press r to get the Rotate handle and rotate the ball see the
logo properly. You may need to tumble around to get this to work.
You can do this in the Karma or Houdini GL view. If you don’t like
the edits then you can delete the node and try again. To deselect
the soccerball_geo, press Ctrl and click on it in the Scene Graph.

06 Now lets add some texture maps to the Backdrop

material. Double-click on the materiallibrary node and
select the backdrop_mat node.
Set the Base Color to 1, 1, 1 because this color will be multiplied
with the texture map. Now click on the Textures tab and under
Base Color click Use Texture. Click the File Selector button and use
$HIP to go to the /tex directory and choose backdrop_color.rat.
You can also add the backdrop_reflect.rat texture to Reflectivity.
You can see in the Scene View that UVs are not set up properly and
the texture map isn’t working.

07 Go back to the Stage level then double-click on the

backdrop node to go to the geometry level. Add a UV
Project node between the grid and the bend. You are putting the
nodes here so the UVs are created before the surface is bent.
On the uvproject node, click on the Initialize tab and click the
Initialize button. Go back to the Transformation tab and set V
Range to 0, -1. This will orient the UVs properly.
IMPORTANT: Set the Display Flag back to the subdivide node.

08 Go back to the Stage level. In the Network View, press

tab > Karma to add a Karma Render Settings and USD
LOOkDEV: MATERIALS

Render ROP node. Wire them into the end of the chain. Select the
karmarendersettings node and on the Image Output > Filters tab set
Denoiser to nvidia Optix Denoiser to turn the denoiser back on.
Select the usdrender_rop node. Click on the Render to Mplay
button. This opens Mplay which shows the rendering as it
progresses. Choose File > Save Frame As to save the image to disk.

63
 PART SEVEN
Rig the Soccerball
In order to create an animation of the ball bouncing, you will start by building a simple rig that will make it easier to keyframe.
This will involve setting up null objects so that you can work interactively in the viewport and adding nodes to the soccerball
geometry network to accommodate the ball rotation along with squash and stretch.

01 Change back to the Build desktop and navigate to the

object level by clicking on one of the path bars and
choosing obj. Now in the network editor, Alt-drag on soccerball_geo
to make a copy of it. Rename this node socccerball_anim.
You will use soccer_anim to build your rig. Now turn off the Display
Flag on the soccer_geo node to hide it. You don’t want to make
changes to the original setup because that object is being used in
Alt-drag SHOT 1 in the Solaris context. This new soccerball will be used for an
animated SHOT 2.

02 On the Create shelf, click on the Null tool then press

Enter to place it at the origin. Name it soccerball_ctrl. Go
to the Misc tab and set Control Type to Circles and Orientation to
ZX Plane. Set the Display Uniform Scale to 4. This creates a handle
for the rig that is easy to select that won’t render later on.
In the network editor, connect the input of the soccerball_anim
object to the output of the soccerball_ctrl null to create a child/parent
relationship. Moving the null will move the ball. Turn off the selection
flag on the soccerball_anim so that you don’t select it by accident in
the viewport while animating. You will use soccerball_ctrl instead.

03 Select the soccerball_ctrl node. In the Parameter pane,

click on the Transform tab then RMB-click on the
Translate X parameter. Choose Copy Parameter.
Dive into the soccerball_anim object. Add a Transform node
between the subdivide and the matchsize nodes. RMB-click on
the Rotate Z and choose Paste Relative References. This places a
channel reference expression in this parameter.
ch(“../../soccerball_ctrl/tx”)
This will connect the movement of the control object to the
rotation on this node.

04 Click on the parameter to expand the channel. You

HOUDINI FOUNDATIONS

are now going to use the ball’s circumference (2πr) to

determine the ball’s rotation as it moves forward.
Edit the expression to read:
-ch(“../../soccerball_ctrl/tx”)*360/(2*$PI*1.1)
First you add a negative (-) at the front. You then multiply the position
of the ball by 360 degrees and divide by 2πr - π being $PI in the
expression. At the object level, move the soccerball_ctrl along the X
axis. The expression will rotate the ball to match the motion. Put it
back at the origin when you are finished.

64
05 In the Viewport, create another Null object at the origin.
Call it squash_ctrl. Go to the Misc tab and set Control
Type to Box and the Display Uniform Scale to 0.2.
Move the null up just above the ball. Translate Y should be about
2.5. In the Parameter pane, choose the Modify Pre-Transforms
menu and select Clean Translates. This sets the Translate Y value
of the null to 0 even though it is above the ground. In order for this
null to drive the squash and stretch, it needs a default value of 0.

06 Parent the squash_ctrl null to the soccerball_ctrl null. This

will ensure that this secondary null moves when you
animate the control null.
RMB-click on the squash_ctrl node’s Translate Y parameter. Choose
Copy Parameter. You will use this parameter to drive the squash
and stretch of the ball. This will allow you to control the squash and
stretch from the viewport interactively.

07 Go into the soccerball_anim object. Add a Bend node

after matchsize. Turn Off the Limit Deformation to
Capture Region checkbox.
Go to a Right View then click the Set Capture Region button. Turn
on Grid Snapping and place a point at the base of the ball and
another at the top. This should set Up Vector to 0, 0, 1, Capture
Direction to 0, 1, 0 and Capture Length to 2.2.
Turn on Length Scale and Preserve Volume then RMB-click on
Length Scale and choose Paste Relative References. Add a +1 at
the end of the expression.

08 Go to the Object level and a perspective view. RMB-click

on the Transform X and Transform Z parameters and
choose Lock Parameter to lock these parameters on squash_ctrl.
RMB-click on the Scale and Rotate parameters and choose Lock
Parameter to lock all three channels.
Select the soccerball_ctrl object. Lock all the channels except
Translate X and Translate Y. Now when you select the controls
you will only see handles for the unlocked channels. This will
make it easier to work with the rig because the animator can only
manipulate the chosen parameters.

09 Now test out the rig by moving it around in X and Y and

using the second handle to squash and stretch it. Once
you are sure that all the parts are working, return all the values to 0
Rig the Soccerball

and get ready to animate.

You may want to turn off Secure Selection in the toolbar on the
left side of the Scene view. This will make it easier to select the two
control nulls while in the Move tool. If not then you will need to
press the S key every time you want to switch selections.
Save your scene file before proceeding.

65
 PART EIGHT
Animate a Bouncing Ball
You can now take the soccer ball rig and use it to animate the ball bouncing. You will learn how to set keyframes, adjust anima-
tion curves and work with time-space handles in the viewport. The bouncing ball is a classic animation exercise that offers a
great opportunity to learning the basics of animating in Houdini.

01 At the bottom left edge of the Timeline, click on the

Global Animation Options button. Set the End to 120
and click Close. This will set the timeline range to 120 frames.
Make sure you are on frame 1. Click on the Pose tool on the
toolbar to your left and select the soccerball_ctrl. Move the ball to
around -15 in X and press K to keyframe it. Move the timeline to
frame 120. Move the ball to the around 15 in X and press K to set a
second keyframe.
Scrub through the timeline to make sure that the ball is animating.
It should be moving and rotating based on the rig design.

02 Move the timeline to frame 12 and press K to set an

intermediate key. Repeat at frames 36 and 60. All of
these keyframes are sitting on the ground.
Now go to frame 1 and lift the ball up in the Y direction. You don’t
need to set another key because this move simply updates the
keyframe you already set at frame 1 .
Move to frame 24 and lift up the ball in the Y direction a little less
than you did at frame 1. Press K to set a keyframe. Move to frame
48 and lift the ball up even less. Press K to set another keyframe.

03 Scrub through the timeline to see that the ball appears to

float whereas you want hard hits when the ball contacts
the ground. Click on the Animation Editor pane tab.
From the Scoped parameter list, click on the Translate Y channel.
Press H to home the view of the curve. Select the three keyframes
where the ball contacts the ground and press the Untie Handles
button on the Functions bar found just above the graph. Now click
in empty space to deselect then start tweaking the tangent handles
to create a sharp bounce at each point. You can also stretch out the
handles at the top to slow the ball down at the peak.

Select
HOUDINI FOUNDATIONS

THE POSE TOOL Secure Selection [~]

You can easily manipulate the two control objects in the soccerball rig
using the Move or Handle tool. The advantage of using the Pose tool Move [T]
is that it gives you access to the Motion Path handle and if you were Rotate [R]
working with Inverse kinematics, there are special handles that you can
use to control the system. Therefore be sure to remember this tool when Scale [E]
you start setting keyframes on your rigs. The Secure Selection does not POSE [Y or Ctrl R]
restrict the Pose tool from selecting different objects.
Handle [Enter]

66
 04 Go back to the Scene view and preview the results.
You now have a sharper hit each time the ball hits the
ground. Make sure you have the Pose tool active and the
soccerball_ctrl node selected. In the top bar, turn On the Motion
Path handle. This shows a path outlined where the ball is bouncing.
Click on each keyframe marker to tweak the bounce. RMB-click on
the handle to Show tangents for more control over the curve.
Go to the soccerball_anim object’s Misc tab, set Onion Skinning to
Full Deformation. Press spacebar-d and from the Scene tab adjust
Frame increment and the Frames Before and After color.

05 To adjust the timing, you can also use the timeline. Press
Shift and drag a bounding box from frame 1 to the last
key in the timeline to select all the keys. Next, MMB drag on the
end of the box underneath the handle to scale the timing of the
bounces to speed them up. You can also select each key using MMB
and then drag with MMB to time each keyframe the way you want.
This is where you will determine the timing of the bounces. Keep
exploring until you get the look that you want. Note that there may
be some awkwardness in the bouncing because of the translate X
MMB Drag Here
values. You will fix that in the next step.

06 Click on the Animation Editor pane tab and you will see
two curves. From the Scoped Parameters list, click on
Translate X for soccerball_ctrl. Now select all the keys except for
the first and the last. Press Delete. Now use the curve handles to
go from a high slope to a low slope. This will have the ball moving
faster at the beginning and slower at the end.
Note that if you go back to tweak the points in X on the motion
path handle, you will get strange results because there are no
longer intermediate keys in that direction - only use it to tweak in Y
from this point on.

07 Go back to the Scene view. RMB-click on the Motion

Path handle and choose Persistent. This will keep it
around as a guide as you set keyframes on the squash and stretch.
Select the squash_ctrl and turn off its Motion Path. Go to the first
bounce and move back one frame. Select the squash_ctrl handle
and stretch out the ball a bit. Set a Keyframe using the K key.
Now go to the bounce frame and move the handle down to create
squash. Set another key. Go one frame forward and stretch the ball
until it is round. Set another key. Repeat for all the bounces. Animate a Bouncing Ball

08 When you are finished, scrub and playback the motion to

preview the results. Make sure the peaks of the bounces
are stretched out. Make sure the Real Time Toggle is On in the
Timeline to properly evaluate the motion.
You can now use the Animation editor to make tweaks to the
squash and stretch. Make sure you keep the keyframes aligned with
the bouncing of the soccer balls.

67
 09 Select the soccerball_ctrl null object. RMB-click on the
Translate Y and choose Motion FX > Noise. A panel pops
up with parameters that you can use to control the noise. A new
subnetwork was created with the CHOP nodes that send their
information back to the soccerball_ctrl’s Translate Y channel.
Set the Amplitude to 5 and press Play to see how this looks. This
adds dramatic up and down motion which make it feel like there is
some serious turbulence. Set the Amplitude to 1. This offers a more
subtle bump to the motion of the ball.

10 Some of this motion goes below the ground. You will

want to focus on creating bumps above the ground.
Go back to the object level. Select the soccerball_ctrl null object.
RMB-click on the Translate Y and choose Motion FX > Limit. Set
Minimum to 0 and Maximum to 6. Now the ball moves flat with
some bumps instead of going up and down the whole way.

11 There should not be any noise while the ball is

bouncing. It is only needed when the ball is rolling.
You can keyframe the Amplitude to turn the noise on and off.
In the newly created motionfx network, select the noise1 CHOP
node. Go to Frame 37 where the ball stops bouncing and starts
rolling. Alt-click on Amplitude to set a keyframe. Go to Frame
1 and set Amplitude to 0. Alt-click on Amplitude again to set
a second keyframe. Go to the Animation Editor and select the
curve. Click on the Constant button in the Functions bar. This
creates a sharp cut from no amplitude to an amplitude of 1.

12 When you are finished, close the window then RMB-

click on the Motion Path handle and turn off Persistent
so that it isn’t visible when you deselect.
In the toolbar at the left side of the Scene view, click on the
Render Flipbook button. Leave the default settings and click Start.
Wait while the sequence is captured and then you will see the
flipbook in an Mplay window. You can Play this back and scrub
through to evaluate your motion.
Save your work.
HOUDINI FOUNDATIONS

MOTION FX
While keyframes and animation curves are stored in the parameters of your
nodes, you can also use channel operators (CHOPs) for a more procedural
node-based approach to working with motion.
Motion FX can be applied to keyframed motion which is extracted and stored
in a Channel CHOP. You can then apply effects such as cycle, noise, smooth,
limit or lag to the existing motion. On the Constraints shelf, you have tools
which let you have one parameter either look at, lag or jiggle behind another.

68
PART NINE
Lights, Camera, Action!
To render out the animated soccer ball, you will need to go back to the Solaris environment and set up a second shot. You will
begin by branching off new LOP nodes from backdrop geometry then adjust the lights and cameras to suit the bouncing soccer-
ball animation. You will also set up motion blur for the deforming geometry.

01 Go to the Object level. Double-click on soccerball_anim

to dive into it. Press n to select all the geometry then go
to the Modify shelf and select Extract. This takes all the motion and
bending of the ball and puts it in one network.
You can see an objectmerge node which is extracting the ball into
a new object called extract_object. RMB-click on the output and
find USD Export then click to place this node. Rename this node
to soccerball_anim. Click on the Export tab then set Valid Frame
Range to Render Frame Range and set Output File to $HIP/geo/
soccerball_anim.usd. Click the Save to Disk button.

02 Go back to the Solaris desktop and point it to /stage.

In the Network View, add a Null node just before the
karmarendersettings node and call it SHOT_01.
Disconnect the three light nodes (not the lightmixer) and move
them above the backdrop node. This will not change how the first
shot looks, but will let you share the nodes between the two shots.
Move the backdrop, light and materiallibrary nodes to the right.

03 Zoom in and add a Reference node to the right under the

backdrop node. Connect the backdrop node to this new
node and set its Display Flag. Next to File Pattern, click on the File
Chooser and find the soccerball_anim.usd file. Rename the node to
soccerball_anim.
Scrub in the timeline, in Houdini GL view, to see the cached
animation which is part of the USD file.

04 Use the Select tool and click on the new animated

Lights, Camera, Action!

soccerball. In the Scene View, press tab > Transform to

add a transform node to the graph.
In the Scene view, use the Transform handle to Move the Ball to
the back of the backdrop in the middle. Scrub the timeline to watch
the ball bounce to the right. Leave it at somewhere around frame
80. Press R to get the rotate handle. Rotate to move the ball so that
it is bouncing at an angle from the backdrop. Scrub the timeline to
see if you like its direction.

69
 05 In the Network view, Select the assignmaterial node
from the SHOT 1 network then Alt-drag to create a
copy of this node. Wire the transform node into the assignmaterial
node then set its Display Flag. This will assign the material to the
backdrop but since the soccerball primitive has changed it needs to
be reassigned.
In the field next to Primitives for the soccerball_mat , change the
primitive name to /soccerball_anim to reassign the material to the
new geometry.

06 Tumble around until you see the ball animated towards

the camera from the top left to the bottom right. In the
LOP Lights and Cameras shelf, Alt-click on the Camera tool to
place a camera from the angle you are currently looking.
Press the Lock Camera/Light to View button so that view changes
can be used to reposition the camera. Now Tumble, Pan and Dolly
in the viewport to tweak the camera to get the framing that you
want for the shot. Scrub the timeline to make sure the camera
works for the whole sequence.

07 Add a Light Mixer node after the camera. On the

lightmixer node you will need to move over the lights. This
will let you use the same light handles you learned about earlier to
make lighting decisions for this shot and use the Karma display in
the viewport to verify your setup. You can also use the lightmixer
node to play with intensity and exposure for this shot.
These edits are being held in the lightmixer node and changes are
not being made to the original lights. The lightmixer lets you tweak
existing lights when working in a multi-shot setup.

08 In the Network view, Alt-drag the SHOT_01 and

karmarendersettings and usdrender_rop nodes. Wire the
lightmixer node into this chain. Select the new karmarendersettings
node and make sure that Camera is set to /camera2. Set Valid
Frame Range to Render Frame Range and set the Output Picture
to $HIP/render/anim/soccerball_anim_$F2.exr. The $F2 adds frame
numbers to the renderings with a padding of two and the /anim/
creates a directory to hold these frames.
On the usdrender_rop node, click Render to Disk. When you
finish, choose Render > Mplay > Load Disk Files and open up the
rendered images to review the final sequence. Save your work.
HOUDINI FOUNDATIONS

KARMA RENDERER
Karma is a physically-based HYDRA renderer built to work with USD files
in the Solaris/LOPS context. This allows it to be used in the Viewport for
interactive updates or to be rendered to disk using a Karma node.
Note: Houdini 19 includes a preview for an upcoming Karma XPU render
engine. This hybrid GPU/CPU renderer is Alpha with many features still
under development and is for testing purposes only. You can choose XPU
in the Scene view’s Display Options or in the Karma node.
Viewport Menu Display Options

70
PART TEN
Set up a Rigid Body Simulation
While traditional animation is great for animating a single soccer ball, dynamics would be a better option if you want to animate
a bunch of soccer balls. Dynamics requires a simulation so that the solver can go frame by frame determining how each of the
participating objects interact with each other. You will use packed geometry to get an efficient result for this simulation.

01 Change back to the Build desktop and navigate to the

object level. Hide all of the animation rig nodes and the
extract_object node by turning off their display flags. Turn on the
soccerball_geo display.
Select the soccerball_geo node then from the Modify shelf click
on the Extract tool. This creates a new object with the soccerball
object merged. Jump up one level and rename extract_object to
soccerball_sim. Hide the soccerball_geo object.
Dive back in to the soccerball_sim object to work with the geometry.
Add a Match Size node to center the ball around the origin.

02 In the Network view, press tab > Box then place it to the
right of the matchsize node.
Set the following on the box node:
ƒ Center to 0, 8, 0
ƒ Rotate to 45, 45, 45
ƒ Primitive Type to Polygon Mesh
ƒ Uniform Scale to 6
ƒ Axis Divisions to 3, 3, 3
This puts it in the right position for the simulation.

03 In the Network view, add a Copy to Points node just

below the other nodes. Wire the matchsize node into the
first input and the box node into the second.
Turn ON the Pack and Instance option. This will create a faster
simulation because the geometry is being instanced to the points of
the cube. Set the copytopoints node’s Display Flag.
In the Network view, press tab > Mountain and place the node
between the box and the copytopoint nodes. Turn Off the Noise
Along Vector option then set Amplitude to 2 and Range Values to Set up a Rigid Body Simulation
Zero Centered. This will jiggle the points on the box.

04 Make sure you are on Frame 1. Add a RBD Bullet

Solver node after the copytopoints node. Click on the
Collision tab, scroll down to Ground Collision, and set Ground
Type to Ground Plane to Ground Plane. Press Play to test out the
simulation. The sim is cached which lets you scrub in the timeline to
review the results.
Under the Collisions tab, set Bounce to 0.8. Under the Properties
tab, set Density to 10, Bounce to 1.1. At the top of the Parameter
pane for this node, click the Reset Simulation button and then
press Play to resim. Scrub to evaluate.

71
 DOPS hidden inside SOPS
In Houdini, simulations are processed using the Dynamic Operators or DOPs.
With the RBD Bullet Solver node in the Geometry/SOP context, you are
working with a node that has a Dynamics network buried inside it.
This makes it easy to set up at the geometry level with all the DOP nodes are
wired up and ready to go but hidden from view. For simpler setups, working at
the geometry level will give you a proper simulation. If you need more control
over the different solvers then you would need to work directly in DOPs.

05 At the end of the chain, add a USD Export node, set its
Display flag and rename it to soccerball_sim.
Set Valid Frame Range to Render Frame Range and set the Output
File to $HIP/geo/soccerball_sim.usd.
Click on the Save to Disk button and this will save the USD file into
your geo directory. You will reference this cached asset into the
Solaris setup as a third shot.

06 Change your Desktop back to Solaris and set the path to

/stage. Make sure you are choose Houdini GL from the
persp menu.
Alt-drag the soccerball_anim Reference node and set its Display
Flag. Set File Pattern to $HIP/geo/soccerball_sim.usd.
Rename this node to soccerball_sim.

07 In the Network view, Select the assignmaterial node from

the SHOT 2 network then Alt-drag to create a copy of
this node. Wire the soccerball_anim node into it the assignmaterial
node then set its Display Flag. This will assign the material to the
backdrop but since the soccerball primitive has changed it needs to
be reassigned.
In the field next to Primitives for the soccerball_mat , change the
primitive name to /soccerball_sim to reassign the material to the
new geometry.

08 Tumble around until you see the balls animated towards

HOUDINI FOUNDATIONS

the camera. In the LOP Lights and Cameras shelf, Alt-

click on the Camera tool to place a camera from the angle you are
currently looking.
Press the Lock Camera/Light to View button so that view changes
can be used to reposition the camera. Now Tumble, Pan and Dolly
in the viewport to tweak the camera to get the framing that you
want for the shot. Scrub the timeline to make sure the camera
works for the whole sequence.

72
 09 Add a Light Mixer node after the camera. On the
lightmixer node you will need to move over the lights. This
will let you use the same light handles you learned about earlier to
make lighting decisions for this shot and use the Karma display in
the viewport to verify your setup. You can also use the lightmixer
node to play with intensity and exposure for this shot.
These edits are being held in the lightmixer node and changes are
not being made to the original lights. The lightmixer lets you tweak
existing lights when working in a multi-shot setup.

10 In the Network view, Alt-drag the SHOT_02 and

karmarendersettings and usdrender_rop nodes.
Wire the new lightmixer node into this chain. Select the new
karmarendersettings node and make sure that Camera is set to /
camera3.

11 On the usdrender_rop node, set Valid Frame Range to

Render Frame Range and set the Output Picture to $HIP/
render/sim/soccerball_sim_$F2.exr. Click Render to Disk.
When you finish, choose Render > Mplay > Load Disk Files and
open up the rendered images to review the final sequence.

CONCLUSION
You have now built a scene from scratch, touching on many different
aspects of Houdini. You have modeled, set up textures, animated, ren-
dered and simulated. Along the way you have learned about the different
Houdini contexts and how to navigate back and forth between them.
While this lesson doesn’t result in blockbuster VFX,
it introduces you to fundamental skills which
you will carry with you as you dive deeper
SET UP A RIGID BODY SIMULATION
into Houdini and begin exploring its
comprehensive toolset.
There is a wealth of learning material
available on the SideFX website to
help you take your next steps.
Best of luck on
your journey!

73
HOUDINI FUNDAMENTALS

NODES, NETWORKS &

DIGITAL ASSETS
A great way to understand Houdini’s node-based workflow is to explore it in the context of a project. It is
important to start learning how to think and work procedurally. In this lesson, you will learn how to create
your own custom brickify tool using procedural nodes and networks to define its function and interface.
Along the way you will get to use different aspects of Houdini’s workspace. Be sure to refer to the
overviews in the introduction to remind yourself of how these UI elements work together. The lessons will
then give you a chance to put your ideas into practice, which is one of the best ways to learn.

LESSON GOAL
ƒ To create a custom tool that turns any given 3D shape into toy bricks.

WHAT YOU WILL LEARN

ƒ How to model a plastic interlocking brick
ƒ How to break down a default rubber toy shape into a grid of points.
ƒ How to use packed primitives and instancing to speed up interaction.
ƒ How to use attributes to color the bricks using a texture map.
ƒ How to work with nodes and networks to control the flow of data
ƒ How to create a Digital Asset to package up and share
your solution with others.
ƒ How to animate the bricks appearing over time.

LESSON
COMPATIBILITY
Written for the features in Houdini 19.5+

The steps in this lesson can be

completed using the following
Houdini Products:

Houdini Core ✔
Houdini FX ✔
Houdini Indie ✔
Houdini Apprentice ✔
Houdini Education ✔

Document Version 3.0 | July 2022

© SideFX Software
75
 PART ONE
Create a Single Brick
To get started, you will build a single brick model that you will later copy onto points to create the
brickified shape. You will create this shape using a combination of polygon modelling tools. Along
the way you will see how each action you take creates a node in Houdini that creates a recipe of the
steps taken to create the geometry.

01 Select File > New Project. Change the Project Name to

brickify_lesson and press Accept. This creates a project
directory with subfolders for all the files associated with this shot.
Select File > Save As... You should be looking into the new brickify_
lesson directory. Set the file name to bricks_01.hip and click Accept
to save.

02 In the viewport, press c to bring up a radial menu. From

this menu, choose Create > Geometry > Box. Your cursor
now shows the outline of a box waiting to be placed in the scene.
Press Enter to place it at the origin. In the Operation Controls bar,
set Size to 0.2, 0.2, 0.2 and Axis Divisions to 3, 2, 3.
You can see that there is a box_object in the Network view. This
object level node contains the transformation information for this
shape. The Operation Controls show you parameters from another
box node one level down.

03 Click on the Select tool then Press 4 to go to primitive

selection mode. Select the top four faces on the box.
Press c to bring up a radial menu and choose Model > Polygons
> PolyExtrude. In the Network view you can see the box node
feeding into a polyextrude node.
In the Parameter pane, use the slider to set Inset to 0.04 which
creates new polygons on the top surface of the box. Each node
contains parameters relevant to the purpose of that node. These
are geometry nodes that are otherwise known as surface operators
or SOPs.

04
HOUDINI FUNDAMENTALS

Next, press the T key to call up the Move tool. This adds
an edit SOP node into the network. In the viewport
RMB-click in empty space to bring up a menu and select Make
Circle to round out the selected polygons.
This menu is associated with the edit node. Every node has its own
interface that you can access as long as a relevant tool is active. In
this case, the Move tool gives you access to the handle. In other
cases the Handle tool would be used to access the interactive
handles for a node.

76
 05 Press c to bring up a radial menu and choose Model >
Polygons > PolyExtrude. In the Scene View pane, use the
handle to drag up the polygons and set Distance to 0.05.
Tumble around and press s to go into select mode then select the
bottom four polygons of the box. Press q to repeat the PolyExtrude
tool. In the Parameter pane, use the slider to set Inset to 0.025.
Press q to repeat that tool and set a Distance of -0.175. When you
finish, tumble back to see the top of the brick.

06 Press 3 to go to edge selection and press n to select all

the Edges. In the Scene view press tab and start typing
Group... Select Group and in the Parameter pane, set the Group
Name to bevel_edges.
Next, set Enable to OFF under Base Group and then set Enable to
ON in the Include by Edges section. Turn on Min Edge Angle and
set it to 89 and then turn on Max Edge Angle and set it to 91.

07 Press s to go to the Select tool. Press 9 to turn on the

“Select Groups” option. In the popup window, click on
the bevel_edges group.
In the viewport, press c to bring up a radial menu. From this menu,
choose Model > Polygons > PolyBevel. This adds a polybevel node
and automatically fills in the Group field with bevel_edges.
Now set the Bevel Offset to 0.006. Under Fillet, set Shape to
Round and Divisions to 3.

08 Go to the object level and in the Network view, rename

the object to single_brick. With the brick selected,
press Shift + to turn on subdivision surface display for this shape.
Deselect the brick to see the model subdivided. If you see wire lines
on your object, press v and from the radial menu, choose Shading >
Smooth Shading to hide them.
Save your work.

SUBDIVISION DISPLAY
Create a Single Brick

You can use Shift + and Shift - to turn subdivision display on and
off on selected polygonal objects. This creates a viewport subdi-
vide that lets you see what the shape would look like subdivided.
These hotkeys set the Display As parameter which can be found
at the object level on the object’s Render tab.
Your object will not render with subdivisions unless you turn on
Render Polygons As Subdivision Surfaces on the same tab.

77
 PART TWO
Copy Bricks to a Point Cloud
You are now going to create a cloud of points that match the shape of a particular piece of
geometry. You are then going to instance the bricks to the 3D grid to create a brickified version. The
instancing is generated by packing the brick geometry then instancing them to the points.

01 In the Scene View, press tab and start typing Test... then
choose Test Geometry: Rubber Toy. Press Enter to place
it at the origin.
Press i to dive into the test geometry object. Set the following:
ƒ Uniform Scale to 3
Now add a Match Size node and set Justify Y to Min. This raises the
toy so that it sits on the ground.

02 In the Network editor, RMB-click on the output of the

matchsize node and type Points... then select Points from
Volumes and place it’s node in the network. Now set its Display
flag to focus on the output of this new node.
Press s to go to the select tool and press 2 to get point selection
and then press n to select all the points which will highlight in
yellow. You will copy the bricks to these points.

03 Press u to go back to the Object level and shorten the

name of the rubber toy to rubbertoy. In the Network
view, press the shift key and click on the rubbertoy and then the
single_brick.
From the Modify shelf tab, select Combine to bring these objects
together. You are taken to the geometry level and the nodes are
feeding into a merge node.

04 In the Network view, select the display_merge node and

HOUDINI FUNDAMENTALS

press the delete key. Now the polybevel node chain is

displayed and the other chain is not. Houdini lets you choose which
node you want to display which would be the node that will be
visible when you go back to the object level.

78
 05 RMB-click on the output of the polybevel nod, start
typing copy... and select Copy to Points node. Click to
place the node below the two chains and set its Display Flag.
Turn on the Pack and Instance option. This is important because it
will display the copied bricks much faster than if this option is off.
At this point there is an error on the node because we haven’t
connected the second input.

06 Click on the dot under the pointsfromvolume node then

connect it to the second input on the copytopoints node.
The bricks appear to be overlapping. Go back to the
pointsfromvolume node and set Point Separation to 0.2.
Now you are copying bricks to the points of the grid.

07 Click on the copytopoints node. Some of the bricks may

appear dark grey and don’t display the proper brick. This
is because of geometry culling in the viewport. You can fix this with
a display setting.
In the Scene View, press spacebar-d to bring up the Display
Options. Click on the Optimize tab and either set Scene Polygon
Limit to a number higher than 50 million or set Distance-based
Packed Geometry Culling to OFF. Close the floating panel. Now
you can see all of the bricks being copied to the points as Packed
Instances.

08 Before saving your work, let’s Organize the network.

Select the nodes that make up the single brick and press
Shift-O to create a network box around them. Click on the title bar
of the box and enter single brick. You can then collapse the box and
move it down to de clutter the network a bit.
Save your work so far.

Copy Bricks to a Point Cloud

PACKED INSTANCES
If you leave Pack and Instance setting OFF on the copy to points
node then you will end up with a large model with over a million
points and primitives. This would make it very slow to manipu- PACK AND INSTANCE | OFF
late in the Viewport because instancing is not being used.
If you turn it ON then the 338 points on the brick model are
packed up and instanced which leaves a much more efficient
point count for the copy to points node. PACK AND INSTANCE | ON

79
 PART THREE
Add Color and Switch to a Teapot
You are now going to add color to the points which will then be picked up by the instanced bricks.
At first this transfer of color only applies in the viewport but setting up a proper material lets you
set up the point colors for rendering the bricks. You will then set up a switch between the rubber
toy and a teapot to make sure that your network works with different shapes.

01 Add a color node between the pointsfromvolume node

and the copytopoints node. This will add color to the
points which will get copied to the bricks. Change the color to red
to help the bricks stand out against the background.
Click on the Render Region tool and drag a bounding box over the
rubber toy. This kicks off a test Render and you can see that the
bricks are not rendering as red. You need a material assigned that
picks up the color.
Click the x button in the top right of the render region to close it.

02 Press tab in the Network view and type Mat... Choose

the Material Network tool and click to place the node
down in the network.
Go to the Material Palette and close /mat and open up /matnet.
Drag a principled shader down into the matnet. Name it brick_mat.

03 Go back to the geometry network using the back button

in the Network view. RMB-click on the output of the
copytopoints node and type Material... Select material and place it’s
node in the network and set its display flag.
Click on the Operator Chooser button on the far right of the
Material parameter. From the pop-up window, navigate to and
highlight brick-material. Turn on the Export Relative Path Option
and click Accept.
The material is assigned but the bricks still don’t render.

04 Go back to the Material Palette and select the brick_

HOUDINI FUNDAMENTALS

material. Click on the Surface tab then set the Base Color
to 0.5, 0.5, 0.5 and turn ON the Use Packed Color option.
Click on the Render Region tool and drag a bounding box over the
rubber toy. The rendered bricks should now be red.
If not then click Render in the top bar of the Scene view to make
sure the changes have been applied.
Save your work so far. Click the x button in the top right of the
render region to close it.

80
 05 Go back to the geometry network using the back button
in the Network view. Press s to go to the select tool. In
the Network view, press tab and type the word Switch then from
the Sourcing folder drag it into the Network view.
Next, drag it over the line connecting the rubbertoy node to the
matchsize node. This inserts it into the network between the two
nodes.
This node will make it easier to switch between different incoming
shapes.

06 In the tool shelf, go to the Create menu and drag the

Platonic tool down to the network view. This places
a platonic node at the geometry level. Tools can be dragged into
the network view as long as the node type makes sense for your
current network level.
Set the platonic node’s Solid Type to Utah Teapot. Set Radius to
4.2. The volume from points node will update the points to match
the new volume to generate a different configuration of bricks.

07 Connect the output of the platonic node to the input of

the switch node. Now you can select the switch node and
change its Select Input to 1. The platonic shape has been cubified
using the same setup as the rubber toy.
This is one of the big benefits of a procedural system. Later you will
package up this network into a custom tool called a digital asset.
As a digital asset, it will be easier to share the network with others
and to manage the tool as it gets deployed in a studio environment
where changes and updates are inevitable as the tool evolves.

08 Use the switch node and set the Select Input to 0 to go

back to the rubbertoy. You can now go back and forth
between these two shapes and even add more shapes to see them
brickified.
Save your work.

Add Color and Switch to a Teapot

SWITCH NODE
The Switch node is a great tool for providing options within a node
network. This node lets you quickly explore multiple options without the
need to wire and unwire different chains.
This node is also very useful when you wrap up your network into a
Digital Asset because you can promote the switch to the asset as either a
menu or a slider for quick access to different options.

81
 PART FOUR
Color the Points using a Texture
Earlier you added a color attribute to the points which affects the coloring of the bricks instances.
Instead of using a single color, you are now going to use a texture map to create a more interesting
look for the bricks. This will involve some special nodes to turn the texture into point colors.

01 Turn on the Display Flag on the testgeometry_rubbertoy.

Select the rubbertoy node and in the Parameter pane turn
Off the Add Shader parameter.
To hide the UV display in the perspective view, go to the Display
Options bar and turn Off the Show UV Texture when UV’s Present
button.
You will bring back the color on the bricks using a different method
that involves pulling the color from a Texture Map on disk.

02 From the Asset menu, choose Edit Asset Properties >

Rubber Toy. In the Properties window, click on the Extra
Files tab and select toylowres.jpg.
Click the Save as File button and save it into the tex folder. The
texture was stored in the digital asset so that it could be shared
along with the asset. You will use the texture on disk to add color to
the bricks.

03 In the Network view, press tab and start typing

Attribute... Select the Attribute VOP node and place it
into the network beside the pointsfromvolume node.
Feed the output of the matchsize node into the first input of the
attributevop node. Set the Display Flag on this new node.
In the Parameter pane, set Run Over to vertices.

04
HOUDINI FUNDAMENTALS

Double-click on the attributevop node to dive down and

use the tab key to add a Texture VOP. Feed it into the Cd
input of the geometryvopoutput node.
Add a UV coordinate node and feed it into the UV input on the
texture node.

82
 Gear
05 Select the texture node. Click on the Gear icon on
the far right of the Texture Map parameter and from the
menu, select Promote Parameter. This will add this parameter to
the upper level of this node.
Click on the little knob that appears next to map. In the parameter
pane, change the Label to Texture Map.

06 Press u to go up one level. You can see the Texture Map

parameter. Leave it set to the default Mandril.pic texture
for now. You will add the toylowres texture map at the end.
Add an Attribute Promote to the end of this chain. Set Original
Name to Cd and Original Class to vertex. Leave New class set to
Point.

07 Add an Attribute Transfer node. Wire the

pointsfromvolume into the first input and attribpromote
into the second. In the Attributes tab click on arrow on the right
side of the Points field and choose Cd.
Add a switch node after the color node and wire the attribtransfer
node into the switch. Rename this node texture_switch.
Set the Switch to 1. Set the Display Flag on the copytopoints node.
Now you can see the colors from the texture map being transferred
to the copied points but the bricks are rotated. Turn off Transform
Using Target Point Orientation to straighten up the bricks.

08 Set the Display Flag on the material node.

Select the attribvop node and click on the File button
on the far right of the Texture Map parameter and navigate to the
toylowres.jpg texture and click Accept. You can see that the texture
map colors are now being assigned to the vertices.
Save your work.

COLOR THE POINTS USING A TEXTURE

ATTRIBUTE TRANSFER
ld
When you want to get attributes from one piece of geometry to another, es
ho
hr
you can use Attribute Transfer which uses a Distance Threshold along nc
eT
sta
with other parameters to get the attributes copied over. Di

In the case of the rubber toy, you are transferring the Cd attribute from
the points on the geometry to the point cloud that is used to copy the
bricks. You could also copy attributes such as UVs or Capture Weights.

83
 PART FIVE
Create a Brickify Digital Asset
Now that the brickify recipe is working and the nodes are wired together properly, you are going to wrap up some of the nodes
to create a single Houdini Digital Asset [HDA] node. Now you can share the network with parameters from inside the asset
promoted to the top level to create an interface that can generate unique results each time the asset is used.

01 In the Network view, drag the platonic node off to the

side. Select all the other nodes in the network and from
the Assets menu, select New Digital Asset From Selection.... This
will collapse them into a single node.
Set the Operator Name to brickify which in turn changes the
Operator Label to Brickify. Click on the button on the far right
of Save to Library. Select $HIP in the Location sidebar and then
double click on the HDA directory. Press Accept then click Accept
again in the Create New Digital Asset dialog.

02 This brings up the Type Properties window and make

sure the Basic tab is visible. Set Minimum Input to 0
so that you don’t “require” an input for the asset to work. The
Maximum Inputs parameter is set to 1 which defines how many
inputs nodes you will allow.
This is the input that is currently connected to the platonic node.
When you use this digital asset later, you will use this input to point
it to a different shape.
Press Apply. DON’T press Accept because it will close the window.

03 In the Network view, rename the new asset node

brickify_asset and double-click to dive into it. Click on the
switch node that is switching between the testgeometry_rubbertoy
and the Subnetwork Input node. The input node is bringing in the
platonic teapot shape from the level above.
In the Type Properties window, click on the Parameters tab.
Click on the Select Input parameter name and LMB-drag it to the
Existing Parameters list in the Type Properties window. Drop it on
the root to add it to the UI. Click Apply which adds the parameter
but doesn’t close the Type Properties window.
HOUDINI FUNDAMENTALS

WHAT IS AN .HDA FILE?

When you saved your asset, it creates a .hda file on disk. HDA stands for
Houdini Digital Asset and the asset definition is stored in this file then ref-
erenced into your scene. This file contains information about the nodes, the
promoted parameters, UI elements and more. This file can be referenced by
multiple people into different shots for a shared experience.
The assets being referenced into your scene can be managed by going to
the Assets menu, selecting Asset Manager... and opening Current HIP File .

84
RMB-Click
04 RMB-click on brickify_asset in the Network View’s path
bar and choose Parameter and Channels > Parameters.
Now you have a floating Parameter window with the two brickify
asset parameters. These are the parameters that will be available to
anyone who uses this asset. Let’s add some more.
The brickify node has a new parameter. Change its value from 0
to 1 and back to see how it affects your scene. The problem is the
name isn’t very appropriate and a menu would work much better
than a slider in this case, so you are going to refine the UI using the
Type Properties.

05 In the parameter list, click on the Select Input parameter.

To the left are options for refining how people see it.
Change its Name to shape and its Label to Shape.
Now click on the Menu tab and turn on Use Menu. Now under
menu items, type 0 under Token and Rubber Toy under Label then
press enter. Next type 1 under Token and Custom Shape under
Label. Press Apply.
Now in the floating parameter pane, you see a Shape parameter
with a menu. Try it out to see how it works.

06 Select the second switch node that sits just under the
color and attributetransfer nodes and promote the Select
Input parameter to the parameter list. Change its Name to look and
its Label to Look.
Now click on the Menu tab and turn on Use Menu. Now under
menu items, type 0 under Token and Color under Label then press
enter. Next type 1 under Token and Texture Map under Label. Press
Apply.

07 Now select the color node in the network and promote

the Color parameter to the parameter list. This brings
over the color widget and the color wheel as well as the RGB fields.
Press Apply in the Type Properties window to see what this
parameter looks like in the brickify_asset parameter interface.
Note: If you press Accept by mistake the new parameter is saved
to the asset but you will need to use the Asset menu and choose
Open Edit Asset Properties... > brickify to reopen it.
Create a Brickify Digital Asset

08 Now select the attributevop node and promote the

Texture Map parameter to the parameter list.
In the Parameter description section, click on the Channels tab and
change the default value to Mandril.pic. This is a default texture
map that doesn’t require a path to load and is a more reliable
default. Later you will reload the toylowres.jpg texture map.
Press Apply in the Type Properties window and you will see this
parameter in the brickify_asset parameter interface and the Mandril
texture map is now coloring the bricks.

85
 09 To make it clear which parameters are associated with
which shape, you can disable and enable them based on
the menu choice. Click on the Color parameter and in the Disable
When field, enter { look != 0 }.
This tells this parameter to disable whenever Color has not been
chosen in the Look menu. Next, click on Texture Map and in the
Disable When field enter { look != 1 }.
Press Apply and test the results using the Shape menu. You can see
the parameters disabling when they are not needed. You could also
hide them with the Hide When option but disabling is fine for now.

10 The brickify_asset texture map parameter now defaults to

Mandril.pic which is a more versatile default because it
is a texture map that will always be available. To go back to the toy
map, you would need to click on the file selector next to Texture
Map and again click on $HIP then dive into the tex directory and
select the toylowres.jpg file.
The asset is now using the Mandril.pic because it is the default.

11 Now press Accept to save the changes to your asset and

close the Type Properties panel.
In the Network view, press u to go back up one level. With the
brickify_asset node selected, choose Assets > Lock Asset > Brickify
from the main menu. Press Save Changes if prompted.
If you double-click on the brickify_asset node to dive down, you can
see that the network is greyed out and these nodes are protected.
You can only manipulate this asset using its parameters. You will
have to unlock the asset to make changes to its inner workings.

12 Go to the Object level. Now Save your scene file to

preserve the work you have done so far.
You now have the scene file and the .hda file which is being
referenced into your scene to create the asset. You can use this
library to create other instances of the asset in this scene or to add
the asset to another scene.
In the next section, you will test out your asset and find out if it is
working properly. It is always good to have one working asset and
one for testing to make sure it is doing what you want it to do.
HOUDINI FUNDAMENTALS

LOCKING AND UNLOCKING ASSETS

You can lock and unlock selected assets using the Assets menu. When
the asset is locked, it references the HDA file to determine how the asset
behaves. If the asset is unlocked the active definition is in your scene file.
When you lock it you will be prompted to save if there are changes.
If you RMB-click on the asset node you can Allow Editing of Contents
to unlock the asset or Match Current Definition to lock the asset but be
careful because there is no prompt to save and changes might be lost.

86
PART SIX
Test the Digital Asset
A Digital Asset can be instantiated more than once in a single scene file. You are going to use this asset on a different piece of
geometry to test how it works. It is always good to have a test version available so that changes set to the first asset can be
quickly verified. The asset can also be used in other scene files once you have it working properly.

01 Use the tab key to get the Squab Test Geometry. Press
Enter to place it at the origin then use the handle to
move it to the side away from the rubber toy.
Double-click on the new object node to dive down to the geometry
level. Select the node and set Scale to 3 and Translate Y to 1.5. This
will make the Squab a bit bigger than the rubber toy.

02 RMB-click on the output of the test geometry node and

start typing brickify... then select the brickify asset from
the menu. This places the asset into this new network.
Set its display flag and you will see another Rubber Toy colored
Red. This is because these are the defaults for this asset.

03 On the brickify asset node, set the Shape parameter to

Custom Shape and how the squab is being brickified. The
new shape is running through the node network inside the asset
to create a unique result. The shape will be lifted above the ground
some more because of the Match Size node inside the asset.
This is how digital assets become tools that you can put into your
pipeline to package up multiple actions into a single node. This is
an approach that speeds up your workflow and helps you achieve
more consistent results.

04 Select the testgeometry_squab node. From the Asset

menu, choose Edit Asset Properties > Squab. In the
Test the Digital Asset

Properties window, click on the Extra Files tab and select squab_
diffuse.jpg. Click the Save as File button and save it into the tex
folder. The texture was stored in the digital asset so that it could be
shared along with the asset.
Now use this texture on disk to add color to the bricks using the
brickify node’s Texture Map parameter.
When you are finished, go to the object level and name this object
squab and the other one rubbertoy. Save your work.

87
 PART SEVEN
Animating the Bricks
It is possible to continue adding features to the asset and you will create an automating build-up animation for the bricks. This
will involve adding more nodes to our network to make sure the asset has this new functionality. Once the results are saved
into the .hda file, the features will be available for use by anyone using this asset in their work.

01 Hide the Squab object and dive into the rubbertoy object.
Select brickify_asset and choose Assets > Unlock Asset
> Brickify. Double-click on brickify_asset node then RMB-click on
the output of texture_switch and select Group by Range. Place the
node and set its Display flag then set the following parameters:
ƒ Group Name to hide_points
ƒ Group Type to Points
ƒ Range Type to Start and Length
ƒ Length to ($F-1)*20
ƒ Under Range Filter, leave Select to 1 and Of to 1

02 RMB-click on the output of the grouprange node and

select Polygon > Blast. Place this node down then using
the arrow next to Group, select the hide_points group. Now turn on
Delete Non Selected to delete points outside the group. Set the
Display flag on the blast node.
Press play to watch as the points grow with every frame. Now set
the Display flag back to the material node at the end of the chain
and watch the bricks grow over time.

03 Right now the bricks are coming in from one side instead
of from the ground. This is because the points are
appearing based on their point numbers. To control this, you need
to reorder the points to create the look you want.
RMB-click on the output of the texture_switch node and type Sort...
then select the Sort tool. Place this node down and change Point
Sort to Along Vector. With this set to 0, 1, 0, the points start at the
bottom and go up.
Playback to see this result. Test out different vectors to see how it
affects the animation.

04 To let you choose whether you want to animate the

HOUDINI FUNDAMENTALS

brickify effect, you can add another switch node. In the

Network view, Press tab and start typing Switch... Choose Switch
and place the node down. Rename it animation_switch.
Click on the output of the texture_switch node and feed it into the
input of the switch node. Repeat for the blast node. This makes the
original shape the first option and the animated effect the second
option. Changing Select Input to 1 will reveal the animated bricks
but for now keep it at 0.

88
 05 From the Assets menu, select Edit Asset Properties >
Brickify. Go to the Parameters tab and drag a separator
from the Create Parameters section to the bottom of the list.
Next, drag the animation_switch node’s Select Input from the
Parameter pane to just under the new separator. Set its Name
to animate_bricks and its Label to Animate Bricks. Next, change its
Type to Toggle which limits to you an on[0]/off[1] setting.
In the Parameter Description section, click on the Channels tab
and set the default value to 0 [off].
Click Apply to save changes.

06 From the Create Parameters section in Type Properties,

drag an Integer parameter under the animate_bricks
parameter. Set its Name to build_speed and its Label to Build Speed.
Turn on the Range option then set the first value to 1 and the
second value to 20. Click on the lock next to 1 to make sure the
number never gets smaller than 1.
In the Parameter Description section, click on the Channels tab
and set the default value to 1.
Press Accept to save and close the window.

07 This parameter isn’t attached to anything yet but you

will now use it to drive the grouprange node’s Length
expression. Select the new grouprange node and change the Length
expression to: ($F-1)*ch(“../build_speed”)
At the end of the network there is an output node. This will ensure
that you get the right output for your asset even if the Display flag
is on another node in the network.
With the brickify_asset node selected, choose Assets > Save Asset >
Brickify from the main menu. This lets you save this expression to
the .hda file without re-opening the Type Properties window.

08 With the brickify_asset node selected, choose Assets >

Lock Asset > Brickify from the main menu. You now
have the brickify effect wrapped up into a custom tool which can
be used on different shots by different artists.
Go to the Squab network where the new features are available on
the brickify node. Turn on the Animate Bricks toggle and set the
Build Speed which might need to be set to around 30 because of
the large number of bricks in this shape.
Playback to see the results.

CONCLUSION
Animating the Bricks

You have now created a shareable tool which was built without writing scripts
using a node-based workflow which is easily accessible to artists. By saving the HDA to
disk, it becomes an asset on disk that can be referencing into multiple shots.
Houdini Digital Assets offer a powerful way for artists to share these kinds of tools in
support of a studio-level production. These procedural assets make it easy to
automate repetitive tasks and to stay focused on the creative needs of your project.

89
 PART EIGHT
Loading HDAs into other Applications
Once you have a Houdini Digital Asset [HDA] saved on disk, it is possible to load that asset into a host application using the
Houdini Engine plug-ins. These let you share assets with colleagues who can load them directly into 3D apps such as Autodesk
Maya or 3DS Max or into game editors such as Unity or Unreal Engine.

01 To use the Houdini Engine in a host application, you need

to first install the plug-ins using the Houdini installer or
Launcher. This will make the plug-ins available but you may need to
take further steps to make the Houdini Engine available within your
session.
Visit the following page for details: SideFX.com/engine
Click on the Engine Plug-ins tab and then click on the desired
plug-in for more information. Once installed, you will find a Houdini
Engine menu in the host application. You can use this menu to load
assets.

02 Once you have the plug-in installed, you can load the
asset using the Houdini Engine menu. This will bring the
brickify asset into the viewport while the asset parameters become
available for manipulation.
You can also turn on animation for Maya or 3ds Max and play it the
sequence using the timeline.

03 In Unreal, you can use the Import button to bring

digital assets into your scene. You can also set Shape to
Custom Shape and connect the asset to geometry within the host
application and the brickification will be applied to that object.
HOUDINI FUNDAMENTALS

WHAT HAPPENED TO THE BRICK COLORS?

You may notice that none of the brickified shapes are showing the brick colors within the various host applica-
tions. This is because the plug-ins don’t always process information the same way as Houdini would. The point
colors are still part of the asset but the host application is not receiving this information.
And while the animation works in Maya and 3ds Max, it would not make sense in Unity and Unreal Engine be-
cause Houdini assets can not become part of the runtime experience of a game therefore the built-in animation
would be ignored. It is important to tailor your assets for the capabilities of the host application.

90
HOUDINI FOUNDATIONS

SMASHING WINE GLASS

In this lesson, you will smash a wine glass then slow down time to hold onto the big splash. This effect
involves both an RBD simulation for the shattered glass and a fluid simulation for the wine. You will
learn how to set up the dynamic network and output the simulations. Visual effect shots often involve a
combination of different kinds of dynamic solvers and Houdini’s dynamic network is designed to achieve a
unified result with these different solvers.
You will also use a retime node to slow down the simulation when it is at its most explosive then reverse
time to return to the starting point. You will then move the simulations into Solaris/LOPS, set up lights and
a camera, then render the shot using the Karma renderer.

LESSON GOAL
To simulate a bullet smashing through a wine glass causing the glass
to break and the liquid to splash.

WHAT YOU WILL LEARN

ƒ How to model the Wine Glass, Bullet and Liquid geometry
ƒ How to pre-fracture the glass geometry using Booleans
ƒ How to run a Rigid Body Simulation of the bullet smashing the glass
ƒ How to run a FLIP Fluid Simulation of the liquid splashing
ƒ How to retime the simulation to slow it down then go back in time
ƒ How to export the results as USD for use in Solaris/LOPS
ƒ How to set up lights and materials in Solaris/LOPS
ƒ How to render the final shot with Karma

LESSON
COMPATIBILITY
Written for the features in Houdini 19.5+

The steps in this lesson can be

completed using the following
Houdini Products:

Houdini Core ✘
Houdini FX ✔
Houdini Indie ✔
Houdini Apprentice ✔
Houdini Education ✔

Document Version 2 | Sept 2022

© SideFX Software
91
 PART ONE PROJECT FILES
Model the Wine Glass Go to the fluids tutorial page on
SideFX.com, where you got this document
To start, you will draw a polygonal curve then revolve it to create the wine and download the fluids_lesson_start direc-
glass. Creasing will be used to sharpen some edges and then you will subdi- tory. Rename it fluids_lesson then put it into
vide to create denser geometry for fracturing. You will then extract geometry the Houdini Projects directory.
from the wine glass to create a shape that you will use to simulate the fluid.

01 Select File > Set Project. Find the fluids_lesson directory

(see instructions above) and press Accept. This makes
the project directory and its sub folders the focus for all the files
associated with this shot.
Select File > Save As... You should be looking into the new
fluids_lesson directory. Set the file name to wineglass_01.hip and
click Accept to save. Now you will be able to access the reference
images in the Tex folder.

02 In the Scene view, Press v to bring up a radial menu and

choose Viewport Layout > Four Views Move your cursor
over the Front panel and press spacebar-b to expand it.
Hit D with the mouse in the viewport. Click on the Background tab
and on the Front tab , use the file picker to navigate to $HIP then
tex>wineglass_profile.jpg. Set the following:
ƒ Make sure Auto-Place is turned OFF
ƒ Image Offset to 0, 3
ƒ Image Scale to 5, 5
Now if you dolly or pan the view, the background moves with it.

03 On the Polygon shelf, click on the Curve Polygon tool.

This creates a Curve node with the Primitive Type set
to Polygon. Press x and choose Grid to turn on grid snapping then
click on point A and then to the second point which also sits on a
grid point. Turn Grid snapping to OFF then keep tracing the image.
Grid Snap
last four points Turn Grid snapping back ON for the final four points at the base of
the glass to make sure they are aligned. When you finish at point B,
Grid snap press Enter to complete the curve. Turn Grid snapping to OFF.
first point

Grid Snap
You can then click on the Edit Mode button in the Operation
second point Controls bar to move any points that missed their mark.

04 Press Spacebar-B to go back to a four view layout then

HOUDINI FOUNDATIONS

mouse over the perspective view and press spacebar-b

again to expand it. Now you can see the curve in 3D.
Press s to get the Select tool and press n to select all the primitives
on the curve. Press c and choose Model > Curves > Revolve. This
Shift-select will turn it into the wine glass model.
these four
edges Press 3 to go to edge selection then double click on the top edge
of the glass then press Shift and double click on the second top
edge and the two edges of the base. Press tab > Crease then set
Crease to 0.75.

92
05 Press 4 to switch to primitive selection then press n to
select all. Now press tab > Subdivide. Set Depth to 2.
This will subdivide the model with the creased edges set up to be
sharper than the other areas of the model. A higher Crease weight
could make these edges completely sharp but for now a softer look
works better.
In the Network editor, add a null node to the end of the chain and
set its display flag. Rename this node to GLASS_OUT.
Go back up to the object level and rename the object to wine_glass.

06 With the wine_glass node selected, press n to select all

the primitives then go to the Modify shelf and click on
the Extract tool. This will create an objectmerge of the wine glass
geometry and place it into a new object.
From the Front view, select the faces at the top of the wine glass
and then press delete. This adds a Blast node into the network
that removes the faces.
Note: You can still see a ghosted version of the original wine
glass because the Scene View is set to Ghost Other Objects. This
setting is useful to add context to your work.

07 Now go back to a Perspective view and double-click on

the base of the wine glass and press delete. This adds a
second blast node. Now you are left with interior faces where you
want the liquid to be. The faces of the wine geometry will appear
dark on the outside. This means that they are the back sides of the
primitives.
Press n to select all the primitives. Press tab > reverse to reverse
the normals so that they are facing out. The darker side of each
primitive will now be on the inside of the shape.

08 Press 3 to change to edge selection then double-click on

the edge of the shape to select the open edge. Press tab
and start typing polyfill. With Polyfill highlighted, press Enter.
Set the following:
ƒ Fill mode to Quadrilateral Grid
ƒ Tangent Strength to 0.
This creates a closed shape that will become the source for the FLIP
Fluid later in the lesson.

09 Press 4 to change to primitive selection then press n

Model the Wine Glass

select all of the primitives. Press c and choose Model >

Polygons > PolyExtrude. Set Distance to 0.01 to create an overlap
with the wineglass to help the fluid render properly.
In the Network editor, add a null node to the end of the chain and
set its display flag. Rename this node to FLUID_OUT.
Go up to the object level, rename this object to wine.

93
 PART TWO
Model the Bullet
To create the bullet geometry, you will start with a primitive sphere and slice it in half. Next you will polyextrude the open end
then use polyfill to close the shape using quad topology. You will then subdivide to define the final shape. This object will be
moving very quickly therefore lots of detail isn’t required.

01 Go to the object level and in the Network view, turn off

the Display Flags for the wine and wine_glass objects. In
the Scene View, press c and choose Create > Geometry > Sphere.
Press Enter to place it at the origin then dive inside and set the
following:
ƒ Orientation to X axis
ƒ Radius to 0.2, 0.125, 0.125
ƒ Columns to 12
Press tab > clip in the Scene view then press n to select all the
primitives and press Enter and set the Direction to 1, 0, 0.

02 Click on the Select tool then press 3 to turn on edge

selection and double click on the open end of the sphere.
Press c and choose Model > Polygons > PolyExtrude.
In the Extrusion tab, set the following:
ƒ Transform Extruded Front to ON
ƒ Translate Z to 0.04
ƒ Scale to 0.7, 0.7, 0.7
This adds a little extra geometry before you remove the triangles at
the front of the bullet then close the shape.

2. Press A then
03 Tumble around and press s to get the select tool and 4 to
get face/primitive selection. Select one of the triangles at
the tip of the bullet then press and hold the a key and then middle
Polyfill to Close
MMB-click Press Delete
click two triangles over to select all the triangular faces. Press the
Delete key to remove them. This adds a blast node to the network.
In the Network View, press tab > polyfill and place the node after
the blast. Set Fill mode to Quadrilateral Grid and Corner Offset to
1 then turn on its display flag. This will close both ends of the bullet
1. LMB-click with good quad topology.

04 Now add a Subdivide node at the end, set Depth to 2

HOUDINI FOUNDATIONS

and then set its display flag.

In the Network editor, add a null node to the end of the chain and
set its display flag. Rename this node to BULLET_OUT.
Go back to the object level and rename this object bullet. Turn on
the display flag for the wine_glass. Translate the bullet to about -20
in X and 5 in Y. You may want to go back to the Front orthographic
view to make sure it is aimed at the desired impact point on the
wine glass.
Save your work.

94
PART THREE
Fracture the Wine Glass
To define the cracks in the wine glass, you will create natural looking lines using the draw curve tool then extrude them into
sheets of geometry. You will then agitate the surfaces using the mountain tool. This chaotic looking shape can then be merged
into the wine glass object where you will set up a boolean operation that uses the sheets to shatter the glass.

01 In the Scene View, go to the object level then press c and

choose Create > Geometry > Grid. Press Enter to place it
at the origin then i to dive inside and set the following:
ƒ Orientation to YZ Plane
ƒ Center Y to 4
ƒ Size X to 5
ƒ Size Y to 8
ƒ Rows to 16
This will create a drawing surface for the Draw Curve tool that
matches the shape of your wine glass.

02 Go to a Right view and press v > Shading > Wireframe

to go into wireframe mode. You can see the bullet sitting
behind the grid and the wine glass. Press n to select the whole grid
then go to the Create shelf and click on the Draw Curve tool.
Draw curves over the wine glass that converge where the bullet
hits the glass. Also add some curves across the stem of the glass
since that will shatter as well. If at any point you draw a curve you
don’t like, press Ctrl-Z to undo it.

03 Go to the Select tool then Press n to Select all the

curves then press c and choose Model > Polygons >
PolyExtrude. Set Divisions to 4.
In the Extrusion tab, set the following:
ƒ Transform Extruded Front to ON
ƒ Transform Space to Global
ƒ Translate X to -4.5
Next, click away then select all the geometry and press tab >
Transform. Set Translate X to 2.25 to center this geometry around
the origin.

04
Fracture the Wine Glass

Press s to bring up select tool, then press 4 to go to

primitive selection. Double click on one of the sheets to
select the whole thing. Press tab > Duplicate to create a copy. Use
the Rotate [r] handle to reorient the duplicated sheet to be almost
orthogonal to the others, making sure it cuts through the cup but
avoids the stem. This will break the cup in the other direction to
create more realistic slivers of glass.
Repeat with another sheet to create another surface for cutting at a
different angle.

95
 BOOLEAN SHATTER
The Boolean node is often used to create traditional booleans such as
Union, Intersect and Subtract. While these work well with closed shapes,
you can use the Shatter option to slice up the geometry with sheets.
Houdini has a Voronoi Shatter tool that can also be used but it will not give
you the jagged look you need for broken glass. There is also an RBD Material
RBD Material
Fracture node that can create glass-like fractures. These work best with flat Fracture
Voronoi
surfaces which is why it wasn’t used in this lesson for the wine glass.
Boolean Shatter

05 Select all the geometry and then press tab > Mountain to
add some noise to the points on the different sheets. Set
Amplitude to 0.75. This will create more interesting cracks in the
glass.
In the Network view, add a Null node to the end of the chain and
rename it FRACTURE_OUT. Set the display flag on this node.
Go to the object level and name this node to fracture_geo then turn
OFF its display flag to hide it.

06 Next dive into the wine_glass object. In the Network

view, press tab > object merge and place the node
down. Click on the node selector next to Object 1 and navigate to
fracture_geo > FRACTURE_OUT and select that node. Make sure
that Transform is set to Into Specified Object .
This brings the agitated sheets into the wine glass geometry
network where you can use it to boolean shatter the glass.

07 In the Network view, press tab > boolean and click to

place down the new node. Wire the subdivide node
(not GLASS_OUT) into the first input and the object_merge into the
second. Set its Display flag then set the following:
ƒ Set B: Treat As to Surface
ƒ Operation to Shatter (Pieces of A)
To visualize the cracks, add an exploded view node at the end of
the chain. If you want to change how things work then you can go
back to the fracture_geo object and edit the sheets. Those edits will
update procedurally.

08 In the Network view, add a Null node that bypasses the

HOUDINI FOUNDATIONS

exploded_view node and rename it GLASSFRACTURE_

OUT. Set the display flag on this node.
You now have two output nodes for this network. GLASS_OUT
gives you the wine glass in its original form and GLASSFRACTURE_
OUT gives you the shattered glass. You will use them both down
the line to complete the shot.
Save your work.

96
PART FOUR
Set up the RBD Simulation
You are now going to create a rigid body simulation using shelf tools. This will add a DOP (Dynamic Operator) network that
brings together geometry, forces and a solver node. In the wine glass geometry network, nodes will be added to prepare the
geometry for simulation. You will use convex proxies so the Bullet RBD solver can handle the odd-shaped pieces of glass.

01 Go up the object level, select the wine_glass object then

from the Rigid Bodies shelf, click on the RBD Convex
Proxy tool. This will set up the initial dynamics network for you. It is
called AutoDopNetwork. Press v > Shading > Smooth Wire Shaded.
This tool is used to break down the parts of the wine glass into
convex shapes that can be used to create complex collisions.
These will appear rougher than the source geometry but after the
simulation, you can go back to visualizing the cleaner topology.

02 Go to the Collisions tab and click on the Ground Plane

shelf tool. This will create an infinite ground plane for
your geometry to collide with.
You can turn Off the Display flag on the groundplane_object to hide
it from the scene view. It will still function as a colliding surface in
the simulation.

03 Dive into the AutoDopNetwork node. Select the wine_

glass node and in the Physical tab set Density to 2000,
which is approximately the density of glass.
Press Play in the playbar to see what happens. The glass falls and
hits the ground. Right now gravity is the only force acting on the
pieces.
You could set up a glue network to hold the pieces together until
impact but the bullet will be so fast that connecting the parts is not
necessary. SET UP THE RBD SIMULATION

CONVEX DECOMPOSITION
For RBD simulations, Houdini uses the Bullet solver which prefers
convex shapes in order to maintain fast simulation speeds. Convex 2
Decomposition lets you take shapes that are concave and break 1
4
them down into convex shapes that are connected together. These 3

are then simulated as one composite piece by the Bullet solver. 5

Since the wine glass fragments come in a variety of shapes, convex

Original Geometry Convex Decomposition
decomposition ensures accurate collisions for all pieces.

97
 04 Go up to object level and dive into the wine_glass object.
A number of nodes have been added to create the proxy
geometry and this chain ends with the dopimport node which is
currently displayed.
On the convexdecomposition node, you can change Max Concavity
to 0.05 to adjust your collision geometry.
Press Play to watch it simulate.

05 Go up to object level and select the bullet object and

then go to the Rigid Bodies shelf and click on the RBD
Objects tool. This will create a packed rbd object from the bullet
and add it to the dynamics network.
Navigate into AutoDopNetwork and select the bullet node.
ƒ In the Initial state tab, set Velocity to 400, 0, 0.
ƒ In the Physical tab set Density to 20000
This is the density of lead and should work well for the bullet.

06 Before you simulate, you want to make sure that the base
of the glass stays on the ground. Navigate into the wine_
glass object and in the display bar, turn on primitive numbers. You
can see that in this case the base packed primitive is 171 - yours
will probably be different.
Add an Attribute Create node between the create_packed_primitive
and proxy_geo nodes. Set the following:
ƒ Group to !171 (or whatever number your base is)
ƒ Name to active
ƒ Value to 1, 0, 0, 0

07 Next, set the display flag on the transform_hires node and

open up Global Animation Options and set End to 50.
Now play back the sim. You can now see the source geometry
being animated to match the proxies but the collision is as dramatic
as it could be.

08 The Bullet sim defaults to 10 substeps on the solver. This

HOUDINI FOUNDATIONS

isn’t enough to resolve collisions for the speed the of the

bullet. Go up to the object level and on the AutoDopNet node set
Substeps to 5. This will add substeps on top of the solver setting.
This may take longer to simulate but will increase accuracy and
generally make the simulation more active.
Play the simulation again to see that the collisions are looking much
better. Go back and tweak settings if you want too change things.
You can even go back and reposition some of the cracks.
Save your work.

98
PART FIVE
Add Fluids to the Simulation
Now that the bullet is smashing the glass, it is time to convert the wine object into a fluid that will become part of the inte-
grated simulation. This means that the RBD and Fluid simulations take place in the same DOP network and will work as one
system. At first, the fluid will be represented by particles that can then be surfaced to visualize the fluid.

01 Go to frame 1. Select the wine object in the Network

view and from the Particle Fluids shelf, click on the FLIP
Fluid from Object tool.
This turns the fluid into a volume of fluid particles. In AutoDopNet,
select the flipfluidobject node and set Particle Separation to 0.05.
This creates more particles which in turn adds more detail to the
simulation.

02 Select the flipsolver1 node and click on the Particle

Motion tab, under Behavior set:
ƒ Add ID Attribute to ON
Under Reseeding set:
ƒ Reseed Particles to OFF
In Volume Motion, set:
ƒ Velocity Transfer to (APIC) Swirly
Under Surface Tension, set:
ƒ Enable Surface Tension to ON
ƒ Surface Tension to 500

03 Press Play to run the simulation. With extra substeps it

will take a little longer but will give you more accurate
results. Simulate to just past frame 10 and press Escape to stop the
sim. Go to frame 10 to preview the fluid so far.
To see it as a surface, go up to the object level and go into the wine_
fluid object. Set the display flag on the Render null node to see the
surfaced fluid. This will take longer to cook but will show you what
the surface is going to look like at each frame.
Add Fluids to the Simulation

DISPLAY AND RENDER FLAGS

When the wine_fluid geometry network is first created, the
Display flag is on the dopimport node which shows particles
while the Render flag is on the null node which flows from
a particle surface node. This set up is meant to give you fast
performance in the viewport and if you render then you see the
final surface. In this lesson, you will be caching out the surface
therefore these nodes will not be used to render.

99
 PART SIX
Cache and Retime the Simulations
For this shot, you only need to compute 10 frames of the simulation. You will save this to disk then use a retime node to create
a longer shot where the fluid slows down then reverses in time. The retimed fluid particles will then be surfaced and output as
a 50 frame sequence which will define the final shot for rendering. The wine glass and bullet will then be retimed to match.

01 At the object level, delete the wine_fluid_interior node

that the shelf tool created. Dive into the wine_fluid object.
Delete all the nodes except the import_wine, compressedcache node
and the particlefluidsurface nodes.
On compressed_cache, set the following:
ƒ Base Name to wine_fluid
ƒ Base Folder to $HIP/geo/fluid/
ƒ Turn Off the Version check box
ƒ End to 10 (RMB-click > Delete Channels first)
Hit Save to Disk.

02 When it finishes, stay at the geometry level. From the

Visibility menu in the top right of the Scene view, select
Hide Other Objects.
Set Load from Disk to On and Scrub through the geo sequence.
It will only play for 10 frames. You are now going to retime the
sequence to stretch it out over 50 frames and slow down the
effect.

03 Add a Retime node after the compressed_cache node and

set its Display flag. Set the following:
ƒ Evaluation Mode to By Frame.
ƒ Turn ON the Scale Velocities option
RMB-click on the Frame field and choose Delete channels.
ƒ At frame 1: set Frame to 1 and Alt-click to keyframe
ƒ At frame 5: set Frame to 1 and keyframe
ƒ At frame 10: set Frame to 7 and keyframe
ƒ At frame 40: set Frame to 10 and keyframe
ƒ At frame 45: set Frame to 1 and keyframe

04 Click on the Animation Editor pane to see the animation

HOUDINI FOUNDATIONS

curve you just created. Press h over the graph to home

the view and then RMB-click-drag to zoom out a bit.
Click and drag on the curve handles to get a shape similar to what
you see here. If you need to break tangents on a curve, select the
key and press T to untie the tangents. Then select and drag on
each end individually. The goal here is to have the liquid splash out
quickly then slow down until is freezes for a short time then speed
out into a fast reconstruction back to its original shape.

100
RETIME
The most interesting part of the smashing wine glass is the first 10 seconds. To
emphasize this part of the simulation, you are going to save out the fluid particles
for the 10 frames then uses the retime node to stretch out the sequence in a sort
of “bullet-time” effect that snaps back to the original wine glass by reversing time.
Once you have this set up for the fluid, you can surface the points and save out a
longer sequence. The same retime node can then be copied and pasted to be used
on the smashing glass and the bullet.

05 Make sure that the retime node is wired into the

particlefluidsurface node then set it’s display flag. This
will give you your final fluid based on the retiming. Select the
particlefluidsurface node and keep:
ƒ Method set to Average Position
ƒ Set Union Compressed Fluid Surface to Off.
In the Filtering tab turn:
ƒ Dilate to ON and set it to 2.
ƒ Smooth to ON and set it to Laplacian Flow

06 Add a USD Export node to the end of the chain. Rename

it wine_surface. Set the following
ƒ Valid Frame Range to Render Frame Range
ƒ Output File to $HIP/usd/wine_surface.usd
Hit Save to Disk.

07 Select the retime node and press Ctrl-c to copy it. You will
paste this in another network to retime the shattering
wine glass. This will ensure that the keyframes match in both of the
networks.
Save your work.

Cache and Retime the Simulations

08 Go to frame 1 and navigate into the wine_glass network.

Beneath the transform_hires node, wire in a File Cache
node and set the following:
ƒ Base Name to glass_pieces
ƒ Base Folder to $HIP/geo/
ƒ Turn Off the Version check box
ƒ End to 10 (RMB-click > Delete Channels first)
Hit Save to Disk. Set Load from Disk to On and Scrub through
frames 1 to 10 to make sure that it looks correct.

101
 USD and SOLARIS
To support the look dev stage of this project, you are caching out the fluid, the
wine glass and the bullet to USD. By doing this, you can focus on rendering
without worrying about simulations being recalculated. Here you will display the
caches in the same scene file as the simulations but another option would be to
start a new scene file and import the caches into that file. This would let you focus
on lighting and rendering your shot but would make it harder to go back and tweak
the sim.

09 Press Ctrl-v to paste in the retime node and place it after

the filecache node and set its Display flag. Now you have
the same timing on the glass that you have on the fluid. You can
scrub through the timeline to watch the timing of the pieces.
After the retime node, add an Attribute Delete. From the arrow
next to Primitive Attributes, select name to remove this attribute
from the geometry.

10 Add a switch node into the network. Feed the GLASS_

OUT into it first and attribdelete node into it second.
Set Select Input to $F>5 && $F<45. Scrub in the timeline to
see how this works. This expression makes the switch from the
unbroken glass to the broken glass at frame 5 and then back at
frame 45.
You are putting these shapes together so that the glass is unbroken
before and after impact in the wineglass USD file.

11 Add a USD Export node to the end of the chain. Set the
following
ƒ Valid Frame Range to Render Frame Range
ƒ Output File to $HIP/usd/wineglass.usd
Hit Save to Disk.

12 Go to frame 1 and select the bullet object. From the

HOUDINI FOUNDATIONS

Modify menu, select Extract. This gives us the world

space position of the geometry. Dive into extract_object and, Paste
the retime node beneath the object_merge node then connect them.
Add a USD Export node to the end of the chain. Rename it bullet.
Set the following
ƒ Valid Frame Range to Render Frame Range
ƒ Output File to $HIP/usd/bullet.usd
Hit Save to Disk.

102
PART SEVEN
Set up and Render the Shot
To render the shot, you will reference the USD files into the Solaris Stage then add a backdrop. Solaris is the context of Houdini
that uses LOP nodes to set up the USD Scene Graph. Next, you will add and position a camera and then an environment light.
The Karma renderer will then be invoked in the viewport to create a preview render of the shot.

01 Change the desktop to Solaris. Choose Stage from the

path bar.
In the Network View press tab > Reference then click to add a
Reference node. Next to Reference File, click on the File Chooser
and find the wineglass.usd file. Rename the node to wineglass. Set
the Primitive Path to /geo/$OS - this will use the node name and
place it into a group called geo.
In the Scene View, use your view tools such as spacebar-h for
homing the view to get a better look at the wine glass.

02 Alt-drag on this node to make a copy and again to make a

second copy. For the first copy, click on the File Chooser
and find the wine_surface.usd file. Rename the node to wine.
Repeat to make another copy, find the bullet.usd file. Rename the
node to bullet.
Add a Merge node to the network and wire all three reference
nodes into it. Set its display flag then scrub to see the results.
In the Scene Graph, you can see a geo entry and underneath that
you will find the three referenced USD files.

03 In the Network view, press tab and type out Grid. Click
to place down the node, rename it backdrop and wire it
into the merge node. Set Import Path Prefix to /geo/$OS. Double-
click on the backdrop node to dive down to the geometry level.
Select the Grid node and set the Size to 200, 200 and Rows and
Columns to 20. RMB-click on the grid node’s output and type
Bend. Click to place a bend node and set its Display Flag then set:
Bend to 75, Capture Origin to -40, 0, 0, Capture Direction to -1, 0,
0, and Capture Length to 20. RMB-click on the grid node’s output
and type Subdivide. Set its Display Flag then set Depth to 2.
SET UP AND RENDER THE SHOT

CACHING OUT SIMULATIONS

To support the look dev stage of this project, you cached out the fluid, the wine
glass and the bullet to geometry (USD) sequences. By doing this, you can focus on
rendering without worrying about simulations being recalculated.
This is a typical workflow when working on VFX shots that can consume LOTS OF
HARD DRIVE SPACE. You should be aware of this before sending off huge simula-
tions with lots of particles - make sure you have somewhere to store the various
intermediate stages.

103
 VIEWPORT RENDERING
You are now going to render the sequence using Houdini’s renderer
Karma. At first this will render using settings you can find in the Display
options. Press d in the Scene View to bring it up. You can turn on the
denoiser here, set Pixel Samples and the Image Resolution.
Later when you set up a Karma LOP, there will be render settings on
that node which will be used to create the final output.

04 Use your view tools to look at the wineglass from the

front. From the LOP Lights and Camera shelf, Ctrl-click
on the Camera tool. This adds a camera node into the network and
you are now looking through the camera in the viewport.
Press the Lock Camera/Light to View button so that view changes
can be used to reposition the camera. Now Tumble, Pan and Dolly
in the viewport to reposition the camera so the wineglass is on the
left and the splash moves to the right. Scrub the timeline to make
sure the camera works for the whole sequence.

05 From the LOP Lights and Camera shelf, Ctrl-click on the

Environment Light tool. This adds a domelight node to
the end of the chain.
Select the domelight node and from the Base Properties tab, click
on the File Chooser button next to Texture. Click on the $HFS/
houdini/pic/hdri listing in the sidebar then select the HDRIHaven_
skylit_garage_2k.rat file. Click Accept.
On the Display Options bar. Click on the High Quality Lighting
with Shadows button.

06 From the Persp menu, choose Karma to render with

Karma in the viewport. You can move to different frames
in the timeline and the viewport will update quickly.
Karma is designed to work with USD which is why everything in the
LOP context is converted to the USD scene graph. You can only use
the Karma renderer from this part of Houdini.

07 To get a cleaner image when you render, you can turn

HOUDINI FOUNDATIONS

on Denoiser if you have an Nvidia graphics card and you

have installed the latest drivers. You can turn it on in the Display
Options bar.

Turn ON
Denoiser

104
PART EIGHT
Assign Materials and Render a Sequence
Now you can add materials to the wineglass, wine and bullet. These materials will become part of the USD Scene Graph and
will get assigned to the geometry using a LOP node. You can then use a Karma LOP to prepare your render settings including
the use of the Nvidia Optix Denoiser. After you render, you can load the sequence into Mplay to review the results.

01 In the Network view, press tab > Material Library. Wire it

into the end of the chain then set its Display Flag.
Go to the Material Palette pane. Click on the arrow next to /stage/
materiallibrary to open up this area. Scroll through the material
gallery on the left of the palette and drag two Glass materials into
the materiallibrary working area for the wineglass and wine.
Now find a copper material and drag it into the materiallibrary
working area. You will use this for the bullet.

02 Select the second glass material and rename it to wine.

Set Inside IOR to 1.3443 which is an IOR for wine. Next,
set Reflectivity to 0.2 to reduce how much the surroundings are
reflected.
Set Transmission Color to 0.2, 0, 0 which will create a reddish wine
look. Next set At Distance to 0.3.

03 Go back to the Stage level. After the Material Library

node, add an Assign Materials node.
From the Scene Graph, drag the wineglass to the Primitives field

Assign Materials and Render a Sequence

then click on the arrow next to Material Path and choose the glass
material for this primitive.
Now click the Plus Sign next to the check box twice to add two
new entries. Use the same method to assign the wine material to
the wine primitive and the copper material to the bullet primitive.

SCENE GRAPH
Just like the geometry and lights, the materials you add using LOP
nodes are added to the Scene Graph. When you used the Material
Library LOP, the default setting for the Material Path Prefix was
/materials/ and that is where they are placed in the graph. You
could choose to organize them differently but this is the default.
This material path is the one you used in the Assign Materials LOP
to place the materials onto the geometry.

105
 04 In the Network View, press tab > Karma to add a Karma
Render Settings and USD Render ROP node. Wire them
into the end of the chain. Select the karmarendersettings node and
on the Image Output > Filters tab set Denoiser to nvidia Optix
Denoiser to turn the denoiser back on.
Select the usdrender_rop node. Set Valid Frame Range to Render
Frame Range and set the Output Picture to $HIP/render/
wineglass_$F4.exr. The $F in the name is needed to add frame
numbers to the renderings and the 4 is the padding of the frame
number.

05 The denoiser from the viewport will not affect the output
from this node therefore you much choose it explicitly.
Select the karmarendersettings node and on the Image Output
> Filters tab set Denoiser to nvidia Optix Denoiser to turn the
denoiser back on.
The nVidia Optix Denoiser will match the one used in the viewport.
There is also an Intel OIDN denoiser which is only available when
rendering to disk.
Save your work. Select the usdrender_rop node and click on Render
to Disk.

06 When you finish, choose Render > Mplay > Load Disk
Files and open up the rendered images to review the
final sequence.
Later you can branch off another Karma node to up the resolution
and render settings for your final rendering. You can go back to
Convergence Mode set to Variance, up the sample count and turn
off the denoiser. It is always good to complete test renderings at a
lower resolution first to make sure that everything is working the
way you expect it to.

CONCLUSION
You have now created a complete VFX shot using the Bullet RBD and FLIP
Fluid solvers to smash a wine glass. You used the retime node to slow
down then reverse time so that the wine glass finished in the same posi-
tion as it started then you cached out the results to USD files.
These were then used to set up and light your shot using the Solaris/LOPS
context. Materials were created and assigned to the primitives get the
right look for your shot.
This project shows how you can use Houdini’s dynamic nodes and net-
works to integrate different kinds of effects while using geometry nodes to
set up and output the simulations.
HOUDINI FOUNDATIONS

Now that you have an understanding of the nodes and networks used to
create VFX shots. This will assist you as you dig deeper into Houdini to
achieve your own effects.
Enjoy!

106
HOUDINI FOUNDATIONS

DESTRUCTION FX
One of the things that makes visual effects fun is that you get to blow things up without causing any real
damage. In this lesson, you will light a fuse using particle sparks then explode a cartoon bomb using rigid
body dynamics for the shell of the bomb and Pyro FX for the fire and smoke. This lesson will teach you how
to set up dynamic simulations using a variety of shelf tools and network nodes.
To give you a complete understanding of the shot being developed, you will build all the elements from
scratch, then simulate the effects. This will help you understand how the simulation nodes work within the
wider context of a Houdini scene. In the end, you will render out the shot using the Karma renderer.

ACES | OPENCOLORIO SETUP

For more accurate color display when working with Pyro
FX, you should use the Academy Color Encoding System
(ACES). To use it, bring up the Correction Toolbar from the
viewport (persp) menu in the scene view. From the arrow button on the right, choose OpenColorIO. This
will give you a Display of sRGB and an output of SDR Video - ACES 1.0. This setting only works for your
current session and will need to be turned back on each time you open Houdini.

LESSON GOAL
Model then blow up a bomb using particle sparks, rigid body dynamics and Pyro FX.

WHAT YOU WILL LEARN

ƒ How to model the bomb and animate the fuse
ƒ How to animate a camera to set up the shot
ƒ How to set up a particle soot trail and sparks for the fuse
ƒ How to shatter then explode the bomb geometry
ƒ How to set up a Pyro FX simulation for the explosion
ƒ How to set up materials and textures
ƒ How to render the effects with Karma in the Solaris context

LESSON
COMPATIBILITY
Written for the features in Houdini 19.5+

The steps in this lesson can be

completed using the following
Houdini Products:

Houdini Core ✘
Houdini FX ✔
Houdini Indie ✔
Houdini Apprentice ✔
Houdini Education ✔

Document Version 1.0 | July 2022

© SideFX Software
107
 PART ONE
Model the Bomb
To create the bomb geometry, start with a primitive sphere and modify it to define the opening at
the top. This will involve a few poly extrudes and bevels to define the geometry you will need for
the final shape. Later in the lesson, you will fracture the bomb.

01 Select File > New Project. Change the Project Name

to destruction_lesson and press Accept. This creates a
project directory with sub directories for all the files associated with
this shot.
Select File > Save As... You should be looking into the new
destruction_lesson directory. Set the file name to destruction_01.hip
and click Accept to save.

02 In the viewport, press c to bring up a radial menu. From

this menu, choose Create > Geometry > Sphere. In the
viewport, press Enter to place it at the origin. In the Operation
Control bar at the top, set Radius to 0.3, 0.3, 0.3.
Press s to get the select tool then 3 to invoke edge selection. Press
Spacebar 2 to go to a Top view. Box select the edges at the top and
bottom of the circle and press delete. This dissolves the edges and
leaves two circular polygons in their place Press Spacebar 1 to go
back to a perspective view.

03 Press s to get the select tool then 4 to switch to primitive

(face) selection. Select the circular polygon at the top of
the sphere.
Press c to bring up the radial menu and select Model > Polygons >
Polyextrude. Move the handle up by a Distance of about 0.075. Set
Output Front to Off.

04 Select the circular polygon at the bottom of the

HOUDINI FOUNDATIONS

sphere and press Delete. This will add a blast node

to the network. Press s to go to the Select tool and 3 to change
to edge selection. Double click on the edge of the hole you just
created to select all the edges.
Press tab > Polyfill. In the Parameter pane, set Fill Mode to
Quadrilateral Grid and Smooth to 270. This creates a cleaner
topology for the bottom of the sphere that doesn’t go to a
single point.

108
SURFACE NORMALS
Every primitive has a normal direction where one side is the inside and
one is the outside. When you polyextrude the bomb geometry it will be
inside out at first. This is indicated with the blue color on the faces. You
can then use a Reverse node to redirect the normals.
You can see the normals on the surface using the Display Primitive
Normals button found in the Display Options bar on the right side
of the Scene view pane.

05 Press n to select all Faces and again get the Polyextrude

tool. Extrude to a Distance value of around -0.04. In the
Parameter pane in the Extrusion tab, turn on Output Back.
Press n to select all Faces and again press tab and start typing
Reverse. This node reverses all the polygon normals. Now they are
pointing in the right direction.

06 Press s to go to the Select tool and 3 to change to edge

selection. Double click on the edge at the top of the
bomb and then press Shift and double click to select the inner
circle at the top.
Press c to use the radial menu to go to Model > Polygons >
Polybevel. Set Distance to 0.005. Set the Shape to Round and
Divisions to 3.

07 Press s to go to the Select tool. Double click on the edge

where the circular part of the bomb meets the extruded
section at the top.
Press q to repeat the last tool which was Polybevel and set
Distance to 0.01, Shape to Round and Divisions to 3.

08 In the Network view, press tab > Transform and add it

to the end of the network. Set Translate Y to 0.3 and
Rotate X to around 27 degrees.
Add a Null node. Wire the end of the polybevel to the null then set
MODEL THE BOMB

the display flag on the null node to display it. Double click on its
name and change it to BOMB_OUT.
Go to the Object level and rename the object to bomb_geo since it
holds the bomb’s geometry.

109
 PART TWO
Model the Fuse
To create the Fuse, start with a Bezier curve that emerges from the top of the bomb. You can then snake the
curve on the ground to create a longer fuse. Reverse the curve direction to get ready for animating the fuse
then add a Polywire node to give the fuse thickness.

2
01 Press spacebar-b to see the bomb from all angles. Move
over the Right view and press spacebar-b to expand it.
Press c to bring up the radial menu and choose Create > Geometry
1 > Curve. Click drag up to create the first point and tangent handle
for the curve. Next add a point and drag down to quickly draw the
curve pointing down.

02 Press x and choose Grid to turn on grid snapping. Click

and drag on the ground to create a third point with its
tangent handle aligned along the ground.

03
the view.
Press spacebar-b to go back to 4 views then mouse over
the perspective view and press spacebar-b again expand

5 Turn off Grid Snapping then turn on the Construction

Plane using the second button at the top of the Display
Options bar. This will make sure that any edits you make
4
stay on the ground.
Draw three new points with their tangents dragged out to define
6
the curve’s shape.
MMB-click to complete the curve.
HOUDINI FOUNDATIONS

TOOL HINTS
The curve tool comes with tool hints that display in the Scene view as
you work. These provide different shortkey options for this tool and
help you get familiar with how it works.
You can collapse it using Shift + F1 and only the tool name will display.
There are a growing number of tools that use tool hints and you will see
more of them in future versions of Houdini.

110
04 In the Operation Control bar, change the curve Mode to
Select/Edit. Now you can click on the edit points on the
curve and make changes to refine the shape of the curve.

Select and edit the points on the ground to get the desired look for
the curve. Tumble around to make sure that your curve stays above
the ground plane.

05 In the Scene View, press tab > Reverse. Press n to select

the whole curve and press Enter. Since the curve was
drawn from the bomb out, it will not animate in the direction you
need. This will put the start of the curve where the fuse begins.

06 Turn on the Display Points option in the

Display Options bar. Add a Resample node.
On the resample node, set the Maximum Segment
Length to 0.025 to add more detail. The resample node
evens out the points.

07 Add a Polywire node to add thickness to the wire. Set

the Wire Radius to 0.0075 and the Divisions to 8.
Between the reverse and polywire nodes, add a Transform node. Go
to the polywire node and RMB-click on the Wire Radius parameter
and choose Copy Parameter. Now go back to the transform
node and RMB-click on Translate Y and choose Paste Relative
References.
This will lift the whole fuse up so that it is not halfway under the
ground grid.

08 Add a Blast node Set Group to 0. This deletes the end of

the fuse geometry. Next add a Normal node to the end
of the chain.
After the normal node, add a Null node and name it FUSE_OUT.
Model the Fuse

This gives you a node representing the whole fuse geometry.

Go back up to the object level and rename this object to fuse_geo.

111
 PART THREE
Animate the Fuse
Animate the fuse using a Carve node that lets you control the length of the curve over time. Add a round cap to the fuse that
will be used to emit soot and sparks. You need to set up tangents on the curve to ensure that the cap follows along properly.
Next add some NULL objects to make it easier to export the cap for use in emitting particles.

01 Dive back into the fuse_geo object. Between the

transform node and the Polywire node add a Carve node.
Drag on the First U slide to see how this affects the curve.
Set First U to 0. Alt click on First U to set a keyframe at frame 1.
The parameter box will change color to indicate that it has been
keyframed and that there is a key at the current frame.

02 Go to frame 180. Set First U to 0.999. This will set a

keyframe. You can see that the fuse will. Go to frame
200. Set First U to 1.0. This will set another keyframe.
In the bottom left of the Playbar, toggle realtime playback on so
that it doesn’t play back too fast and press Play.
The fuse now animates into the bomb geometry where you will set
up the explosion.

03 Click on the Animation Editor Pane tab. Select the

animation curve and click on the Straight button at the
top of the panel. This will straighten the curve and the fuse will
animate evenly from start to finish instead of speeding up then
slowing down at the end.
Go back to the Scene View pane tab and playback the animation to
see how this change affects the motion.
Save your work.

04 In the Network View, add a Sphere node and set its

HOUDINI FOUNDATIONS

Display Flag. Set the following:

ƒ Radius to 1, 1, 1
ƒ Center Y to 0.0075
ƒ Uniform Scale to 0.0075
Press Spacebar-F to focus on it.
ƒ Orientation to Z Axis
ƒ Rows to 9 and Columns to 8
In the Scene view, press n to select all then tab>Clip node and set
its Direction to 0, 0, -1.

112
05 Tumble around and press s to get the select tool and 4 to
get face/primitive selection. Select one of the triangles at
the tip of the sphere then press and hold the a key and then middle
click two triangles over to select all the triangular faces. Press the
Delete key to remove them. This adds a blast node to the network.
Press 3 to go bring up edge selection then double click on the edge
of the area you just blasted. Press tab > polyfill which places the
node after the blast. Set Fill mode to Quadrilateral Grid then turn
on its Display flag. Set Smooth to 100 and Tangent Strength to 0.
This will create quad topology at the tip of the sphere.

06 Add a Copy to Points node into the network. Feed the

polyfill node from the sphere into the first input and the
carve node into the second input. Set Target Points to 0.
Add a Merge node. Feed the blast node and the copytopoints node
into it then wire it into the normal node.
The cap it positioned properly at the end of the curve but it isn't
oriented correctly. You need to add normals to the curve to allow
for proper alignment.

07 Add a Orientation Along Curve node between the

reverse and the carve nodes. Under Output Attributes
turn off the Y Axis option. Leave Tangent (Z axis) set to N. This adds
normals to the curve that will align the end cap as it moves along
the fuse.

08 Add a Color node after the polyfill node and set the Color
to yellow. Add another Color node after the fuse’s blast
node and set its Color to Dark Gray. These colors will be helpful for
visualizing the fuse as you work and can also be used to affect the
materials being assigned later.

09 Branch off a Null node after the copytopoints node and

place it to the side. Name the null node CAP_OUT. You
will use this later to extract the cap into another network that you
Animate the Fuse

will reference to emit particles. Set its Display Flag to see only
the half sphere. You can scrub the Playbar to see it move with the
carve.
When you are finished, set the Display Flag back to the FUSE_OUT
null.

113
 PART FOUR
Create an Animated Camera
As you develop this shot further, it would be helpful to have a camera set up to frame the final shot. This camera rig will be
built by constraining a null object to a curve then using an aim constraint to point the camera to the null object. This will give
you a camera that follows the end of the fuse to make it easier to evaluate the particles as they are being emitted.

2
01 Dolly out to see the whole scene from above. Make
sure the construction plane is on. Press c to bring up the
radial menu and choose Create > Geometry > Curve. Click-drag
a point near the start of the fuse and drag forward to extend the
tangent.
Next, click-drag a second point behind the bomb and drag to pull
out the tangent to create an s shaped curve. MMB-click to finish
and use Select/Edit mode if you want to tweak the shape.

02 On the Curve node, turn on the Z Axis option and set

it to N. This will create normals that will assist with
animation along the curve.

03 Add a Null object at the origin. Rename this node

camera_lookat_null. From the Constraints shelf, click on
the Follow Path tool. This accepts the null as the starting object.
Select the curve as the path object then press Enter. Press Enter
twice more since you don’t need a look-at object or a look-up
object.
Now if you scrub in the timeline, you can see that the null object is
moving along the Path from the first frame to the end frame at an
even pace.

04 A constraint network was added to the null object and a

HOUDINI FOUNDATIONS

path node is created. Go to frame 1. With the Path node

selected, RMB click on the Position parameter and choose Delete
Channels to remove the expression that is animating the null along
Straight Curve Drag Tangent Handle the path. Set Position to 0. Alt click on Position to set a keyframe.
Go to frame 195 and set Position to 1. Alt-click on Position to set a
second keyframe.
Click on the Animation Editor tab and press h to see the whole
curve. Select it and press the Straight button to make it linear. Grab
the tangent on the second point and adjust to smooth out the end.

114
 05 In the Network view, press tab > Camera and press Enter
then click to place it at the origin. Now use the Move tool
to move it in front of the fuse and a bit to the right. Next move it up
1 Move Along Z along the Y axis to raise it from the ground by about 0.75 units.

2 Move Along X

3 Move Up

06 From the Constraint shelf, click on the Look At tool. This

will use the selected Camera as the look at object. Select
the null object as the look at object and press Enter. Press Enter
again to not assign anything as the look up object.
Now the camera is looking at the Null object. From the Camera
menu, select cam1 to look through this camera.

07 Go to Frame 1. Select the camera and make sure the

Handle tool is active. This brings up a camera handle
which you can use to reposition the camera to have a better view
of the fuse's starting point.
Go to Frame 195. Use the same handles to reposition the camera
to make sure the bomb in nicely positioned in the frame.
You may need to scrub back and forth to tweak the camera to work
properly throughout the sequence. Make sure you can see the fuse
throughout the whole sequence.

08 Go to the object level and select the curve, camera_

lookat_null object and cam1 node and align them then
put them into a network box. Double click on the box’s title bar
and name the box Camera Rig. Turn off the display of all the parts
so that you don’t see them in the scene view as you work. You can
collapse this box if you want or keep it open if you want to work
with it further.
Save your scene.
Create an Animated Camera

CONSTRAINTS
To get the null object to follow the path and then to get the camera to look at the path,
you used animation constraints found on the Constraints shelf. The are accomplished
using a special node type called Channel Operators or CHOPS. You can find these
nodes inside the null and camera nodes. You can use these to control how the con-
straints work.
Another way to work with CHOP nodes is to use the Motion FX menu that you find
when you RMB-click on any parameter.

115
 PART FIVE
Create a Soot Trail
To create a soot trail, use the end cap to emit a trail of particles. Learn how to emit these points properly and how to add forces
such as gravity to control the motion of the particles. Learn how to set up collisions where the particles either stick to the
ground or slide off the bomb surface.

01 Dive back into the fuse_geo object node. With the display
flag on CAP_OUT, go to the Modify shelf and use the
Extract tool. Press n to select all the faces then press Enter to
create a new object with the Cap being imported using an Object
Merge node. Go up to the object level and name this object soot_
trail.
Go back to the fuse_geo object and set the display flag to FUSE_
OUT. Now you will see the combined fuse geometry when you
render and the new object will be used to generate particles.

02 Dive into the soot_trail node and add a POP network

node to the end of the chain. Dive in and on the Source
First Input node, set Const Birth Rate to 1000. Press Play. You can
see particles being emitted but they aren’t doing anything.
Go back up one level and on the Simulation tab, set Substeps to 3.
Press Play. You can see particles being emitted more evenly.

03 Alt-drag on the object_merge node to make a second

copy. Set Object 1 to the fuse_geo>resample node. You
are going to use this curve to transfer velocity onto the cap.
Now add a Orient Along Curve node and under Output Attributes
turn on Tangent (Z Axis) and set it to v.
Add an Attribute Transfer node between the original object_merge
and the popnet node. Turn off the Primitives checkbox and set
Points to v.

04 Place an Attribute Adjust Vector node in between the

HOUDINI FOUNDATIONS

attributetransfer and the popnet. Under Adjustment value

set the following:
ƒ Adjustment for to Direction Only
ƒ Adjust with to Noise
ƒ Range Values to Zero Centered
ƒ Amplitude to 0.5
Under Noise Pattern set Element Size to 0.025 and under Post
Process, turn on Enable Post-Process. Press Play to test.

116
 05 Dive back into the popnet and add a Gravity Force node
under the popsolver. Now if you play the simulation the
particles fall below the ground.

06 Go back up to the Object level and create a grid. Set size

to 30, 30 and Rows and Columns to 31, 31.
Rename this to ground. Go back into the popnet and add a Pop
Collision Detect node after the source_first_input node. Set the SOP
Path to the obj/ground/grid1. On the Behavior tab, set Response
to Stick. Keep Color Hits set to Red.
Add another Pop Collision Detect node. Set the SOP to the bomb_
geo geometry object. Set Response to Slide. Change Color Hits to
Green. Press Play.

07 Right now the particles will be emitted throughout the

whole sequence. You need them to stop them just as the
bomb explodes. Add a POP Kill node after the wire_pops_into_here
merge node. Go to frame 1. Set Activation to $F>199. This will kill
the particles at frame 200.
Now under the Rule tab, turn on Enable. Press Play to test.

08 Jump back up to the geometry context and wire the

popnet into an attribute create node. Set Name to pscale
and Value to 0.001.
Add a Color SOP and set the Color to dark grey.
Add a null to the end of the chain and name it SOOT_OUT.

PARTICLE FX
CREATE A SOOT TRAIL

The soot trail is created using particle dynamics. Particles are points that you can
affect using forces such as wind or gravity. Starting with the end of the fuse, you
give birth to points that are then simulated using a number of different techniques.
Particles are simulated using the Dynamics or DOPS section of Houdini then
brought back into SOPS where you can work with them as geometry. In the next
section, you will use particles to create sparks at the end of the fuse.

117
 PART SIX
Create Particle Sparks
To create sparks, start by copying the soot particle object and make changes to the new object to generate sparks. These par-
ticles will have shorter lifespan and will be more active. The Spark Trail node will give you the look you need to add sparks to
the shot. You can adjust parameters on this node to get the look you need.

01 Go to frame 1 then navigate to the Object level. Hide the

Ground object. Alt-drag on the soot_trail object to make
a copy. Name the copy sparks.
This new object already has a popnet and can be modified to
generate the spark particle simulation. In Houdini, it is often a good
idea to re-use a network you already have as opposed to building
everything from scratch.

02 Dive into the sparks object. You will make a few changes
to get the network set up to create sparks.
Delete the atttributecreate and color nodes. These are not needed
for this particle network. Rename the null node to SPARKS_OUT.
On the attributeadjustvector node change Amplitude to 1.75.
Add a new Attribute Adjust Vector node above this one. Turn on
Enable Pre-Process and set Constant Value to 0, 1, 0. Now the
particles will rise up before dropping down. Press Play.

03 Got Frame 1. Add an Attribute Adjust Float node just

before the popnet node. Set the following:
ƒ Attribute Name to life
ƒ Unit Settings to Duration
ƒ Pattern Type to Random.
ƒ Min Value to 2 and Max Value to 4
ƒ Under Random, set Seed to $F
HOUDINI FOUNDATIONS

SPARKS | PARTICLE TRAIL

The Particle Trail SOP takes an animated particle system and generates
motion trails from its particles. These trails can be used for various effects
such as sparks, fireworks, and rain.
This node also allows you to control the look of the trails. This enables
you to fine-tune their look in the SOP context, without the need to render
the points with large amount of motion blur.

118
04 Dive into the popnet and delete the two collisiondetect
nodes. You will not need to worry about collisions at this
point. You will add other nodes back later when needed.
Press Play to see the particles being emitted.

05 Go back up to the Geometry level. Go to frame 1.

In the Network view, press tab > Spark Trail. After a short
delay, this will put down two nodes - a timeblend and a sparktrail
node. Wire the popnet into the timeblend node and the sparktrail
into the SPARKS_OUT null.
Press Play. You will see sparks but they will be lagging behind by
the end of the fuse.

06 Add a Time Shift node after the object_merge which

brings in the cap and the attributetransfer node. Click
on the Frame parameter name to see the expression being used.
Change it to $FF + 1
This will move the cap forward a frame but at the same time align
the sparks with the end of the fuse. Press Play to test.

07 Go to frame 1. Select the sparktrail node. Click on the

Split tab and turn on the Enable Split checkbox. Set
Percent to Split to 40.
Under Shape, set Splits per Point to 4.
Press Play to test.

08 Go to frame 1. Jump up to the Object level.

CREATE PARTICLE SPARkS

Set the camera to cam1 and press the Flipbook button

on the bottom of the toolbar. Click Start on the window that pops
up. This will create an animated sequence of the scene view that
you can use to evaluate the motion of your particles.
Make sure the Realtime toggle is on and playback the flipbook
to see how the shot is progressing.

119
 PART SEVEN
Blow up the Bomb
For the bomb geometry, a rigid body dynamics simulation will be needed. Start by fracturing the geometry then adding
attributes that will create an explosion. You can then control the speed of the moving parts to art direct the look. Once the
simulation is ready, you will cache out the geometry to work more efficiently as you move on to the PyroFX stage.

01 Hide the fuse_geo, soot_trail, and sparks objects. Press

Spacebar-G to focus on the bomb geometry. Go to
Frame 1 in the Playbar. Press Spacebar G to focus on the bomb.
Select the bomb_geo object and from the Simple FX shelf, select
Simple Fracture. When asked to Select Collision Object just press
Enter with nothing selected. It will then take a short time to set up
this network. This object merges out the bomb geometry and sets
up nodes for fracturing and simulating.

02 On the fracture_solver node, set Start Frame to 200. Now

set the start frame in the Playbar to 200. Click the First
Frame button to go to frame 200. You will only need to simulate
the explosion from frame 200 to 240.
On the fracture_solver node go to the Collision tab in the Ground
Collision section, set Ground Type to Ground Plane. Click on the
Advanced tab then in the Constraints >Glue section, remove the
word Glue from the Data Name field.
Press Play to watch the bomb fall apart. Now you want to add a
starting velocity to explode the parts.

03 Go to Frame 200. Add an Attribute Adjust Vector node

and place it to the right of the other nodes. Wire the
third output of the rbd_configure node into the attributeadjustvector
node then wire the attributeadjustvector node into the third input
of the fracture_solver node.
Set the following on this node:
ƒ Turn on Enable Pre-Process
ƒ Turn on Overwrite Initial Value
ƒ Initial Vector to 0, 1, 0

04 Under Adjustment Value, set the following:

HOUDINI FOUNDATIONS

ƒ Adjustment for to Direction Only

ƒ Operation to Spread
ƒ Adjust with to Noise
ƒ Spread Angle Min to 15
ƒ Spread Angle Max to 120
Under Noise Pattern set Element Size to 0.5.
Under Post-Process turn on Post Process then turn on Length
Scale and set it to 20.

120
05 Press Play. Now you have the bomb exploding. Setting
velocity attributes on the geometry feeds into the
simulation which uses the initial velocity of the pieces to propel
them forward.
The manipulation of attributes plays a bit role in many visual effects
setups in Houdini.

06 Select the rbdconfigure node. Open up the Speed Limit

section and turn on the Speed Max and Spin Max
parameters. Set Speed Max to 2 and Spin Max to 30. Press Play.
The simulation slows down a lot from its original speeds.
Now set Speed Max to 10 and Spin Max to 60. Press Play.

07 Select the rbdmaterialfracture node and under Cell Points

set Scatter Points to 25. Press Play. Now you have a lot
more pieces.
Go back to the attributeadjustvector node and set Length Scale to
50 and Initial Vector to 0, 5, 0. Press Play.

08 Go to frame 200. On the fracture_io node, set Base

Name to exploding_bomb and Base Folder to $HIP/geo/
bomb. Click Save to Disk which will then turn on Load From Disk.
Press Play to preview the cached geometry.

09
OUT.
Add a Null node and wire the first output of the fracture_
io node into it. Rename this node EXPLODING_BOMB_
Blow up the Bomb

121
 PART EIGHT
Create the PyroFX Explosion
As the bomb explodes there needs to be an accompanying fireball. You will start with a Simple Fireball that works on the GPU
then make changes to create a look that works for the shot. You can also incorporate the parts of the exploding bomb to push
and influence the PyroFX volume in interesting ways.

01 Go to the Object level. From the Simple FX shelf, click on

Simple Fireball. In the Scene view, press Enter to place it
at the origin. This creates several nodes inside a fireball object.
Set the Display Flag on the pyroburstsource node and on the Burst
Animation tab, set Start Frame to 200. Go to frame 200. This node
represents the initial blast of the explosion.
From the Quick Setup menu, choose Single Input Point. This adds
a single point. Use the transform handle to raise it up to around 0.3
which is the middle of the bomb geometry.

02 On the pyrosolver_fireball node, change Start Frame

to 200. Set Simulation Type to Minimal OpenCL to
simulate using your GPU.
Under the Sourcing tab, open the Limit Source Range option and
set the Frame Range to 200, 240. Turn off Cycle Length.
Set the Display Flag on the RENDER_Simple_Fireball null node.
Zoom out to see more of the scene then press Play to test the
simulation. The Fireball is very big. Zoom out to see the whole
explosion.

03 Go to frame 200. On the pyroburstsource node go to the

Burst Shape tab and set Initial Size to 0.35 and Spread
Angle to 180.
Now go to the fireball Pyro Solver node and click on the Bound
tab and set Size to 15, 12, 12 and Center to 0, 4, 0. This will make
a smaller box which will fit better on the GPU. On the Setup tab,
set the Voxel Size to 0.035 - this will add more detail to this smaller
simulation.
Press Play to test the simulation. Now the explosion fits better in
this scene.

04 The next step is to integrate the Pyro FX with the

HOUDINI FOUNDATIONS

exploding bomb. Add a File node and load the $HIP/geo/

bomb/exploding_bomb.$F.bgeo.sc geometry sequence from disk.
Wire this node into the second input of the fireball Pyro Solver
node.
Press Play to test the simulation. The collision geometry isn’t having
an effect because it hasn’t been prepared properly.

122
05 Go to frame 200. Select the fireball node. From the
Quick Settings menu in the top right, choose Setup SDF
Collision. This adds a vdbfrompolygons node. This node gets it’s
Voxel Size from the Pyro Solver node.
Add an RBD Unpack node between the file and vdbfrompolygons
nodes.
Add an Unpack node between the rbdunpack and vdbfrompolygons
nodes. On the unpack node, set Transfer Velocity to v. Set the
Display Flag on vdbfrompolygons then press Play to see the collision
geometry.

06 Between unpack and vdbfrompolygons, add a Peak node

to make pieces bigger. Set Distance to 0.1. Between the
peak node and the vdbfrompolygons node, add a Attribute Adjust
Vector node. Turn off Adjust Value and turn on Enable Post Process
then turn on Length Scale and set it to 2.
Set the Display flag back to the fireball node. From the Collision
tab, open the Limit Collision Range section and set Range Type to
Frame Range and Frame Range to 200, 240. Turn off Cycle Length.
Press Play to test the simulation. The exploding bomb is now
colliding with and influencing the Pyro FX simulation.

07 Between the fireball Pyro Solver node and the pyrolook

node, insert a Pyro Post Process node. Turn on the
Convert to VDB and the Convert to 16bit Float checkboxes. Next
turn on the Cull Volume and Resample Volumes options and leave
them both set to vel.
This node will make your volume more efficient and save you disk
space when you cache your volume.

08 On the fireball Pyro Solver node click on Quick Setups

and choose Cache Simulation. Move the node to the
side then wire the pyropostprocess node into the sim_fireball cache
node then wire the cache node into the pyrolook node. Set Base
Folder to $HIP/geo/pyrosim/ then click Save to Disk.
Now with the Load from Disk option set to on, you can scrub in the
Playbar to see the PyroFX explosion.

Create the PyroFX Explosion

09 Set the Display Flag on the pyrolook node. This is the

Pyro Bake Volume node that can be used to visualize
the simulation in the Scene View. It is designed to provide a similar
interface to the Pyro Shader you will use later when rendering.
Under the Smoke tab, darken the Smoke Color then under the
Scatter tab set Intensity Scale to 2500. Make other adjustments to
get a look that you like for your simulation.

123
 PART NINE
Export the Geometry to USD
To set up the shot for rendering, you need to export the geometry to USD files that can be referenced into the Solaris context.
While you could import the geometry directly, having it cached out as USD will allow you to lock down your sequences then
focus on lighting and rendering in Solaris. For some of these objects, you will add UVs before exporting to prep for texturing.

01 Go to the Object level and hide the fireball object.

Double click on the bomb_geo and set the Display Flag
on the second polybevel node. This shows you the bomb geometry
before it was rotated into place.
Set the Display to Hide
Other Objects to stay
focused on the contents of
this network.

Inner Loop
02 Go to the Select tool then press 3 to select Edges. Press
w to go into wireframe mode and then press Shift and
double click to select part of the loop at the bottom. Press Shift and
double click to select all four parts. Repeat for the inner loop.
Now go to the loop on the top of the inner sphere and press Shift
Longitudinal
and double click to select. Now pick the latitudinal line that aligns
Loop with the X axis and press Shift and double click to select.
In the Scene View, press tab > Group. Change the Group Name to
Inner Loop uv_edges.
Bottom Loop

03 In the Scene View press tab > UV Flatten which adds

a uvflatten node after the group node. Set Seams to
uv_edges and under Layout Constraints, turn off Enable Manual
Layout. This provides a good looking UV layout that will be perfect
for texturing the bomb.
Set the Display Flag on the BOMB_OUT null node.

04 Add a USD Export node to the end of the chain. Rename

HOUDINI FOUNDATIONS

it static_bomb. Set the following

ƒ Valid Frame Range to Render Current Frame
ƒ Output File to $HIP/usd/static_bomb.usd
Press Save to Disk.

124
 05 Go back to the Object level then double-click on the
bomb_geo_fracture network. Select the fracture_io node
and press Save to Disk to cache out the geometry with the new
UVs. You won't see them yet because the geometry is packed.
After the fracture_io node and before the EXPLODING_BOMB_OUT
node, add an Unpack node then an Attribute Delete node. On the
attributedelete node, enter name next to Primitive Attributes. Turn
off the Point Attributes section.
This ensures that the sequence comes into Solaris as a single mesh.
The name attribute would break the sequence into individual parts.

06 After the attributedelete node, add a Normal node. This

will help the bomb geometry display properly in Solaris.
Add a USD Export node to the end of the chain. Rename it
exploding_bomb. Set the following
ƒ Valid Frame Range to Render Frame Range
ƒ Output File to $HIP/usd/exploding_bomb.usd
Press Save to Disk.

07 Go back to the Object level then double-click on the

ground network. After the grid node, add a UV Project
node and set its Display Flag. Go to the Initialize tab and press the
Initialize button. On the Transformation tab, change Scale X to -1,
Scale Y to 1 and Rotate Y to -90.
This will allow the texture to repeat on the ground surface rather
than create one large texture,

08 Add a USD Export node to the end of the chain. Rename

it ground. Set the following
ƒ Leave Valid Frame Range set to Render Current Frame
ƒ Output File to $HIP/usd/ground.usd
Hit Save to Disk.

Export the Geometry to USD

USD and SOLARIS

To support the look development stage of this project, layout, lookdev and lighting
workflows are set up in the Solaris context. This is represented by LOP networks. The
USD caches you are creating here will go into the Solaris context.
You will reference the USD caches into the LOP network using this scene file but in a
larger pipeline another option would be to start a new scene file and import the USD
files in a fresh scene. This would let you focus on lighting and rendering your shot but
would make it harder to go back and tweak the geometry and simulations.

125
 09 Go back to the Object level then double-click on the
fuse network. Add a USD Export node to the end of the
chain. Rename it fuse. Set the following
ƒ Valid Frame Range to Render Frame Range
ƒ Output File to $HIP/usd/fuse.usd
Hit Save to Disk.

10 Go back to the Object level then double-click on the soot

network. Add a USD Export node to the end of the chain.
Rename it soot_trail. Set the following
ƒ Valid Frame Range to Render Frame Range
ƒ Output File to $HIP/usd/soot_trail.usd
Hit Save to Disk.

11 Go back to the Object level then double-click on the

sparks network.
Add an Attribute Create node to the end of the chain. Set the
Name to width and the Value to 0.0005. This will determine the
look of the sparks when they are rendered.
Add a USD Export node to the end of the chain. Rename it sparks.
Set the following
ƒ Valid Frame Range to Render Frame Range
ƒ Output File to $HIP/usd/sparks.usd
Hit Save to Disk.

12 Go back to the Object level then double-click on the

fireball network. From the sim_fireball node, branch off a
USD Export node. Be sure to bypass the pyrolook node. Rename it
pyro_fireball. Set the following
ƒ Valid Frame Range to Render Frame Range
ƒ Output File to $HIP/usd/pyro_fireball.usd
Hit Save to Disk.
HOUDINI FOUNDATIONS

Scene Import
Another way to get your geometry and simulations into Solaris is using
the Scene Import LOP. This creates a direct connection between the
geometry and objects you are working with and the LOP network. This
approach would require cache LOPs to support motion blur which is an
extra step not needed if you are referencing in USD files.
This tool is going to be used to bring the animated camera from the
object level into the Solaris context.

126
PART TEN
Set up the Shot in Solaris
Learn how to reference all of the USD files into Solaris then import the camera from the object level. Apply materials
to all the elements and start rendering with Karma to evaluate the results. Learn how to add a key light and prepare render
settings to explore the final look of the shot.

01 Go back to the Object level. Press tab > LOP Network to

create a subnetwork to use to set up your shot. Name it
destruction_stage. Double click on the node to dive into it.
Change the desktop to Solaris. Make sure you see obj > destruction_
stage in the Scene view’s path bar. Press D over the scene view and
go to the Background tab and set Color Scheme to Dark.
In the Network View press tab > Reference then click to add
a reference node. Next to Reference Pattern, click on the File
Chooser and find the static_bomb.usd file. Rename the node to
static_bomb. Set the Primitive Path to /geo/$OS.

02 Alt-drag on this node to make three copies. Name them

exploding_bomb, ground and fuse then point the File
Chooser parameter to these USD files.
Feed these into a Merge node and set its Display Flag. After the
static_bomb node, add a Prune LOP. Set the Prune parameter to
$F >200. After the expoding_bomb node, add another Prune LOP.
Set the Prune parameter to $F <199. Now when you scrub in the
Playbar, the bomb will switch between the static and exploding
bomb at frame 200.

03 Press tab > Scene Import (Cameras). Place down

this node and wire it into the merge node. Set the
Destination Path to /cam/. This adds any camera found at the
object level into the Solaris context.
Go to the camera menu choose cam1 to look through this animated
camera. Make sure you have the merge node selected. Scrub in the
Playbar to see the fuse and the exploding bomb animate and then
explode through the lens of the camera you set up at the object
level. If you were to make changes to that camera it would be
reflected here in Solaris.
Set up the Shot in Solaris

04 In the Network View press tab > Reference then click to

add a reference node. Next to Reference Pattern, click on
the File Chooser and find the sparks.usd file. Rename the node to
sparks. Set the Primitive Path to /fx/$OS.
Alt -drag on this node to make a copy. Name it soot then point the
File Chooser and find the soot_trail.usd file. Add a Merge node to
bring these files together then feed them into the main Merge node
and set it's Display Flag. If you scrub in time, you will see the sparks
and the soot persist after they were killed. The USD file holds onto
the particles right until frame 240.

127
 05 After the effects merge node, add a Prune LOP. Set its
Display Flag. Set the Prune parameter to $F >199. Now
the soot and sparks disappear at this point.

06 Add a Material Library node after the merge node. Go to

the Material Palette and drag a Principled Shader and a
Concrete Shader into /stage/materiallibrary.
Go to the Network view and rename the principled shader to
bomb_mat. Alt drag to make three copies of the principled shader
and call them fuse_mat, sparks_mat and soot_mat.

07 Go back up one level and add an Assign Material node

to the end of the chain. From the Scene Graph, drag
static_bomb and exploding_bomb to the Primitives field.
Now click on the arrow next to Material Path and select the bomb_
mat. Press the plus sign to add more sections and repeat these
steps to assign materials to the fuse, sparks and soot.
Assign the concrete Material to the ground.

08 Go to around frame 180. Click on the Light tool in the

LOP Lights and Cameras shelf. This adds a light to your
shot and has you looking through it.
Go to the Base Properties tab and set Intensity to 50.
In the Scene View, click on the Shadow button and click on the
surface of the bomb then Shift click behind it to create a shadow.

09 In the Scene view, select Karma from the Persp menu.

HOUDINI FOUNDATIONS

Turn off the Reference Plane.

Go back into the Material Palette and select the concrete
shader. In the Texture section, set Effect Scale to 0.01.
Turn on the Denoiser in the side bar. This will use your
graphics card (nVidia cards only) to more quickly resolve
the noise in your rendering.

128
 10 Select the bomb_mat shader. Under Surface, set the
following:
ƒ Base Color to Black [0, 0, 0]
ƒ Roughness to 0.7
Under Displacement, turn on Enable Noise Displacement and set
the following:
ƒ Noise Type to Alligator Noise
ƒ Frequency to 30, 30, 30
ƒ Amplitude to 0.01
ƒ Roughness to 0.8

11 For the fuse_mat and soot_mat shaders, set the Base

Color to a Dark Grey.
On the fuse_mat shader turn off Use Point Color to allow the Base
Color to control the look.

12 Select the sparks_mat shader and under Emission set the

following:
ƒ Emission Color to 1, 1, 1 [white]
ƒ Emission Intensity to 10
ƒ Turn on Use Point Color
This will make the sparks shine brighter and even create a bit of
illumination on the ground surface.

13 Go back to the Stage level. In the Network View, press

tab > Karma to add a Karma Render Settings and USD
Render ROP node. Wire them into the end of the chain. Select the
karmarendersettings node and set Primary Samples to 32. On the
Image Output > Filters tab set Denoiser to nvidia Optix Denoiser
to turn the denoiser back on.
On the Advanced tab, go to the Sampling section and set
Convergence Mode to Path Traced.
Set up the Shot in Solaris

RENDER SETTINGS
The Karma Render Settings node adds render settings that become part
of the Scene Graph and will be used to render to the viewport and to disk.
Before you added these settings the viewport settings were being used
when you rendered in the viewport. If you have Karma rendering set up in
the Scene view then Press D to see these settings. If you have render set-
tings in your Scene Graph then they will override the viewport settings.

129
 PART ELEVEN
Render the PyroFX
To complete the shot, add the fireball USD file then assign the proper material. Next, you will set up another camera to create
a wide angle shot of the explosion and then render out the two sequences to achieve the final sequence. You can then preview
the results using the Mplay image viewer.

01 Set the Persp view menu back to Houdini GL. In the

Network View press tab > Reference then click to add
a reference node. Next to Reference File, click on the File Chooser
and find the pyrofx_fireball.usd file. Rename the node to fireball. Set
the Primitive Path to /fx/$OS.
Feed this node into the original merge node. Alt click on the
connecting wire to add a dot and move the dot to the lower right.
With the Display Flag on the Karma node, scrub to around frame
204 to see the explosion.

02 Navigate back to the Object level then into the fireball

object. Select the pyrosolver_fireball node. From the
Quick Setups menu, select Create Render Stage. This adds lopnet_
fireball into this network. You can dive into it to see the suggested
setup. You could use this to render the fireball on its own but you
need it as part of our existing LOP network.
Select the rendergeometrysettings node and press Cmd-C to copy it.

03 Navigate back to the destruction_stage LOP network and

press Cmd-V to paste it into the network. Wire it just
under the fireball reference node.
This node does two things. First it sets up velocity motion blur on
the fireball then it uses the volume to help light the shot.

04 Select the materiallibrary node and click the plus sign

HOUDINI FOUNDATIONS

next to Number of Materials. Click on the Operator

chooser button next to Material VOP on the new listing and
navigate into the fireball object then into the lopnet_fireball then
into the matnet to select the Pyro_Shader. Click Accept.
Even though this material is in a different LOP network, it can be
referenced from its location to this materiallibrary node.

130
05 On the assignmaterial node, add another material
listing and from the Scene Graph drag /fx/fireball to
the Primitives section then click on the Material Path arrow and
choose Pyro_Shader.
Set the Persp view menu back to Karma.

06 Add Prune node to the network and wire it just under

the rendergeometrysettings node. Under Pruning Options,
set Prune to $F<201.
The fireball at frame 200 was poking through the bomb geometry
and this will delay the explosion for one extra frame.

07 On the Karma node, set Valid Frame Range to Render

Frame Range. RMB-click on the End value which shows
240 and select Delete Channels. Change the End value to 210.
You will use the animated camera for the first 210 frames then cut
to a different camera for the last 30.

08 Set the Persp view menu back to Houdini GL. In the

Network view, Alt-drag on the karmarendersettings and
usdrender_rop nodes to the right. Set its Display Flag.

09 Tumble in the Scene view to get a new camera angle

that is looking down on the explosion from above. Press
the Ctrl key and click on the Camera tool in the LOP Lights and
Render the PyroFX

Cameras shelf.
Set its Primitive Path to /cam/$OS and its name to cam2.

131
 10 Lock the Camera to View and tweak the viewpoint to get
the shot you are looking for. Check it at various points
between frame 210 and 240. Turn the Lock the Camera to View
option off when you are ready.

11 Move the cam2 node up above the karmarendersettings2

node. On the usdrender_rop2 node, change the Start and
End to 211 and 240 then change the camera to point to /cameras/
cam2.
Select the first karmarendersettings node and make sure that its
camera is set to /cam/cam1. Otherwise the node will not render
since the default camera1 isn't in our scene.
Select both usdrender_rop nodes, and change the Output picture to:
$HIP/render/bomb/destruction_fx_$F2.exr

12 Select the usdrender_rop1 node. Click on the Render to

Disk button. Repeat for usdrender_rop2.
Go to the Render menu and select MPlay > Load Desk Files. Go to
the render/bomb directory and select the image sequence then click
Load. This will play the images as a single animation.

CONCLUSION
You have now built a destruction shot using
particles, rigid body dynamics and Pyro FX. You
built the complete project from scratch and have
experienced many of the tools and techniques
that Houdini artists use on a daily basis.
You have also had a chance to bring your
work into Solaris where you used USD to set
up the scene graph for rendering to Karma.
Now you can take these skills and begin
exploring your own Destruction FX shots.
HOUDINI FOUNDATIONS

132
HOUDINI FOUNDATIONS

TERRAIN
GENERATION
Houdini includes a dedicated toolset for generating and shaping terrains. These tools represent terrain
using 2D volumes, called heightfields, where each voxel contains the height of the terrain at a particular
grid point. The Houdini viewport lets you visualize 2D heightfields as 3D surfaces. You can also set up mask
fields that can be used to focus your edits to specific parts of the terrain. In this lesson, you will build up
terrains using patterns, noise and erosion then export the results for use in a game engine.

LESSON GOAL
Create a landscape using the Heightfield tools in Houdini and bring it into Unreal Engine or Unity.

WHAT YOU WILL LEARN

ƒ How to create a Terrain using heightfields
ƒ Add patterns, noise and distortions
ƒ Create Masks using terrain features
ƒ How to create Scatter Points on heightfield
ƒ How to set up Instancing using Terrain Scattering
ƒ Export the Terrain as a Digital Asset [HDA]
ƒ Import the HDA into Unreal Engine

LESSON
COMPATIBILITY
Written for the features in Houdini 19.5+

The steps in this lesson can be

completed using the following
Houdini Products:

Houdini Core ✔
Houdini FX ✔
Houdini Indie ✔
Houdini Apprentice ✔
Houdini Education ✔

Document Version 2.0 | July 2022

© SideFX Software
133
 PART ONE
Shape the Terrain using Heightfields
To create terrains in Houdini, you will work with heightfields. You will start with a blank height-
field then add some noise and distortion to define the basic look of the landscape. As you work,
you can tweak parameter values on the nodes while layering in details.

01 Go to the Desktop selector and choose Terrain. This

will give you shelf tools and radial menus focused on
terrain. Turn off the Reference Plane using the button at the top
of the Display options then press d in the viewport and from the
Background tab set Color Scheme to Dark.
Select File > New Project. Set the Project Name to terrain_lesson
and press Accept. Select File > Save As... You should be looking into
the new terrain_lesson directory. If not then click on $JOB in the
sidebar and you will be looking at the directory. Set the file name to
terrain_01.hip and click Accept to save.

02 From the Terrain Tools shelf, click on the HeightField

tool. Press Enter to place it at the origin. Press Spacebar
H to show the whole heightfield.
This defines a 1000 x 1000 grid with a grid spacing of 2.

03 Use the main radial menu (hotkey c) to select Deform >

Noise. Set the following:
ƒ Noise Type to Worley Cellular F1
ƒ Amplitude to 360
ƒ Offset to 20, 0, 300
This kind of noise gives you a good starting point for your terrain.
Use the main radial menu to choose Deform > Blur to access the
Heightfield Blur tool. Set the Radius to 20. This softens the edges
to make them feel worn by time.

04 Use the radial menu to choose Deform > Distort. Set the
HOUDINI FOUNDATIONS

following:
ƒ Amplitude to 40
ƒ Element Size to 220
This node moves the existing values around by advecting them
through a noise field. You can use the parameter values proposed
here or explore on your own to get a look that you like better.
Houdini’s procedural approach would even allow you to come back
later and change the parameter values to see how different settings
affect the outcome.

134
 05 Use the radial menu to choose Deform > Noise. Set the
following:
ƒ Amplitude to 10
ƒ Element Size to 20.
Select the last four nodes in the Network view then click on the
Network box icon to add a Network box. Adjust its edges to shape
the box then click in the top bar and name it Shape the Terrain.
Network boxes make it easier to read the network especially if you
share your file with other artists.

06 Use the radial menu to choose Mask > Mask by Feature.

Set the following:
ƒ Min Slope Angle to 35
ƒ Max Slope Angle to around 60
This lets you focus on areas on the side of the mountain.
If you were to go back and change the shape of the terrain feeding
into this node then the mask would update accordingly.

07 From the radial menu, choose the Erode > Slump. Set
Spread Iterations to 75.
The heightfield_slump node creates a type of erosion that moves
unstable piles of rubble into a more stable configuration. It affects
the Mask layer and also outputs to the Flow and Flow Direction
layers.
If you MMB-click on the node, you can see which layers are being
created. You can MMB-click on an earlier node in the chain for
comparison.

08 From the Terrain Tools shelf, choose the Heightfield

Remap tool. Set Layer to Remap to Flow then click on
the Compute Range button. Set Output Max to 1 to normalize
these values. Use the radial menu to choose Mask > Clear Mask.
This clears the mask layer so that you can use it to create more
layers in your setup.

Shape the Terrain using Heightfields

Select the last four nodes in the Network view then click on the
Network box icon to add a Network box. Adjust its edges to
shape the box then click in the top bar and name it Create Flow
Mask. Save your work.

HEIGHT LAYERS
The data passing through a geometry network can contain multiple height fields.
In the terminology of Houdini’s terrain tools, these are called height layers. For
example, a tool might use one height layer to represent bedrock and another to
represent loose soil on top. The default height layer is named height.
Each voxel contains a “selected-ness” value, called a mask layer. Most terrain
nodes take a second input that can contain a mask layer to control which parts
of the terrain the node will modify. The default mask layer is named mask and is
displayed as a red tint on the 3D surface.

135
 PART TWO
Add and Visualize Mask Layers
You can set up layers on your terrain by first populating the mask then copying that information
to a particular layer. You can do this more than once to add more layers. These layers can be
used later to visualize key aspects of the terrain.

01 From the radial menu, choose Mask > Mask by Feature.

Under Mask by Slope, set the following:
ƒ Min Slope Angle to 0
ƒ Max Slope Angle to around 45.
Turn on Mask By Direction then set:
ƒ Goal Angle to around 136
ƒ Angle Spread to 180.
These settings cover a larger area of the terrain including the
valleys.

02 From the radial menu, choose Layer > Copy Layer. Leave
the Source set to Mask and set the Destination to slope.
By copying the mask to a new layer, it leaves you free to clear the
mask and use it for other tasks.
From the radial menu, choose Mask > Clear Mask. This again clears
the mask layer so that you can use it to create more layers in your
setup. Add a network box to organize your nodes and call the box
Create Slope Mask.

03 Use the main radial menu to select Mask > Mask by

Feature. Under Mask by Slope, set the following:
ƒ Min Slope Angle to 0
ƒ Max Slope Angle to around 70.
Turn on Mask By Curvature and set:
ƒ Max Curvature to 0.5.
Next move the Curvature Ramp points in towards the center to
find the peaks of the terrain. This gives you a very detailed mask
that you can use to find the peaks of all the key features in the
landscape.

04 From the radial menu, choose Layer > Copy Layer. Leave
HOUDINI FOUNDATIONS

the Source set to Mask and set the Destination to peaks.

Afterwords, choose Mask > Clear Mask. This again clears the mask
layer.
You now have three layers that have all been derived from masks.
Add a network box to organize your nodes and call the box Create
Peak Mask.
You will use these in the next step to visualize the terrain.

136
VISUALIZING HEIGHTFIELDS
To visualize your heightfields, you start with a ramp that is assigned to
the overall height of the terrain. You start by Computing the Range that
the ramp will apply to then you can adjust the colors in the ramp itself
to visualize your landscape.
You can then add colors to the various layers that will sit on top of the
ramp. This lets you create a richer look for your scene.

05 From the radial menu, choose Visualize > Heightfield

Visualize. Click on the Compute Range button to align
the visualization with the current heightfield range.
This sets the ramp visualizer to go from the base of the terrain to its
peaks. You can see how the ramp looks in the 3D view where the
mountain tops are highlighted in white.

06 Under the ramp widget, set the following:

ƒ Layer 1 to peaks
ƒ Layer 2 to slope
ƒ Layer 3 to flow
The default colors are now visible in the 3D view using the three
layers you built up using masks. You will now use these to define
the look of the terrain.

07 Set all three of these layers to white. [1, 1, 1] This gives

a snowy look to the landscape. You can use these layers
for any number of different features, but for this mountain, snow is
the look that you want to emphasize.

Add and Visualize Mask Layers

08 In the Height Ramp, select and remove all the markers

except two. Set the one on the right to black and the
one on the left to dark grey. This creates a darker look under the
snowy layers which helps the dark areas pop out visually.
Save your work.

137
 PART THREE
Remap and Erode the Terrain
Right now some of the heightmap is below 0 and some is above. You are going to use a Remap
node to change the range and then use the ramp on that node to add a ridge around the moun-
tain. You will then erode the landscape to add new layers to the terrain.

01 Find the Heightfield Noise node at the end of the Shape

by Terrain Network box. Turn on its Display Flag and click
to select his node.
From the Terrain Tools shelf, click on the Heightfield Remap tool.
Click on the Compute Range button. To reframe the heights, set the
following:
ƒ Output Min to 0
ƒ Output Max to 300

02 Using the Remap ramp, add points in the middle then

add and adjust the following points:
ƒ Point 2 has a Position of 0.25 and a Value of 0.25
ƒ Point 3 has a Position of 0.4 and a Value of 0.27.
This creates a ridge running along the base of the mountain.

03 Choose Erode > Erode. Click on the Visualization tab and

click the Compute Range button. Press Play to watch the
terrain erode. Stop around frame 15.
 

04 Set the Display flag back to Heightfield Visualization

HOUDINI FOUNDATIONS

node at the end of the chain. Set Layer 3 to water which

is a layer that is coming from the erode node and change its color
to a blue. This will bring out some of these areas in the visualization.

138
PART FOUR
Scatter points on the Terrain
To add trees and rocks, you will mask out the new plateau area then set up a special terrain
scatter that will use this mask. These scattered points will then be used to copy instanced cones
designed to represent trees. These will be replaced later in Unreal.

01 Use the radial menu to choose Mask > Mask by Feature.

Set the following:
ƒ Min Slope Angle to 0
ƒ Max Slope Angle to 50
Turn on Mask by Height and set the following:
ƒ Min Height to 70
ƒ Max Height to 85
This should highlight the plateau created with the remap node. If
not then tweak these values until you isolate this area in the Mask.

02 In the Network editor RMB-click on the output of the

maskbyfeature node and start typing scatter…
Place the Heightfield Scatter down and set turn on its Display flag.
Now wire the output of the maskbyfeature node into the second
input of the scatter node to restrict the points to the area defined
by the incoming mask.
Set Coverage to 0.05. Turn off the Keep Incoming Terrain option.

03 In the Network Editor, press tab and begin to type Copy

to Points and click to add this node. Click the Pack and
Instance check box to turn it on. Wire Heightfield Scatter into the
second input of Copy then set its display node.
In the Network Editor, add a tube node down then feed it into the
first input of the copytopoints node. Set the following:
ƒ Radius to 0, 2 and Radius Scale to 1
ƒ Height to 10 and Center to 0, 5, 0.
Add a color node after the tube to make the trees green. Add a Scatter points on the Terrain
Merge node and feed Heightfield Visualize and copytopoints into it.

04 Go back to the heightfield_scatter node and turn off the

Match Normals with Terrain and Match Direction with
Slope. This will ensure that all the trees are pointing up.
Now change Random Up to 10 to create some variation in their
direction and set Randomize Yaw to 20. You cannot see the effect
of this but if you replace the tree later you will see random rotation.
The Scale of the trees is controlled by Variability. Change the Range
to 1, 2 to create random scales these two values.

139
 PART FIVE
Open the Terrain in Unreal
To bring the landscape into game engines such as Unreal Engine or Unity, start by creating a Houdini Digital Asset. Once
you have the Houdini Engine plug-in set up properly, then this asset can be loaded into the game editor with the copied
tree stand-ins importing as instanced objects. When you import the terrain into Unreal Engine, the heightfields will be
recognized as a landscape. You can also import the asset into Unity using the Houdini Engine plug-in.

01 Select all the nodes in the Network editor. Click on the

Create Subnet from Selected Button. RMB-click on the
new subnet and choose Digital Asset> Create New.
Set the Type Name to terrain and turn off Branch and Version. Set
Library Path to HIP File Directory and Library Filename to Node
Name.
Click Create. The Edit Type Properties window opens up. Click
Accept to close this window.

02 Open Unreal Engine and from the main panel click on the
New Project tab and choose the Third Person template.
Click Create Project. When it opens, delete the default geometry
so that it doesn’t get in the way of your terrain.
From the Content Browser, click Import to Game and find the
terrain.hda asset file. Drag the asset from the Content Browser to
the 3D workspace. Set the ThirdPersonCharacter up to a Translate
Z of around 10,000 then press Play and walk around the terrain.

03 Go to the Houdini Instanced Inputs section and expand

terrain1_1. This is the cone instance which you can
replace with content from within Unreal.
In the Content Browser, open StarterContent > Props. Drag the
SM_Bush prop over to the Houdini Instanced Input. Set the Scale
Offset to 5, 5, 5. The geometry is instanced to the points then
randomly scaled and rotated just like the cones.
In the outliner, select the Landscape node under terrain. Beside
Landscape material, click on the menu and find a grass material.
Press Play to explore the Terrain.

CONCLUSION
HOUDINI FOUNDATIONS

You can also use the terrain layers to create texture maps that you can work
with to define the look of your landscape. You can do this using the Heightfield
Output node. You can then use these channels to build a material in the Unreal
Engine that references the features of the landscape.
This quick look at the terrain features in Houdini will open up a wealth of possi-
bilities for artists creating landscapes for their games then populating them with
details such as rocks and trees.

140
CHARACTER FOUNDATIONS

KINEFX RIGGING | FUR DUDE

In this lesson, you will rig, animate and add fur to a two legged character named Fur Dude. Starting with
existing geometry, you will draw the skeleton, capture the geometry then build rig controls for an animation
rig. You will then keyframe a walk cycle and add fur to the surface of the creature.
This lesson uses KineFX, Houdini’s new SOP-based procedural rigging tools. While these tools are used
primarily for retargeting workflows they also include tools for rigging characters and creatures. These tools
are still evolving and this lesson offers a taste of what will be currently possible. In future releases, you will
see the KineFX and animation workflows expanded and refined.

LESSON GOAL
To rig, animate and add fur to the fur dude creature.

WHAT YOU WILL LEARN

ƒ How to build a skeleton using KineFX joints
ƒ How to capture deforming and rigid geometry to the skeleton
ƒ How to wrap up the capture rig into a digital asset
ƒ How to add controls and build an animation rig
ƒ How to animate a walk cycle
ƒ How to add fur to the creature
ƒ How to render using Solaris and Karma.

LESSON
COMPATIBILITY
Written for the features in Houdini 19.5+

The steps in this lesson can be

completed using the following
Houdini Products:

Houdini Core ✔
Houdini FX ✔
Houdini Indie ✔
Houdini Apprentice ✔
Houdini Education ✔

Document Version 2.00 | Nov 2022

© SideFX Software
141
 PART ONE PROJECT FILES
Draw the Skeleton Go to the Fur Dude tutorial page on
SideFX.com where you got this document
Start by opening the scene file and reviewing the Fur dude geometry then and download the furdude_lesson_start
place joints using the Skeleton tool. This tool will let you create, name and directory. Rename it furdude_lesson then
adjust joints to line up with the character you want to animate. put it into the Houdini Projects directory.

01 Select File > Set Project. Find the furdude_lesson

directory that you downloaded earlier and press Accept.
This makes this project directory and its sub folders the place for all
the files associated with this shot.
Select File > Open. You should be looking into the new
furdude_lesson directory. Open the file name to furdude_start.hip.
Select File Save As... and rename the file furdude_01.hip. Click
Accept to save. This way you can go back to the start file later if you
want to redo the lesson.

02 The scene opens with a single object called fur_dude_rig.

You are going to capture this geometry to bones created
using the KineFX toolset.
Double-click on the node to dive down to the geometry level. Here
you can see the File node which is importing the fur_dude_geo.bgeo
file from disk.
This geometry has some information stored with it such as
primitive colors and groups. To see this you can MMB-click on the
node to see Attributes and Groups listed. You will use the groups
later on to help capture the geometry.

03 Move your cursor over the Scene View and press

spacebar-b to go to a four view layout. From the icon
in the top right, turn on Link Ortho Views. Now you can pan and
zoom in the top, front and right views and they are all synced up.
Move your cursor over the Right view and press Spacebar-b again.
This is a good view for drawing joints.

04
CHARACTER FOUNDATIONS

In the Network view, press tab > Skeleton and place the
new node next to the File node. Set it’s Display Flag. Set
the file node’s template flag so that as you are working with the
skeleton node the geometry is visible as a grey wireframe.
Make sure the Handle tool is selected. In the top bar, set Joint
Placement to Freehand. This will draw on the construction plane
without taking the geometry into account. Click to place your first
joint just above the leg and then place six more joints as shown in
this picture.
 Start Here MMB-click to stop drawing joints.

142
 head
05 In the top bar, set Mode to Modify. Now you can edit the
joints instead of drawing them. Click on the first joint and
in the top bar set its Name it COG.
neck2 In the Parameter pane, click on the + sign in the tab area. From this
neck1 choose New Pane Tab Type > Animation > Rig Tree. This brings up
spine3 a pane that shows your skeleton joints. You can double click on the
spine2 second joint and name it spine1. Now either use the Scene View or
spine1 the Rig Tree to name the rest of the joints as shown.
COG

COG
pelvis 06 In the top bar, set Mode back to Create. By default
it will try to draw from the end of any selected joint.
MMB-click to stop this drawing action. Now click on the COG
l_hip
joint in the Right view then draw a pelvis joint just below it.
In the Scene View, press spacebar-b to go back to a four view In
the Front view, draw a hip joint on the left side of the character.

pelvis
07 Now go back to the Right view and draw the final four
joints for the leg as shown here.

l_hip Go back to Modify mode then rename the joints using either the
Rig Tree or by selecting and naming the joints in the top bar. After
the pelvis, all the other joints will have a “l_” prefix since these joints
l_knee
will be used for the left leg.

l_ankle

l_toe l_ball

08 In the Scene View, press tab > Skeleton Mirror. This

creates a mirror copy of all the joints. Go to the
Parameter pane and click on the arrow next to Group. Select only
the leg joints and press enter. Now only the leg is being mirrored.
Under naming, set Find Tokens to l_ and Replace Tokens with r_.
Now you have the right leg joint properly named.
Select File > Save to save your work so far.

KINEFX VS OBJECT LEVEL RIGGING Geometry Object

Level
DRAw THE SkELETON

Level
The KineFX tools in Houdini offer a joint-based workflow that takes place at
the geometry [SOP] level. Houdini’s other character workflow is bone-based 4 Joints 3 Bones
and in this case you work primarily at the object [OBJ] level.
In the KineFX workflow, joints are basically just points on a curve and this
opens up lots of opportunities to use sop-level tools to manipulate rigs. Here
you are going to learn about tools designed specifically for rigging characters
and creatures.

143
 PART TWO
Capture the Geometry
Rigging a character involves capturing geometry to the skeleton joints in such a way that rotating the joints deforms and bends
the geometry. Houdini uses a biharmonic capture method that gives great results with your first capture so that you can start
testing out your rig right away. Later you will paint capture weight s to refine the results to work with your character.

01 Change your Scene view to the perspective using

spacebar-b. Make your Network View a bit bigger so you
have room to work.
In the Network view, press tab > Joint Capture Biharmonic
and place this node down under the skeleton nodes. Wire fur_
dude_geo into the jointcapturebiharmonic node’s first input. Wire
skeletonmirror into the jointcapturebiharmonic node’s second and
third input and then set its Display flag.
You can now see capture weights on the geometry. You will refine
and paint these later to set up the deformation of the geometry.

02 In the Network view, press tab > Bone Deform. Wire the
three outputs of jointcapturebiharmonic into the three
inputs. Set the Display flag on bonedeform.
In this lesson, you are capturing teeth, claws and the eye which
you probably don’t want to deform. You will split these out later to
capture them using a different method.

03 Next press tab > Rig Pose and place down the
node. Move it over the third line connecting
jointcapturebiharmonic and the third input on bonedeform to add
it into the chain. This is where animation goes on the rig and the
rigpose will be used to rotate joints and can also be used to set
keyframes.

04
CHARACTER FOUNDATIONS

Select the rigpose node and make sure the Handle tool is
active in the Scene View. Select and rotate various joints
to test out the deformation. You can reset this later so feel free to
explore.

144
PART THREE
Add More Bones
It would be nice to have more bones in the mouth area. The procedural networks in Houdini let you to go back and add the
joints and all the other nodes including the biharmonic capture will update to reflect the changes. This gives you flexibility
when first setting up your creature’s rig.

01 In the Network view, set the Display flag on the skeleton

node and the Template flag on the original File node. Go
to the Right view in the Scene View using spacebar-b twice.
Set the Display Flag on the Skeleton node. Select the skeleton node
and turn on the Handle tool. Set Mode to Create. This will allow us
to add more joints to the skeleton. With KineFX, you can add bones
and the procedural nature of the other nodes in our network will
allow these changes to be accepted.

02 Click on the neck1 joint to start drawing then click two

joints at the jaw and lower mouth. When you have these
joints in place MMB-click to stop drawing and change Mode back
to Modify. Now either click on the joints and rename them jaw and
lower_mouth or rename them using the Rig Tree view.
jaw

neck1

lower_mouth

03 You can edit the joint positions in Modify mode. Turn

On the Tweak checkbox in the top bar and now you can
click drag on joints to move them. As you move a joint, all of its
children will also move and you may have to adjust them back into
place. To avoid this, you can turn on Child Compensate.
If you RMB-click on a joint, you will see options for splitting,
unparenting, copying and pasting joints. You won’t need to do
those things for this skeleton but it is good to know they are there.
You can even mirror joints here but you are using a different node
for that in this network.

04 Set the Display flag on the bonedeform node and turn off
the Template flag on the file node. The geometry will be
reconfigured then captured using the new bones. Select the rigpose
node and select the Handle tool.
Add More Bones

Click on the new jaw joint and rotate it down. This works but it
manipulates both the lower lip and upper lip. It would be better if
it only affected the lower lip. To fix this you will paint the capture
weights in a later section of the tutorial.

145
 05 Set the Display flag back to the skeleton node and the
Template flag on the original File node.
Select the skeleton node and turn on the Handle tool. Set Mode
to Modify. Click on the head joint and using Tweak Mode, lower it
until it aligns with the eyeball.
Set Mode to Create. Click on the head joint and then click a new
joint right in the center of the eyeball. Switch to Modify mode and
Name this joint eyeball.

06 Set Mode to Create. Click on the head joint and then

click a new joint above the eyeball joint. MMB-click to
deselect then again click on the head joint and then click a new joint
below the eyeball joint.

Switch to Modify mode and Name these joints upper_lid and lower_
lid. You will use these joints to rotate the eyelids during animation
but they need to also be at the center of the eye.

07 Turn off the Template flag on the geometry. Select the

upper_lid joint and use Tweak Mode to drag it down on
top of the eyeball joint. Repeat with the lower_lid to overlap the
eyeball joint.
You now have all three joints in the same spot. Later you will attach
geometry to each of them individually and then you will be able to
animate them independently.

08 Set the Display flag on the bonedeform node. The

geometry will be reconfigured then captured using all the
new bones.
You are not able to pose the new joints yet because the geometry
hasn’t be attached to them in a rigid manner yet. That will come a
couple steps down the line after you paint capture weights on the
body and the tongue.
CHARACTER FOUNDATIONS

JOINT ORIENTATION
With a joint based system where each joint is a point, the joint orientation is
critical because it defines how the joint will rotate during forward kinematics.
KineFX has tools that allow you to point along the chain to the next joint or
you can rotate each joint by hand using the child compensate option to make
sure that re-orienting the joints doesn’t affect the other joints in the chain.

146
PART FOUR
Joint Orientations
When you animate the character, the orientation of the joints plays a significant role in how you manipulate the rig. At this
point, you will orient some of the joints by hand and then use a Re-orient joints node to point all the other joints down the -z
axis. Sometimes when you evaluate your rig at a later stage, you may need to come back and tweak the orientations.

01 Go back and set the Display flag on the skeleton node. In

the Scene View , RMB-click and select Display Joint Axes
to see the joint orientations. Select the COG joint and RMB-click
and select Show Handle. Now if you rotate the joint the whole rig
goes with it. Press Ctrl-Z to undo.
Click on the Child Compensate checkbox and now you can rotate
the joint without changing the rest of the skeleton. You can use the
Ctrl key to constrain to 45 degree increments.

02 In the Scene View, press p to bring up the parameters for

this joint. To orient the COG with the world, Set Rotate
and Local Rotate to 0, 0, 0.
Click in empty space to deselect the COG joint then click on the
Pelvis joint. Set Rotate and Local Rotate to 0, 0, 0.

03 After the Skeleton joint, insert a Orient Joints node which

will by default point the orientations along the positive Z
axis. Click on the arrow next to Orient group and in the Scene view,
select all the joints. Press Ctrl and select the COG, pelvis and neck1
joints to remove them from the selection. Press Enter.
Now all of the joints are oriented along positive Z except for the
three deselected joints.

04 Set the display flag back to the Bone Deform node to

allow all the other nodes to update and accept the new
joint orientations.
JOINT ORIENTATIONS

The results don’t look any different than before you oriented the
joints but it will affect how you pose and manipulate the character
when you are animating.

147
 PART FIVE
Attach Capture Geometry
Right now you are using the skeleton joints to capture geometry and assign capture weights to each point
on your character. To manage this with more control, you can attach curves to the joints that will extend the
influence of that joint. This provides a method for achieving your goals as quickly as possible.

01 In the Network View, press tab > Split and place the node
between the File node and the jointcapturebiharmonic
node. The first output of the split node should be wired into the
first input of the jointcapturebiharmonic node.
Click on the pull down menu in the Group field and choose the
fur_dude_body and fur_dude_tongue groups. These are now feeding
the first output of the split node and all the remaining pieces such
as the eyeball, teeth and claws are going out the second output.
You will use a different method to bind those to the skeleton after
painting weights on this geometry.

02 In the Network View, press tab > Visibility and place the
node between the jointcapturebiharmonic node and the
bonedeform node. Click on the pull down menu in the Group field
and choose the fur_dude_tongue group.
This node hides the selected geometry but does not remove it.
This ensures that point numbers and primitive numbers don’t get
changed which is important when working with the paint capture
weight tool. You don’t want to change this information every time
you hide some geometry during this process.

03 Press S to get the Select tool and 3 to go to Edge

selection. Select an edge in the middle of the lower lip on
the left then press Shift-A and select an edge on the other side.

04
CHARACTER FOUNDATIONS

Press tab > Curve from Edges and the selected edge
will be extracted from the geometry. Rename this node
lower_lip.
The curvefromedges node gets placed in between the file
and jointcapturebiharmonic nodes. Move it to the side and
then reconnect the file node output to the first input of the
jointcapturebiharmonic node. This branches the curvefromedges
node off to the side.

148
05 Press S to get the Select tool and 3 to go to Edge
selection. Select an edge on the upper lip on the left then
press Shift-A and select an edge on the other side.
Press tab > Curve from Edges and the selected edge will be
extracted from the geometry. Rename this node upper_lip.
Branch this node off the same as you did with the other one.

06 Press tab > Merge Packed and place this node down
underneath the two extracted curve nodes then feed the
output of those nodes into the mergepacked node.
Press tab > Attach Capture Geo. Add this node to the network
then feed the skeletonmirror node into the first input and the
mergepacked into the third input.

07 Feed the add_capture_geometry node’s output to the

middle input of jointcapturebiharmonic. Under Advanced
Shape Settings, click the Assign Shapes plus sign button two times.
For the first one, set:
ƒ Group to @name=lower_mouth
ƒ Shape Name to lower_lip
And the second one to;
ƒ Group to @name=neck2
ƒ Shape Name to upper_lip
Turn On keep Shape World Transform for both Shapes.

08 Set the Display Flag on jointcapturebiharmonic. Now

bypass the add_capture_geometry node to show
difference in the capture weights.
When this node is turned on, the attached curves are assisting with
the capture of geometry to the associated joint. This allows for
more control as you set up your character.

Attach Capture Geometry

09 Set the Display flag on the bonedeform node. Click on

the rigpose node and in the Parameter pane, click on the
Clear button next to Transformations. This will reset the joints.
Now with the Handle tool active and in the Scene View click on the
jaw joint. Rotate it to lower the lips down.
You can see that this still deforms parts of the belly. To correct this,
you need to paint capture weights to reassign capture weights to
different joints.

149
 PART SIX
Paint Capture Weights
The biharmonic capture added weights to the character’s geometry that are associated with the different skeleton bones. You
can now use a new node to adjust the capture weights using a brush workflow. For this creature, the goal will be to get the top
of the mouth to not be influenced by the lower mouth joint and to tweak how the feet area is weighted.

01 In the Network View, press tab > Joint Capture Paint and
place the node just above the bonedeform node with all
three connectors getting connected. Set its Display Flag then click
on the pull down menu next to Target Joint and choose the neck2
joint.
In the Scene view, you will also see a big round paint icon on your
cursor which you can use to paint weights. Paint on the head area
and this part of the geometry will be captured to the neck2 joint.

02 You want to capture the head and the upper lip area to
the neck2 joint. The geometry will appear red when it is
at its strongest.
Use big strokes for the top of the head then use the Scroll wheel on
your mouse to reduce the brush radius or go to the Brush tab in the
Parameter pane to change the radius.

03 Paint on the upper lip to increase the influence of the

neck2 joint on this area. Make sure you are painting the
top lip and not the bottom lip. If you mess up, use Ctrl-Z to undo
your stroke. Tumble around, even inside the mouth to get the upper
part of the mouth.
In the Operation Controls panel, you can see options to Display
Deformed Geometry, Display Joints and Display Color. You can
turn these on and off to help you evaluate your capture weights as
you paint.

04
CHARACTER FOUNDATIONS

Press F and choose Smooth to smooth out what you

have painted here with other capture weights. Make your
brush radius a bit bigger to create more smoothing.
Once this is finished, click on the rigpose node with the Handle
tool active and in the Scene View click on the jaw joint. Rotate it to
lower the lips down. You can see that now the top lip isn’t moving
but other parts of the head are being rotated.

150
05 Select the jointcapturepaint node. Click on the pull down
menu in the Capture Region field and choose the jaw
joint. Press F and choose Subtract to take away influence of this
joint on the top of the head and eye area.
You can also use this method to remove the influence of the jaw
and the lower_mouth joints on the belly. That way when the mouth
moves the belly isn’t affected too much.

06 From time to time, go back to the and test out the jaw.
Once you have it no longer influencing the head area
then you are good to go. You can also go back to smoothing to
clean things up once you are finished.
You can now paint weights for other joints in the skeleton. The
Biharmonic should have done a good job on the legs and feet but
you can test the rig using rigpose then paint weights to refine the
results.

07 Select the toe joints and make sure that the ends of the
feet are captured to those joints. Later you will use a
reverse setup and having the geometry attached to the toes .

08 On the visibility node set Apply to Non-Selected

Primitives . Now you only see the tongue.
It should also be weighted fine by default but you can paint weights
if you need to. Select the jointcapturepaint node and paint weights
for the neck1 and jaw joints.
When you are finished then Bypass the visibility node again.

09 On the visibility node set Bypass Flag.

Paint Capture Weights

Once this is finished, click on the rigpose node with the

Handle tool active and in the Scene View click on the jaw joint.
Rotate it to lower the lips down. If you are happy with how all the
parts are working then you are ready to capture the rigid geometry.

151
 PART SEVEN
Capturing the Rigid Geometry
Earlier you split out the eyes, teeth and claw geometry. Now you are going to pack this geometry then assign each part to a
joint using the capture packed geometry node. This is the equivalent of parenting each object to the skeleton since parenting
isn’t an option at the geometry level when using KineFX.

01 In the Network View, press tab > Name from Groups and
place the node between the File node and the split node.
Change the Group Mask to *. This will turn all of the groups into
name attributes.
To see this, click on the Geometry Spreadsheet Tab and click on the
Primitives button at the top left. Scroll down to see that there is a
name attribute and the group names are used as the values for all
of the geometry.

02 In the Network View, press tab > Pack and place the
node down to the right of the split node. Wire the second
output of the split node into the pack node and set its Display Flag.
You can see the body parts in the Scene View.
Turn the Path Attribute checkbox to Off then the Name Attribute
checkbox to On and set Transfer Attributes to name.
Using the arrow in the top right of the Network view, choose Split
In the Geometry Spreadsheet, you can see the 8 packed primitives.
These can now be captured to the skeleton as rigid geometry.

03 In the Network View, press tab > Capture Packed

Geometry and place the node down under the pack
node. Wire the output of the pack node into the first input of the
capturepackedgeo node. Next wire the output of the skeletonmirror
node into the second input of the capturepackedgeo node then set
the Display flag on the capturepackedgeo node.
Nothing has happened yet - you need to associate the bones and
the geometry you want to capture.

04
CHARACTER FOUNDATIONS

In the Parameter pane, click on the plus sign next to

Manual Capture. Click on the arrow next to Capture Geo
and in the Scene view, select the lower teeth. One click is all you
need because the lower teeth are one of your packed groups. Press
Enter to accept.
Now click on the arrow next to Joint. The geometry disappears and
the skeleton is shown as lines and points. Select the lower_mouth
joint and make sure your cursor is over the Scene View then press
Enter to accept.

152
 05 Click the plus sign and then the arrow next to Capture
Geo and in the Scene view, shift select the fur_dude_
upteeth and fur_dude_gums . Press Enter to accept. Now click on
the arrow next to Joint. Select the neck2 joint and press Enter.
Repeat these steps two more times to associate the following:
ƒ fur_dude_rclaws to r_toe
ƒ fur_dude_lclaws to l_toe
ƒ fur_dude_eye to eyeball
ƒ fur_dude_uplid to upper_lid
ƒ fur_dude_lowlid to lower_lid

06 Wire the output of the capturepackedgeo into the first

input of the bonedeform node. Set the Display flag on the
bonedeform node.
Select the rigpose node and then select and rotate joints in the
Scene view. You can see that the geometry is being captured to the
joints without any deformation.
The eye and eyelid joints overlap therefore when you click a little
menu pops up letting you choose - you may need to click a few
times to get the joint you need. You can then use them to rotate
the eye and the eyelids separately.

07 In the Network View, press tab > Merge and place

the node down in-between the capturepackedgeo
and bonedeform. Wire the capturelayerpaint node into the merge
node. On the merge node, press the blue up arrow next to
capturelayerpaint to reorder the inputs.
Now everything is deforming but the body and tongue are grey
while the packed geometry has its colors. There is also an error on
the merge node because the one side has color (Cd) as a primitive
attribute and the other side has it as a point attribute.

08 In the Network View, press tab > Unpack and place the
node down between the capturepackedgeo and the merge
nodes. Set Iterations to 2 and Transfer Attributes to * ^Cd. The *
grabs all the capture attributes and the ^Cd make sure that you do
not remove the original color attribute. Now both sides are using
point colors and the rig looks correct.
Now you can select the rigpose node and then select and rotate
joints in the Scene view. You can see that all the captured geometry
is now working together. Capturing the Rigid Geometry
Save your work.

FLATTENING THE NETWORK

The network that you have created to capture the geometry, paint weights
and prepare the geometry for animation works well but can take a little time to
update when changes are made.
To create a more efficient rig, you will flatten the network into a single file that
has capture weights stored in the geometry. This file will deform efficiently
when fed into a bone deform sop when accompanied by a skeleton with the
same bones that were used to capture the geometry in the first place.

153
 PART EIGHT
Create Capture Rig Digital Asset
The captured geometry and the skeleton can now be wrapped up into a digital asset that can be used as the foundation for the
animation rig. To start, you will export the geometry with its capture weights and the skeleton then embed these into the digital
asset file. This will help make the capture rig more efficient when animating the character.

01 In the Network View, press tab > ROP Geometry Output

and place the node down under the merge node. Wire
the output of the merge node into the input of the rop_geometry
node. In the Parameter pane, set Output File to:
$HIP/geo/furdude_capt.bgeo.sc
Click Save to Disk to save the geometry to your geo directory. This
geometry looks like the geometry you imported at the beginning
of the lesson but now it includes important data such as capture
attributes that allow it to deform.

02 In the Network View, press tab > ROP Geometry Output

and place the node down next to the skeletonmirror node.
Wire the output of the skeletonmirror node into the input of the
rop_geometry node. In the Parameter pane, set Output File to:
$HIP/geo/furdude_skel.bgeo.sc
Click Save to Disk to save the geometry to your geo directory. This
geometry represents your skeleton and can be used to build up an
animation rig asset.

03 Go back up to the Object level. Rename the object to

fur_dude_capture. Turn of it’s Display flag.
Click on the File button on the Create shelf. Click on $HIP then go
into the geo directory and select the furdude_capt.bgeo.sc file. Press
Enter to place it at the origin then rename the new object node to
furdude_rig. Double click to dive into this object. Alt-drag on the
File node to create a second one. Set its Geometry File to:
$HIP/geo/furdude_skel.bgeo.sc
Rename it furdude_skel.bgeo then set the Display Flag on the
furdude_capt.bgeo node.

04
CHARACTER FOUNDATIONS

Select the two File nodes then from the Assets menu
select New Digital Asset from Selection. Set the
following:
ƒ Operator Name to fur_dude_capture_rig
ƒ Operator Label to Fur Dude Capture Rig
ƒ Save to Library to: $HIP/hda/fur_dude.hda
Click Accept to create the HDA file. An Edit Operator Type
Properties window pops up. Set Maximum Outputs to 3 and click
Accept. Rename the subnet to fur_dude_capture_rig.

154
 05 Double click on this node to dive into the subnet. Press
tab > Output and place an output node beneath the fur_
dude_geo_capt File node. Wire the File node into the output node.
Rename it CaptureGeo.
Alt-drag twice to create two new Output nodes. Name the second
one RestSkeleton and set its Output Index to 1. Name the third one
AnimSkeleton and set its Output Index to 2. Wire the fur_dude_skel.
bgeo File node into both the second and third output nodes.
Set the Display Flag on the CaptureGeo output node.

06 From the Asset menu, select Edit Asset Properties >

Fur Dude Capture Rig. This opens up the Operator Type
Properties window. Click on the Extra Files tab.
Click on the chooser button next to Filename in the lower left and
navigate to $HIP/geo/fur_dude_capt.bgeo.sc. Click Accept. Then
click Add File.
Repeat these steps for the fur_dude_skel.bgeo.sc file. Now you have
placed these files inside the digital asset file which will make them
easier to share with other people as a complete package.
Click Accept to finish.

07 On the first file node, click on the Chooser icon next

to Geometry File then click on opdef:/ in the Locations
sidebar then double click on the sop directory then again on the
fur_dude_capture_rig folder. Select the fur_dude_capt.bgeo.sc file
then press Accept. This creates the following opdef expression:
opdef:/Sop/fur_dude_capture_rig?furdude_capt.bgeo.sc
Repeat these steps for the fur_dude_skel.bgeo.sc File node.
From the Assets menu, select Lock Asset > Fur Dude Capture Rig.
Save Changes to protect the contents of this asset. Later you can
unlock it and update these files if needed.

08 Go up one level where you will see the capture rig with
three outputs. Press tab > Bone Deform and place this
node underneath. Wire the three outputs of the fur_dude_capt_rig
into the three inputs of the bonedeform node. Now add a rigpose
node in the middle of the third chain. Set the Display Flag on the
bonedeform node.
Select the rigpose node and make sure the Handle tool is active.
You can again see all the skeleton joints. Pose the skeleton to test
that the deformations are working in the same way as before. CREATE CAPTURE RIG DIGITAL ASSET

HDA
This network of nodes was saved as a Houdini Digital Asset or HDA. This is a
file on disk that is easy to share. The capture rig is an asset that is referenced
off disk and in the next section, you will build the animation rig with the cap-
ture rig nested inside it. Both of them along with the bgeo files will be stored
in a single HDA file.
Changes made to the HDA file on disk will update instances of the asset being
worked on by animators in their scene files.

155
 PART NINE
Create the Animation Rig Asset
You are now going to create a second digital asset that has the capture rig nested inside it. This new asset will be the one that
can be animated to create the final motion of the character. This new asset will contain all of the rigging tools such as inverse
kinematics and aim constraints that assist with animation. In order to test these controls as you add them, you will set up a test
version of the rig that is locked and visible in a second Scene view pane.

01 Select the three nodes then from the Assets menu select
New Digital Asset from Selection. Set the following:
ƒ Operator Name to fur_dude_anim_rig
ƒ Operator Label to Fur Dude Anim Rig
For Save to Library, click on the browse button then click on $HIP
the double click into the HDA directory. Select the fur_dude.hda
file then click Accept. It is now set to: $HIP/hda/fur_dude.hda
Click Accept then Accept again in the Type Properties panel. This
adds the new asset definition to the same HDA file. Rename the
subnet to fur_dude_anim_rig.

02 Go to the Object level. Alt drag on the furdude_rig node

to create a new geometry node and name it test_rig.
Press t and move the test rig over to the left.
Double click to dive into it. Select the fur_dude_anim_rig node and
from the Asset menu, select Lock Asset > Fur Dude Anim Rig.
Go back up one level. You now have two versions of the rig - the
test_rig is locked and shows you what it is like to interact with the
completed asset. Right now you have nothing to interact with. You
will fix that soon.
Click in the Network editor, and press Ctrl 1 to set a quickmark.

03 Navigate back to the furdude_rig object then into fur_

dude_anim_rig. Click in the Network editor, and press Ctrl
2 to set a quickmark. Now you will be able to quickly jump back and
forth between these networks.
Select rigpose then click on the Clear button next to
Transformations. Go to the Asset menu, select Save Asset >
Fur Dude Anim Rig. From the Asset menu, choose Edit Asset
Properties > Fur Dude Anim Rig. Click on the Node tab and set
Default State to: kinefx__rigpose. Click Accept.

04
CHARACTER FOUNDATIONS

In the Network view, press 1 to navigate back to the

test_rig. In the Scene View on the left, expand the
toolbar and click on the Handle tool.
You will now see joints displayed on the test rig. If you click on
any of them, you will see that they can’t be changed because
their parameters have not been promoted to the asset. This is the
beginning of building an animatable interface for this character.

156
PART TEN
Add More Control Joints
To provide more flexibility with the control rig, you can add joints such as a root joint for the whole skeleton, heel joints for a
reverse foot setup or a look-at point for you to target with your eyeball. These joints will have the same names as the ones in
the original rest skeleton and that will ensure that they are used to drive the motion on the character.

01 In the Network view, press 2 to navigate back to the fur_

dude_anim_rig and set its Display flag. Set the Template
flag on the fur_dude_capture_rig node.
Branch off a Delete Joints node from the third output of the fur_
dude_capture_rig node. Click on the arrow next to group then select
fur dude’s right leg joints. Press Enter.
Add a Skeleton node. Change to Create Mode. Turn on Grid
snapping and add a point to the origin. Change back to Modify
mode and rename the new joint to furdude_main.

02 Go to right view. Go back to Create model and MMB-

click to break the connection to the main joint. Click on
the l_toe joint then add a new joint where the heel would be.
Change back to Modify mode and rename the new joint to l_heel.
Press p and with Child Compensate set to On, set the Rotate to 0,
0, 0.

l_heel

03 Pan to the area around the eye. Go back to Create model

and MMB-click to any connections. Now add a new joint
out in front of the eye.
Change back to Modify mode and rename the new joint to eye_
target. Press p and make sure that Rotate is set to 0, 0, 0.
eye_target

ADD MORE CONTROL JOINTS

ADDING EXTRA JOINTS TO THE RIG Rest

Skeleton
These extra joints were added to the stream of nodes on the far right while Skeleton with
the middle rest skeleton will continue with the original joints. When fed into Extra Joints
the bone deform, the extra “phantom” joints are ignored and only the original
joints determine the final output of the rig.
If you were to feed these into the middle rest skeleton input on bone deform
you would get an error because these extra joints would not have correspond-
ing capture weights in the incoming geometry.

157
 04 Add a Parent Joints node. Click the plus sign twice. Click
on the arrow next to Joint1 and in the scene view click
on the COG joint. Press Enter (with your cursor over the Scene
View) to accept. Now use the arrow next to Parent1 and select the
furdude_main joint.
For the second entry, click on the arrow next to Joint2 and in the
scene view click on the eye_target joint. Press Enter (with your
cursor over the Scene View) to accept. Now use the arrow next to
Parent2 and select the COG joint.

05 Add a Skeleton Mirror node to the chain. This creates a

mirror copy of all the joints. Go to the Parameter pane
and click on the arrow next to Group. Select only the left leg joints,
including the new l_heel joint,and press Enter. Now only the leg is
being mirrored.
Under naming, set Find Tokens to l_ and Replace Tokens with r_.
Now you have the right leg joints properly named.

06 Feed the skeleton mirror node into the rigpose node.

Set the Display flag on the bonedeform node. The claws
on the foot appear to have flipped. Go back to the skeleton joint,
select the l_toe and press p to bring up the parameters. Set Rotate
to 0, 0, -90.
Select the four nodes you used to add joints and click on the Add
Network box button. Position the box and center the nodes. Click
on the box’s title and enter Add Joints.

07 From the Assets menu, select Save Asset > Fur Dude
Anim Rig. This saves the changes into the asset definition
which will update on the test_rig. You still can’t do anything with the
test_rig because there are no parameters promoted to the top level.
You are now going to setup the main controls and promote
parameters to start bringing the test_rig to life.
CHARACTER FOUNDATIONS

WHAT IS THE ROLE OF THE TEST RIG?

The rig you are working on is an unlocked asset which gives you the ability to work with all
the joints inside the asset. But when the asset is published to an animator, they can only
work with parameters that have been promoted to the top level of the character.
The test_rig is a second locked version of the asset and until you start promoting parameters
and building controls, you will not be able to manipulate it. That is what makes it a great tool
for verifying that the asset is ready for animation - if you can’t work with the test rig then
the animator will not be able to pose the character.

158
PART ELEVEN
The Main Controls
To add kinematics, you need to break the current hierarchy where the feet are under the COG. You can break off a few joints
and reparent them to build the hierarchy you need. This reparenting happens off to the side then you will blend the results back
into the original skeleton hierarchy which is important to make sure the bone deform functions properly.

01 Branch off a Delete joints node from the skeletonmirror

node. Click on the arrow beside Group and in the scene
view select the furdude_main, the COG, the pelvis, the l_heel and r_
heel joints. Press Enter then set Operation to Delete Non-Selected.

02 Add a Parent Joints node after the deletejoints node.

Click on the + sign to add a joint listing. Click on the
arrow next to Joint1 and in the scene view select the two heel
joints. Press Enter.
Click on the arrow next to Parent1 and in the scene view select the
furdude_main joint. Press Enter.
Now the heel joints are free from the COG and can be used to drive
the IK on the legs down the line. But all of the joints are parented to
the furdude_main joint except for the pelvis joint which is parented
to the COG joint.

03 Wire the parentjoints node into the rigpose node. Clear

any existing joints from this node then set the Display
flag on it. Now click on the furdude_main joint, the COG joint, the
pelvis joint then the two heel joints.
If you now set the Display flag on the bonedeform node everything
is mixed up because most of the joints are now missing. This can be
fixed with a skeleton blend.

04 Add a Skeleton Blend node into the Network editor.

Wire the skeletonmirror node into the left input and the
rigpose into the right input. Then wire the skeletonblend node into
The Main Controls

the third input of the bonedeform node.

In the skeletonblend Parameter pane, set the World Space checkbox
to On and weight1 to 1.

159
 CONTROL GEOMETRY
When you add a joint to a rigpose, it can be promoted to the top level of the
asset to select and animate. The Attach Control Geometry node lets you as-
sign geometry you build to different parts of the rig to make it easier to select
joints for manipulation. You can create any shape you want for these controls.
A good example where this will help is the eyeball and the two lid joints which
overlap. Control Geometry will make it easier to select these when you set up
that part of the rig. For now you will use it for the main controls.

05 From the Assets menu, select Save Asset > Fur Dude
Anim Rig. This saves the current setup. In the Network
view, Press 1 to navigate back to the test_rig where you can see
that only the five joints listed in the rigpose are visible.
You cannot select and move them because the parameters haven’t
been promoted yet.

06 In the Network view, press 2 to navigate back to the fur_

dude_anim_rig. Add a Circle node to the network. Set
Orientation to XZ. Set Uniform Scale to 0.2. Set Divisions to 36
and Arc Type to Open. Add a Color node and set Color to yellow.
Follow with a Merge Packed node and set Name 1 to circle_ctrl.
Place an Attach Joint Geometry node between the parent node and
the rigpose node. Wire the mergepacked node into the second input.
Set the Display Flag on this node and press Enter in the scene view.
Select all the visible joints then press G and use your scroll wheel to
find the circle_ctrl geometry. It gets assigned to all the joints.

07 In the Operation Control bar at the top, change the

Mode to Tweak Shapes.
Select the COG and Pelvis joints then press G to bring up a
transform handle. Press E to get the scale handle then click drag on
the middle handle to scale in all three directions until these controls
are a bit smaller (around 0.67 in the parameter pane).
Select the two heel joints then press G to bring up a transform
handle. Press E to get the scale handle then click drag on the
middle handle to scale in all three directions until the heel controls
are much smaller (around 0.3 in the parameter pane).

08
CHARACTER FOUNDATIONS

Select the rigpose node then select the four joints using
the control geometry. To make this work on the test rig
you need to promote parameters.
From the Assets menu, select Edit Asset Properties > Fur Dude
Anim Rig. Click on the Parameters tab.
On the rigpose node go to furdude_main, RMB-click on Scale and
choose Lock Parameter. Now drag the on Translate and move it
over to the Parameter list under root. Set it’s Label to Main Translate.
Repeat for the Rotate parameters and name them Main Rotate.
Click Accept to finish and save the results to the asset.

160
 09 In the Network view, press 1 to navigate back to the
test_rig. The test rig now updates to show the new
controls - as long as you have the Handle tool active. Select the
furdude_main joint using its control geometry and now a transform
handle appears that you can use to move the rig around.
Undo when you finish to put it back at the starting point..

10 In the Network view, press 2 to navigate back to the fur_

dude_anim_rig. From the Assets menu, select Edit Asset
Properties > Fur Dude Anim Rig. Drag the Translate and Rotate
for the COG, l_heel and r_heel joints and the Rotate for the pelvis.
In each case, lock the unused Scale or Translate (for the pelvis)
parameters for each joint.
Click Accept to finish and Save the results to the asset.

11 In the Network view, press 1 to navigate back to the test_

rig which has been updated to show the new controls.
Select the COG and heel joints using the control geometry and
transform the parts around.
Undo when you finish to put all the parts back to their original
location.

12 In the Network view, press 2 to navigate back to the

fur_dude_anim_rig. Select the nodes you used to set
up the main controls and click on the Add Network box button.
Position the box and center the nodes. Click on the box’s title and
enter Main Controls.
From the Assets menu, select Save Asset > Fur Dude Anim Rig. This
has no effect on the rig but keeps the asset up-to-date. You may
want to Save your Scene file too.

ORGANIZING YOUR NETWORK

THE MAIN CONTROLS

Lining up nodes and adding network boxes are extra steps that are worth the
added effort. The more organized your network is, the easier it is for you to
work with later or for others to read what your intentions are.
You can also add comments to each node and display them in the network or
you can use sticky notes to explain larger blocks of nodes. Communication is
always beneficial when creating networks in a team setting.

161
 PART TWELVE
Inverse Kinematics for the Legs
To animate a character, Inverse Kinematics allow you to set up the legs so that moving either the feet or the hips causes the
knee to bend appropriately. You are again going to pull out some joints from the main skeleton and set them up using KineFX.
You will again blend the results back into the original hierarchy.

01 Rename the skeletonblend node to skeletonblend_controls.

You are going to use these nodes a lot and it will be
helpful to be able to identify them.
Branch off a Delete joints node from the skeletonblend_controls
node and set its Display flag. Click on the arrow beside Group and
in the scene view select the left and right hip, knee and ankle joints.
Press Enter then set Operation to Delete Non-Selected.

02 Branch a Parent Joints node and set its Display flag.

Click the + sign to add a Joint and set Joint1 to *. Leave
Parent1 left blank. Now all of the joints are unconnected. This will
leave you free to use them independently.

03 Place an IK Chains node into the Network editor. Feed

the skeletonblend_controls into the first input and the
parentjoints node into the second one. Set its Display Flag.
In the Parameter pane, click on the + sign and in the folder click on
the arrow next to Root name. Select the l_hip joint and press Enter
(with your cursor over the Scene View). Set the Mid Name to l_knee
and the Tip Name to l_ankle. You can use the arrow to select the
joints or just type in the name. Set Match by Name to On and
Blend to 1 then set Orient Tip to On.
Click the + sign again and do the same for the right leg.

04
CHARACTER FOUNDATIONS

Add a Rig Pose node in-between the parent node and the
ikchains node. Click on the ankle joints and move them
around to see the Inverse Kinematics at work. Because you used
Match by Name the ankle joints are acting as the end effectors for
the IK chain.
You will see some flipping as you move the ankles. This is because
the knee joints are being used as the twist effectors and they are
not positioned very well - you need to move them in front of the
legs.

162
 05 Add another Rig Pose node in-between the first rigpose
node and the ikchains node. Press the Shift key and select
both the knee joints and move them in front of the character. Now
any flipping you had in your IK will have flipped back.
Name the first rigpose node to rigpose_ankles and the second one
to knee_offset.

06 You can now go back to the rig_ankles node and test out
the ankle joints. They don’t flip now that you have the
knees offset. You can also go to the knee_offset node and move the
knees to act as a twist effector for the IK chain.

07 Add Skeleton Blend node and wire the skeletonblend_

controls node into the first input and the ikchains into
the second input. Then plug the output of the new Skeleton Blend
node into the third input of the bone deform.
Rename the node to skeletonblend_ik and set the World Space
checkbox to On and weight1 to 1.

08
surface.
You can now go back to the rigpose_ankles node and test
out the ankle joints to see how it affects the captured

At this point, you are not going to save the digital asset to push
these changes to the test rig. The rigpose_ankles node will not be
how you control the ankles in the final rig. You are now going to
build a reverse foot setup that will allow you to drive the whole foot
setup and then the whole leg setup will be saved to the asset.
Inverse Kinematics for the Legs

FK/IK BLENDING
There are a couple of ways to blend IK and FK but you wont’ be using
them in this lesson because you will only be using IK for the legs. To get
blending to work, you would need to promote parameters for the leg
joints to the asset using a rigpose then you could either use the Blend
attribute on the IK Chains node or use a Skeleton Blend between the
IK solution and the rotating joints in the rigpose. If Fur Dude had arms
then it would make more sense to set this up.

163
 PART THIRTEEN
Reverse Foot Setup
To control the feet, you will create a classic reverse foot setup where the heel becomes the root then the toe, ball and ankle are
parented to that. This can be easily accomplished in KineFX and the results blended back into the original skeleton. In this case
you will completely rebuild the right foot but since the joint names align everything works properly.

01 Delete the rigpose_ankles node. You will control the

ankles using a reverse foot setup.
Branch off a Delete joints node from the skeletonmirror node and
set its Display flag. Click on the arrow beside Group and in the
scene view select the l_ankle, l_ball, l_toe and l_heel joints. Press
Enter then set Operation to Delete Non-Selected.
Note: If you want redirect the connecting wire to go around the
network box then you can Alt-click on it to add dots.

02 Add a Parent Joints node and set its Display flag. Click
the + sign to add a Joint and set Joint1 to *. Leave
Parent1 left blank.
Branch a second Parent Joints node and set its Display flag. In
the Scene View click on the l_heel joint then the l_toe joint then
the l_ball joint then the l_ankle joint. MMB-click to finish. In the
Parameter pane they will be displayed in the following order:
ƒ Joint1 to @name=l_toe | Parent1 to @name=l_heel
ƒ Joint2 to @name=l_ball | Parent2 to @name=l_toe
ƒ Joint3 to @name=l_ankle | Parent3 to @name=l_ball

03 Add a Skeleton Mirror node to the chain. Go to the

Parameter pane and under naming, set Find Tokens to l_
and Replace Tokens with r_. This creates us the reverse foot for the
right leg.
Because all the joints in this hierarchy have the same names as the
original skeleton it will transfer the information properly when this
rig is posed.

04
CHARACTER FOUNDATIONS

Now add a Skeleton Blend between the deletejoints and

parentjoints that you set up earlier for the leg. Rename
the node to skeletonblend_reversefoot and set the World Space
checkbox to On and weight1 to 1. Click on the arrow next to Group
and select the two ankle joints. Press Enter.
Add a Rig Pose node after the skeletonmirror. Rename it rigpose_
foot. Feed it into the second input of the skeletonblend_reversefoot
node.

164
 05 Set the Display flag on the ikchains node. Now go to
the rigpose_foot node and select the l_heel joint. Now
when you move it you are moving the whole foot and the leg chain.
Select the l_toe joint and rotate it. Again the reverse foot works and
the IK chain is activated.
When you finish press Clear to remove all the joints. You will add
some of them back later. For the heel joints, you want to use the
heel joints you set up earlier as part of the main controls.

06 Now add a Skeleton Blend between the reverse foot

skeletonmirror and rigpose node you were just using.
Rename the node to skeletonblend_heels and set the World Space
checkbox to On and weight1 to 1. Now feed the output of the
skeletonblend_controls node into the second input.
Click on the arrow next to Group and select the two heel joints.
Press Enter. Now you can select the rigpose for the controls and
move the left or right heel to control the whole setup. You can also
grab the COG joint and if you move it up and down the IK chain
works properly.

07 Put an Attach Joint Geometry node between the

skeletonblend_heels node and the rigpose_foot node. Wire
the mergepacked node from the Main Controls network box into
the second input. Set the Display Flag on this node and press Enter
in the scene view.
Select the toe and ball joints then press G and use your scroll wheel
to find the circle_ctrl geometry. In the Operation Control bar at the
top, change the Mode to Tweak Shapes. Select the toe and ball
joints then press G. Press E to get the Scale handle then click drag
on the middle handle to scale the controls to around 0.3.

08 Set the Display flag on the bonedeform node. Pose the

l_ball using the rigpose_foot node. You can see that rolling
the ball causes the toe to point down instead of bending.

REVERSE FOOT SOP

REVERSE FOOT SETUP

There is a Reverse Foot SOP that you could use to control the
feet and drive the leg kinematics. The Reverse Foot SOP has
sliders for controlling the roll of the foot an individual controls
for all the parts. But you are not using it in this lesson.
Instead you will create a hand-built reverse foot solution to learn
how joints can be manipulated to give you the control you need.

165
 09 Now add a Skeleton Blend between the ikchains and
skeletonblend_ik node you were just using. Rename the
node to skeletonblend_toes and set the World Space checkbox to
On and weight1 to 1. Now feed the output of the rigpose_foot node
into the second input. Click on the arrow next to Group and select
the toe and ball joints. Press Enter.
Now the toe will point in the right direction if you roll the ball joint.

10 Select the rigpose_feet node. Make sure only the left and
right ball and toe joints are listed on this node. For all of
them, RMB-click > Lock Parameter on the Translate, Rotate and
Scale parameters. Then go to the Rotate Y for each of them and
RMB-click > Unlock Parameter.
Now if you select any of these four joints you will only get a Rotate
Y handle.

11 Go back to the Main Controls Network box. Add a Box

node to the network. Set its Uniform Scale to 0.02.
Follow this with a Color node and set it to a red color. Feed this
node into the mergepacked node and set Name 2 to box_ctrl.
Put an Attach Joint Geometry node between the parentjoints
node and the knee_offset node. Wire the mergepacked node into
the second input. Set Mode to Tweak Shapes then select the
knee joints then press G and use your scroll wheel to find the box
geometry.

12 Go to the knee_offset rigpose node. Make sure that only

the two knee joints are listed. For both of them, RMB-
click > Lock Parameter on the Rotate and Scale parameters. You
will control these joints by moving them around.
Set the Translate values to -0.15, 0, -0.05 for both knees.
Add a Network Box around all of these nodes used to define the
spine and head joints and name it Leg Controls.
CHARACTER FOUNDATIONS

THE ROLE OF RIG POSE

Up to this point the Rig Pose node has been used to test out the rig. When
building an animation control rig, this node is also used to set up the param-
eters that will be promoted to the top level and define which joints will be
visible when working with the digital asset.
As you set up these nodes it is easy to accidentally add parameters that you
don’t need which will add extra joints at the top level. You can also make the
mistake of deleting parameters that you do need by clicking on the x button.

166
PART FOURTEEN
Promote the Leg and Spine Controls
To make all of the leg and spine controls available to the animator, the parameters need to be promoted to the top level of the
asset. This is an important step that is always needed to give the animator the control they need. This also means that you can
keep certain parameters hidden that you don’t want animators to work with.

01 From the Assets menu, select Edit Asset Properties > Fur
Dude Anim Rig. Click on the Parameters tab.
Create four folders. Name them Head, Body, Legs and Main. Drag
the COG parameters onto the Body folder and the Main parameters
onto the Main folder.
Put two folders inside the Legs folder and name the Left and Right.
Add the Left Heel parameters into the Left folder and the Right Heel
parameters into the Right Folder.

02 From the knee_offset rig pose node, drag the l_knee

translate values into the Left folder in the Leg folder.
Rename it Left Knee Offset. Click on the Channel tab and set the
default to -0.15, 0, -0.1.
From the rigpose_foot node, drag the Rotate Y from the l_ball joint,
name it Left Ball Rotate and set its Range to 0 to 30. Rotate Y from
the l_toe joint, name it Left Toe Rotate and set its Range to 0 to
20. Drag a Separator in between the Knee Offset and Ball Rotate
parameters.
Repeat for the right knee and the right foot.

03 Click Accept and now the new parameters and controls

are saved into the asset. Go up one level to see the asset
parameters laid out on the furdude_rig.
These parameters are for the unlocked rig that you are working on.
Rather than test out these parameters, it is a better idea to play

PROMOTE THE LEG AND SPINE CONTROLS

around with the test_rig.

04 In the Network view, press 1 to navigate back to the test_

rig which has been updated to show the new controls.
Play with the test rig in the Scene view using the handles or using
the parameters in the floating panel.

167
 05 Branch off a Delete joints node from the skeletonblend_ik
node and set its Display flag. Rename the node
deletejoints_spine.
Click on the arrow beside Group and in the scene view select the
spine1, spine2, spine3, neck1, neck2 and jaw joints. Press Enter then
set Operation to Delete Non-Selected.

06 Add a Rig Pose node and rename it rigpose_spine. Now

add a Skeleton Blend between the skeletonblend_ik and
bonedeform node. Rename the node to skeletonblend_spine and set
the World Space checkbox to On and weight1 to 1. Now feed the
output of the rigpose_spine node into the second input.
Add a Network Box around all of these nodes used to define the
spine and head joints and name it Spine Controls.

07 Set its Display Flag then in the Scene view press and hold
the S key then select all the joints. This will add them to
the rigpose list.
Lock the Translate and Scale parameters for all of these joints. You
will only be rotating them.

08 From the Assets menu, select Edit Asset Properties > Fur
Dude Anim Rig. Click on the Parameters tab.
Drag the neck1, neck2 and jaw onto the Head folder and name
them Neck 1 Rotate, Neck 2 Rotate and Jaw Rotate. spine1, spine2
and spine3 onto the Body folder and name them Spine 1, Spine 2
and Spine 3. Add a separator between the COG parameters and the
spine parameters.
Click Accept. This will save the new controls to the rig. You can now
explore them using the test_rig.
CHARACTER FOUNDATIONS

SPINE CONTROLS
For this rig, you set up the spine using joint rotations. This is known as
Forward Kinematics and just like other parts of the rig, you pulled these
joints out from the main skeleton then set them up in a Rig Pose to be
promoted to the top level.
You did not use control geometry for these parts because the joints are
easy to select in the Scene view. You don’t need to have control geo for all
the joints.

168
PART FIFTEEN
Eye Controls
The next step is to set up the eyelids with control geometry to make it easier to select these overlapping joints. You will also
set up the eye target joint as a look at for the eyeball using a different section of Houdini called VOPS. When these parts are
rigged, you will again promote the appropriate parameters to the character’s asset.

01 Branch off a Delete joints node from the skeletonblend_

spine node, set its Display flag and name it deletejoints_
eyelids. Click on the arrow beside Group and in the Rig Tree select
upper_lid and lower_lid. Move your cursor over the Scene View and
press Enter then set Operation to Delete Non-Selected.
Add a Rig Pose node and rename it rigpose_eyelids. Set its Display
Flag then in the Scene view press and hold the S key then select all
the joints. This will add them to the rigpose list.

02 Now add a Skeleton Blend between the skeletonblend_

spine and bonedeform node. Rename the node to
skeletonblend_eyelids and set the World Space checkbox to On and
weight1 to 1. Now feed the output of the rigpose_eyelids node into
the second input.

03 Place an Attach Joint Geometry node between the

deletejoints node and the rigpose node. Wire the
mergepacked node into the second input. Set the Display Flag on
this node and press Enter in the scene view. Make sure the Mode
is set to Assign Shapes. Select all the two eyelid joints then press G
and use your scroll wheel to find the circle_ctrl geometry.

04 In the Operation Control bar at the top, change the

Mode to Tweak Shapes.
Select the eyelid joints then press G to bring up a transform handle.
Press E to get the scale handle then click drag on the middle handle
to scale in all three directions until these controls are a bit smaller
EYE CONTROLS

(around 0.5 in the parameter pane).

Now select the upper_eyelid joint and move it up by 0.02 then
select the lower_eyelid joint and move it down by 0.02.

169
 05 From the Assets menu, select Edit Asset Properties > Fur
Dude Anim Rig. Click on the Parameters tab.
Lock the Translate, Rotate and Scale parameters for all of these
joints. Unlock the Rotate X parameters. From the rigpose_eyelids
node, drag the Rotate X from the upper_lid to the Head folder. Set
the Range from -10 to 30. Next, drag the Rotate X from the lower_
lid to the Head folder. Set the Range from -20 to 20.
Add a Separator to divide the eye controls from the head controls.
Click Accept. This will save the new controls to the rig. You can now
explore them using the test_rig.

06 Branch off a Delete joints node from the skeletonblend_

eyelids node, set its Display flag and name it deletejoints_
eyes. Click on the arrow beside Group and in the scene view select
the eyeball and eye_target joints. Press Enter then set Operation to
Delete Non-Selected.
Add a Rig Pose node and rename it rigpose_eyetarget. Set its
Display Flag then in the Scene view select the eye_target joint. This
will add it to the rigpose list. You do not need the eyeball joint - it
will be controlled by a look at constraint. Lock the Rotate, and Scale
parameters for the eye_target joint.

07 Add a Rig Attribute VOP node. Wire the deletejoints_eyes

node into the first input and the rigpose_eyetarget into
the second one.
Now add a Skeleton Blend between the skeletonblend_eyelids and
bonedeform node. Rename the node to skeletonblend_eyeball and
set the World Space checkbox to On and weight1 to 1. Now feed
the output of the rigattributevop node into the second input.
Set the Display Flag on the bonedeform node.

08 Double click on the rigattributevop node to dive into it.

In the Scene view, click on the eyeball joint then drag
the eyeball (deletejoints) version into the Network editor. This gives
you a Get Point Transform node that focuses on the eyeball joint
coming from the First Input.
Click on the eye_target joint then drag the eyetarget (ripose_
eyetarget) version into the Network editor. This gives you a Get
Point Transform node that focuses on the eye-target joint coming
from the Second Input.

09
CHARACTER FOUNDATIONS

Click on the eyeball joint then drag the eyeball version

into the Network editor. This gives you a Set Point
Transform node that focuses on the eyeball joint.

170
 10 Now press tab > Look At (KineFX) and place the node
in the middle. Wire the xform output on the eyeball
getpointtransform node into the from input on the lookat node. Wire
the xform output on the eye_target getpointtransform node into
the to input on the lookat node. Wire the outxform output on the
lookat node into the xform input on the eyeball_set node.
The eyeball geometry flips. Select the lookat node and set the Look
At Axis to Z to match the orientation you set on the rig when it was
first set up.

11 Put an Attach Control Geometry node between the

deletejoints_eyes node and the rigpose_eyetarget node.
Wire the mergepacked node from earlier in the network into the
second input. Set the Display flag on this node and go to the
Handle tool. Make sure the Mode is set to Assign Shapes. Select
the eye_target joint in the 3D view then press G and use your scroll
wheel to select the square_ctrl.
Set the Display Flag on the bonedeform node. Click on rigpose_
target. Select and move the eye_target joint. This orients the eyeball.
Undo to set it back to it’s original position.

12 Add a Network Box around all of the nodes used to

define the eyes and name it Eye Controls.
From the Assets menu, select Edit Asset Properties > Fur Dude
Anim Rig. Click on the Parameters tab.
From the rigpose_eyetarget node, drag the Translate parameters
from the eye_target to the Head folder in the eye section. Name
them Eye Target Position.
Click Accept. This will save the new controls to the rig.

13 You can now explore them using the test_rig.

You now have all of the parts finished for this control
rig. It is now ready for you to animate a walk cycle. To do this you
will create a second copy of the test rig and animate using that
network.
This digital asset can create multiple instances in multiple scene
files and if you later need to come back and make a change, all of
the assets will update. This is the pipeline advantage of working
with a digital asset rig.

RIGGING IN VOPS
The Rig Attribute VOP offers a range of different solutions beyond what you have seen in
this lesson. IK chains can be built using this approach and the IK chain SOP you used earlier
has one of these inside it.
EYE CONTROLS

You can also use VOPS to set up a curve solver, realistic shoulder, reverse foot and more.
The ability to drag joints from the Scene view to the VOP network is a unique workflow that
helps speed up workflow.

171
 PART SIXTEEN
Animate the Rig
It is time to keyframe a walk cycle for the fur dude character. This will involve new tools such as the Channel List to pin down
channels for blocking out the motion. The results will be a quick and dirty walk cycle designed to see fur dude in action. The
goal is to map out a basic keyframe workflow to learn a bit about how to animate KineFX rigs.

01 Go to the object level. Press tab > Geometry and place

down the node. Rename it walkcycle. Turn off the Display
Flags on all of the objects.
Double-click to dive into walkcycle and in the Network view, press
tab > Fur Dude Anim Rig. Press Enter to place it at the origin. This
is a new locked version of the fur dude rig that you will animate
from scratch.
This puts another version of the character asset into the scene. You
can have multiple versions in this scene file or in other scene files
and they will all reference the same asset definition on disk.

02 From the Desktop menu (the one that currently says

Build), choose Animate. This gives you panes that are
designed to work with a keyframe workflow. You may need to go
back into the walkcycle object.
The Channel List on the left will play a key role in blocking out the
animation of the character. The Animation Editor lets you display
and edit animation curves. In this lesson you will block out the
motion and won’t be doing any curve editing.

03 In the Network view, select the fur_dude_anim_rig inside

the walkcycle object. Go to the Parameter pane and
click on the box icon in the top right. Choose Parameters and
Channels > Create Nested Channel Groups. Click Close in the pop
up window. The parameters from your asset are now listed and
organized based on their folders.
Click on the Pin icon next to the fur_dude_anim_rig channel group
to pin these channels. Make sure your timeline is set to frame 1
then press k to keyframe all the channels.
CHARACTER FOUNDATIONS

HOW CHANNELS WORK

When you select a joint, the channels load into the channel list. Press k to keyframe them.
If you want to keep them loaded in the list for blocking purposes, you can pin them or add
them to a channel groups to pin them all together. You can build the Channel groups directly
from the asset. They are organized based on how you built the UI for your character. You can
also build your own groups to pin down specific channels.

172
04 Set the Timeline to start at 10 and end at 50. Go to
frame 10. With all the channels pinned, press k to set
another keyframe. You want to keyframe first then pose. Any
posing will update the value for that keyed frame.
You want to create a pose where the left heel moves forward
and rotates up. The COG will go down a bit and the right ball rolls
forward.

05 Go to frame 15 and press k.

Roll the left heel flat but don’t move it. Move the COG to
align with the left foot. Move the right heel up in line with the other
foot. Rotate the right ball back to flatten the foot.
Rotate the COG a bit towards the left foot. You can also add some
rotation to the three spine joints to emphasize this tilt.

06 Go to frame 20 and press k.

Now you want to create a pose where the right heel
moves forward and rotates up. The COG will go down a bit and the
left ball rolls forward. This is the opposite of the pose at frame 10.
Rotate the COG and spine joints back to center them.

07 Go to frame 25 and press k.

Roll the right heel flat but don’t move it. Move the COG to
align with the right foot. Move the left heel up in line with the other
foot. Rotate the left ball back to flatten the foot.
Rotate the COG a bit towards the right foot. You can also add some
rotation to the three spine joints to emphasize this tilt.

08 Go to frame 30 and press k.

You want to create a pose where the left heel moves
forward and rotates up. The COG will go down a bit and the right
ball rolls forward.
Animate the Rig

Rotate the COG and spine joints back to center them.

173
 09 Continue this pattern up until frame 50. You can repeat
the same poses to keep the walk cycle moving forward.
At this point you can go back and tweak any pose to refine the
motion. You can also explore creating some overlapping action
using extra keyframes. You can animate some eye movement and
maybe the eyelids blinking. You can also go beyond 50 frames if
you would like a longer animation.

10 Connect the output of the fur_dude_anim_rig node into a

ROP Geometry node. This will allow you to export out a
cache of the fur dude geometry. Set the Output File to
$HIP/geo/furdude_walk.$F.bgeo.sc
Next Set Valid Frame Range to Render Frame Range. RMB-click on
the Start/End/Inc parameter and choose Delete Channels. Set the
Start to 1 and the End to 50.
Click the Save to Disk button to store the cache to disk. You will
use this to add fur.

11 Branch an Attribute Delete node off of the fur_dude_

anim_rig node. Under Primitive Attributes choose Cd.
This removes the color from all the body parts.

12 Connect the output of the attributedelete node into a

USD Export node. This will allow you to export out a the
fur dude geometry to the USD format. Set the Output File to
$HIP/usd/furdude_walk.usd
Next Set Valid Frame Range to Render Frame Range. RMB-click on
the Start/End/Inc parameter and choose Delete Channels. Set the
Start to 1 and the End to 50.
Click the Save to Disk button to store the cache out the USD file.
You will use this later in the rendering process.
CHARACTER FOUNDATIONS

CACHING OUT ANIMATION

Because of Houdini’s procedural nature it is not really necessary to cache out the anima-
tion. You could reference it from this network to another network for grooming or to
Solaris for conversion to USD. The advantage of caching is that you lock down your ani-
mation and work with a flattened file on disk. This is a very production-friendly approach.
In Solaris a USD file referenced from disk is also more efficient. Since Houdini
references files from disk, you are always free to change your animation and
output the new sequence and it will be picked up automatically.

174
PART SEVENTEEN
Add & Groom the Fur
Fur dude gets his name for a reason and you are going to work with a variety of grooming tools to add and shape the hair. Using
a desktop designed for grooming, you will add frizz, clumping and hair dynamics which will simulate as fur dude walks. The end
result will be ready to be exported for rendering.

01 Got to the Grooming desktop. Put the four existing

objects into a network box and call it box Rig & Animate.
In the Network editor, press tab > File. Place the node then double
click to go into it. Click on the browse button next to Geometry File
and go to $HIP/geo. Select furdude_walk.$F.bgeo.sc. Press Accept.
Add a Blast node, set it’s Group to fur_dude_body and turn On
the Delete Non-Selected checkbox to focus on the body. Set the
Display Flag then go to the Object level and rename this fd_anim.
Alt drag to create a copy it and name it fd_rest. Dive in and change
Geometry File to $HIP/geo/ furdude_walk.1.bgeo.sc.

02 Now move the time slide forward a bit. One object has a
static version of fur dude and the other is animated. Click
the Add Fur button . Select the fd_rest object and press Enter.
Now select the fd_anim object and press Enter. Turn off the Display
Flag on fd_rest_anim, fd_rest_deform and fd_rest_hairgen.
Turn on the Display flag for fd_rest and fd_rest_groom.

03 Select the fd_rest_groom node and from the Hair Tools

shelf, click on Set Guide Length. Turn the Randomize
Button to On. Set Min Length to 0.03 and from the menu on the
right select Texture.
Use the file browser button to select $HIP then go into the tex
directory and choose fur_length.jpg.
Now set Max Length to 0.15 and again choose Texture. You can
use the arrow on the right to select the fur_length.jpg image.
You now have the eye, lip and the bottom of the feet masked out
and the rest of the fur with random lengths.

04 From the Hair Tools shelf, click on Bend Guides. Set

Add & Groom the Fur

Angle to 45 to add some bending to the guides.

Next, click on the Frizz Guides tool. Set the following:
ƒ Frequency to 15
ƒ Amplitude to 0.005
ƒ Random Amplitude to 0.02
This keeps the hairs from looking too straight when rendered. You
can add more frizz if you want a more tangled look.

175
 05 From the Hair Tools shelf, click on Clump Guides. Set
Clump Size to 0.02 to Tightness to 0.5.
Next change the Clump Profile so that it stays high for a bit longer
then tapers off near the end.
Now go to the Object level and turn off the display on the fd_rest
and the fd_rest_groom and turn on the display on the fd_anim and
fd_rest_sim and fd_rest_hairgen.
Select the fd_rest_groom node and set Density to 20000.

06 Select the fd_rest deform node and from the Hair Tools
shelf, click on Simulate Guides. On the fd_rest_sim node,
go to the Vellum Constraints tab and under Bend set Stiffness to 5.
Alt-drag on the fd_anim node to make a copy and rename it
fd_collision. Dive into this node and on the blast node set Delete
Non Selected to Off. Add the tongue, upper teeth and gums to the
Group selection then add a Null node and name it COLLISION_OUT.
At the object level, select fd_rest_sim then under Vellum Collisions,
set External Collisons to On then set Collider SOP to ../fd_collision/
COLLISION_OUT.

07 Click on the Caching tab and set Set Valid Frame Range
to Save Frame Range. RMB-click on the Start/End/Inc
parameter and choose Delete Channels. Set the Start to 1 and the
End to 50. Click Save to Disk to run the simulation.
Now set the Load from Disk checkbox to On. The cache will now
be used to define the fur instead of the hair calculating for each
frame.

08 Select the fd_rest_hairgen node. Under Distribution, set

Density to 1000000. Scroll down to Guide Interpolation
and set Clump Crossover to 0.25 to create a it of overlap between
clumps. This gives you an idea of what furdude looks like with a full
head of hair.
These are not the hairs that you will render in Houdini’s lighting
context called Solaris. You will instead bring in the guide hairs and
render using a hair procedural at render time.
CHARACTER FOUNDATIONS

HAIR BRUSHES
The grooming desktop also has hair brush tools that let you work interactively
on the character’s surface. You can lengthen, smooth, cut and extend hair.
These tools were not needed to set up fur dude’s grooming but later you may
want to explore them to further style the final look.

176
PART EIGHTEEN
Set up an Render the Shot
To render the shot, you will reference the USD files into the Solaris Stage then add a backdrop. Solaris is a Houdini context that
uses LOP nodes to set up a USD Scene Graph. Next, you will import the fur then add and position a camera and a light. The
Karma renderer will then be invoked to create a preview render of the shot then render out the animated sequence.

01 Change the desktop to Solaris. Choose Stage from the

path bar. In the Network View press tab > Reference
then click to add a Reference node.
Next to Reference File, click on the File Pattern and find the
furdude_walk.usd file. Rename the node to furdude. Set the
Primitive Path to /char/`@sourcename` - this will use the node
name and place it into a group called char. In the Scene Graph Tree
expand char then furdude to see all of the named primitives.
In the Scene View, use your view tools such as spacebar-h for
homing the view to get a better look at the walk cycle.

02 Press tab > Material Library. Wire it into the output of

the reference node then set its Display Flag.
Go to the Material Palette pane. Click on the arrow next to /stage/
materiallibrary to open up this area. Scroll through the material
gallery on the left of the palette and drag a Principled Shader
materials into the materiallibrary working area.
Go to the Network view and Alt-drag this material to create four
more. Rename the five materials body_mat, eyeball_mat, eyelid_mat,
teeth_mat, and tongue_mat. You can also see the materials in the
Scene Graph Tree.

03 For furdude_body_mat, under the Surface tab, set Base

Color to 1, 1, 1. Click on the Textures tab and under Base
Color click on Use Texture then use the button next to Texture to
call up the file window. Click on $HIP in the side list then click on the
tex folder to open it and then click once on skin_color.jpg to select
it. Click Accept to assign the texture to the material. Next set the
Roughness to 0.5 and Reflectivity to 0.
Assign eye_color.jpg and eye_lid.jpg to their materials using the
same method. Set tongue_mat to a redish pink and teeth_mat to a
yellowish-white.
Set up an Render the Shot

04 Go to the Stage level. After the Material Library node, add

an Assign Materials node. From the Scene Graph, drag
the fur_dude_body to the Primitives field then click on the arrow
next to Material Path and choose body_mat. Now click the Plus
Sign next to add four new entries. Assign them as follows:
ƒ fur_dude_eye > eyeball_mat
ƒ fur_dude_lowid/uplid > eyelid_mat
ƒ fur_dude_lowteeth/upteeth/claws > teeth_mat
ƒ fur_dude_tongue/gums > tongue_mat

177
 USD SCENE GRAPH
When working in Solaris, the geometry and materials you add using
LOP nodes are added into the Scene Graph and converted to USD.
When you add a light and a camera, they will also become part of the
USD Scene Graph.
It is not necessary to completely understand USD to light and render
in Houdini as an artist but once you start thinking about the pipeline of
your project, USD will become a useful tool in managing your shots.

05 In the Network view, press tab and type out SOP Import.
Click to place the node. Rename it hair. Set Import Path
Prefix to /char/$OS. Click on the node icon net to SOP Path and
navigate to the fd_rest_hairgen node.
Add a Merge node in-between the furdude and materiallibrary
nodes. Wire the hair node into it.

06 Go back to the Material Palette pane. Open up /

stage/materiallibrary. Drag the Hair material into the
materiallibrary working area.
Leave the Root Color and Tip Color set to the defaults for the hair.
Next click on the Secondary Reflection tab and set Root Color to
dark grey and Tip Color to medium grey.
Go back to the Assign Material node, From the Scene Graph, click
on the arrow next to Primitives and select the fur curves. Next,
click on the arrow next to Material Path and choose hair.

07 In the Network view, press tab and type out Grid. Click
to place the node. Rename it backdrop and wire it into
merge. Set Import Path Prefix to /geo/$OS. Double-click on the
backdrop node to dive down to the geometry level.
Select the Grid node and set the Size to 50, 50 and Rows and
Columns to 10. RMB-click on the grid node’s output and type
Bend. Place the bend node and set its Display Flag then set: Bend
to 75, Capture Origin to 0, 0, -10, Capture Direction to 0, 0, -1, and
Capture Length to 10. RMB-click on the bend node’s output and
type Subdivide. Set its Display Flag then set Depth to 2.

08
CHARACTER FOUNDATIONS

Go to object level and set Rotate Y to -45 degrees. Add

material and assign it. You can leave the default grey or
add your own base color.
Set time range from 10-50. You will be rendering this sequence
after the hair has had 10 frames to settle down.

178
 09 Use your view tools to look at furdude from the front.
From the LOP Lights and Camera shelf, Ctrl-click on the
Camera tool. This adds a camera node into the network and you are
now looking through the camera in the Scene View.
Press the Lock Camera/Light to View button so that view changes
can be used to reposition the camera. Now Tumble, Pan and Dolly
in the Scene View to reposition the camera so furdude starts on the
left and the moves to the right. Scrub the timeline to make sure the
camera works for the whole sequence.

10 Now turn Off the Lock Camera/Light to View button

then tumble around until you are looking down on Fur
Dude. From the LOP Lights and Camera shelf, Ctrl-click on the
Area Light tool. This adds an arealight node to the end of the chain.
Select the arealight node and from the Base Properties tab, set the
Intensity to 2.

11 From the Persp menu, choose Karma to render with

Karma in the Scene View. You can move to different
frames in the timeline and the Scene View will update quickly.
Turn ON Karma is designed to work with USD which is why everything in the
Denoiser LOP context is converted to the USD scene graph. You can only use
the Karma renderer from this part of Houdini.
To get a cleaner image when you render, you can turn on the
Denoiser if you have an Nvidia graphics card. Be sure to install the
Denoiser from the Render menu and then turn it on in the Display
Options bar.

12 Press tab > Karma to add a Karma Render Settings and

USD Render ROP node. Wire them into the end of the
chain. Select karmarendersettings and on the Image Output > Filters
tab set Denoiser to nvidia Optix Denoiser. Set the Output Picture
to $HIP/render/walk/furdude_walk_$F2.exr. The $F in the name is
needed to add frame numbers to the renderings and the 2 is the
padding of the frame number.
Select usdrender_rop. RMB-click on the Start/End/Inc parameters
and choose Delete Channels. Set the Start to 10 and the End to 50
Select the usdrender_rop node. Click on Render to Disk.
Set up an Render the Shot

KARMA RENDERER
You are now going to render the sequence using Houdini’s renderer Karma. Karma is
designed to render USD and is known as a Render Delegate. At first you will render in the
Scene View. Press d in the Scene View to bring up display options to control your rendering.
You can turn on a denoiser, set Pixel Samples and define the Image Resolution.
Later when you set up a Karma LOP, there will be render settings on that node that you will
use to create the final output being saved to disk.

179
 13
sequence.
When you finish, choose Render > Mplay > Load Disk
Files and open up the rendered images to review the final

Later you can branch off another Karma node to up the resolution
and render settings for your final rendering. It is always good to
complete test renderings at a lower resolution first to make sure
that everything is working the way you expect it to.

14 If you want to tweak the hair and fur settings, you can pin
the Scene View to the LOP network then go back to the
object level and from the Simulate Guides node, set the Load from
Disk checkbox to Off. This will allow any changes to flow through
to the final rendering.
Here you can see that the hair was shortened and made much
frizzier. You can do anything you want. When you are finished be
sure to re-cache and then turn Load from Disk back to On.

15 You can then go back to the Solaris network and maybe

change the colors on the fur and re-render.
For the final render you may want to up the resolution to
1920x1080 and change some of the quality settings. For instance
you can set Pixel Samples to 128 and Light Sampling Quality to 16.

CONCLUSION
You have now rigged, animated and rendered the Fur Dude char-
acter using the KineFX tools in Houdini. Along the way you have
touched on various important steps including the creation of a
capture rig then the layering of a animation control rig on top
of that. You have packed up the character into a Houdini Digital
Asset and then keyframed a traditional walk cycle.
You then added fur with various grooming tools then rendered
using Karma. This gives you a complete workflow for creating
a shot using this character. You can then go back and refine any
CHARACTER FOUNDATIONS

aspect of the procedure to create multiple iterations as you seek

out the perfect result.

As mentioned before, the KineFX toolset is currently designed for

retargeting and motion editing which are not covered in this lesson.
These rigging and animation tools will continue to evolve and this
lesson provides a taste for what is coming in Houdini’s procedural
rigging workflow in future versions.

180
HOUDINI FOUNDATIONS

PROCEDURAL ASSETS
FOR UNREAL
To create game assets using Houdini’s node-based workflow, it is important to start learning how to think
and work procedurally. In this lesson, you will learn how to create game assets using procedural nodes and
networks then deploy them directly into Unreal using the Houdini Engine.
Along the way, you will get to use different aspects of Houdini’s user interface. You will learn how the
different UI elements work together to support you as you build your game assets.
This lesson was completed using Unreal Engine 5. Note that while the lesson focuses on Unreal, you can
import the same assets into Unity using the Houdini Engine.

LESSON GOAL
To create assets that can be imported into Unreal as game art.

WHAT YOU WILL LEARN

ƒ How to work with Nodes and Networks to control the flow of data
ƒ How Houdini Digital Assets can be used to package and share your solution with others
ƒ How to load Houdini Digital Assets into the Unreal Editor
ƒ How to instance objects to points and control the orientation and scale of the objects using attributes
ƒ How to create a game asset with Collision Geometry for use in Unreal
ƒ How to export a Rigid Body Simulation as an FBX file for Unreal.

LESSON
COMPATIBILITY
Written for the features in Houdini 19.5+

The steps in this lesson can be

completed using the following
Houdini Products:

Houdini Core ✔
Houdini FX ✔
Houdini Indie ✔
Houdini Apprentice ✘
Houdini Education ✔

Document Version 3.0 | July 2022

© SideFX Software
181
 PART ONE
Create a Simple Building
Learn how to build a simple building using a procedural network of nodes. Some of the nodes
will be created by interacting in the Scene View and others in the Network view. You will use
channel references to connect parts of one node with other nodes in the system. This creates a
procedural solution that you will wrap up into a Houdini Digital Asset.

01 Select File > New Project. Call it hengine_lesson and

press Accept. Select File > Save As... Set the file name to
hengine_01.hip and click Accept to save.
In the viewport, press c and choose Create > Geometry > Grid.
Press Enter to place it the grid at the origin. In the Operation
Controls bar at the top of the Scene View, set:
ƒ Size to 5, 3
ƒ Rows to 4
ƒ Columns to 6
Press v and choose Shading > Smooth Wire Shaded.

02 Press c and choose Create > Geometry > Box. Press

Enter to place it at the origin. Double click on the box_
object in the Network view. This puts you inside the object at the
geometry level. Set the following:
ƒ Size to 0.1, 1, 0.1
Rename the box node to column. In the Scene view press tab
and start typing Match Size then select Match Size . Press n to
select the box and then press Enter. This adds a new node. In the
Parameter pane, under Matching, set Justify Y to Min. This raises
the box up so that it sits on the ground.

03 Press u to go back up to the object level or click on obj

on one of the Network Path bars. Click on the Select tool
and click in empty space to deselect all the objects.
On the Modify shelf, click on Copy to Points. Select the column as
the geometry to copy and press Enter then click on the grid as the
geometry on whose points to copy and press Enter.
In the Parameter pane, set Transform Using Implicit Target... to Off.
Now the columns will be pointing up. Turn on Pack and Instance
to ensure instanced geometry in Unreal. Press 4 to go to primitive
selection mode. This will hide all the corner points.

04 Press tab > PolyExtrude in the Network View. Place it to

HOUDINI FOUNDATIONS

the side. Wire the grid node into the polyextrude node
and set its Display flag. Set Distance to 0.1 and under Output
Geometry and Groups turn on Output Back. Set the Template Flag
on the copytopoints node.
Add a Match Size node to the network and wire polyextrude into
the first input and copytopoints into the second. Set Justify Y to
Max to Same, Offset by 0.1. Now the extruded shape will sit on top
of the columns.
Add a Merge node and wire copytopoints and matchsize into it.

182
 05 In the Network view, press tab > group. Wire it between
polyextrude and matchsize. Change Group Name to edges.
Alt drag to create a another group node and set its Display flag.
Keep Group Name set to edges and then set Initial Merge to
Subtract from Existing. Turn off Enable under Base Group and turn
on Enable under Keep by Normals. Set Direction to 0, 1, 0 and
Spread Angle to 0. Alt drag on this node to make a copy and wire it
into the chain. Change Direction to 0, -1, 0. Now only the sides of
the box are in the edges group.
Add a PolyExtrude node. Set Group to edges and Direction to 0.15.

06 Set the merge node’s Display Flag to see the first floor of
the building complete. Turn off the Template Flag on the
copytopoints node.
In the Scene view, press n to select all and then press tab > Copy
and Transform. Select the column node and RMB-click on Center Y
and choose Copy Parameter. Go to the copy node and RMB-click
on Translate Y and choose Paste Relative References. Add a + 0.1
to the expression.
Now increase Total Number to 4 to add three more floors.

07 Select all the nodes in the Network editor. From the

Asset menu, choose New Digital Asset from Selection.
This will collapse the network into a subnetwork then use that
subnet node to create the Digital Asset.
Set the Operator Name to building which will change the Operator
Label to Building. Click on the button to the far right of Save to
Library. In the Locations sidebar, click on $HIP/ and then double-
click on the hda directory. Press Accept and then Accept again to
save the asset to disk.

08 The Edit Type Properties window opens up. This panel is

where you will build the user interface for your asset. You
will revisit this window later. Click Accept to close this window.
Rename the node in the network view to building.
The nodes you have used to build the asset will remain a part of the
asset even after you save it. This will allow you to continue making
changes even after you have started using it in your Unreal game
levels.
CREATE A SIMPLE BUILDING

WHAT IS AN HDA FILE? .hda file

Houdini nodes and networks can be encapsulated into single nodes SAVE AS

called Houdini Digital Assets which let you share your techniques Houdini
Digital
with colleagues. These assets are saved to disk inside files known as Asset
.hda files.
Asset files created in older versions of Houdini may have a different
extension .otl which means Operator Type Library - both these file
types work the same way.

183
 PART TWO
Import the Asset into Unreal
Now that you have a digital asset file on disk, you can import it into the Unreal Engine. This is possible because of the
Houdini Engine plug-in which connects the two applications. Houdini Digital Assets that are loaded into Unreal are
cooked using Houdini under the surface. The Houdini Engine for Unreal plug-in must be installed to complete this les-
son (See Sidefx.com/unreal for instructions). Please note that Houdini Apprentice does not work with Engine.

01 IN UNREAL – Set up a Third Person Template with

Blueprint support. Delete the TextRenderActor. Open the
Content Drawer and click on the Import button. You may want to
dock the Content Drawer. Navigate to your Houdini project and
find the building asset. Click Open.
Set the Scale to 2, 2, 2 and move it into the corner. If you press the
Play button you will see the building during gameplay.

02 IN HOUDINI – To add normals and collisions to your

geometry you need to add some nodes to the network.
Dive into the building network and between the polyextrude and
matchsize nodes, add a Normal node. This will add normals for the
geometry to display properly in Unreal. Under the normal node add
a Group node and set the Group Name to rendered_collision_geo.
This will turn your geometry into collision geometry.
Copy and Paste these two nodes and insert the new nodes after
the column node.

03
the scene.
IN UNREAL – In the Generate section of the building
asset’s Details panel, press Rebuild. Press Play to explore

The normals are working properly and now you will collide with the
columns if you try to run through them.

04 IN HOUDINI – Select Assets > Edit Asset Properties

HOUDINI FOUNDATIONS

> Building. In the properties window, click on the

Parameters tab. Now select the column node and drag the Size Y
parameter to the Parameter tab. Rename it Floor Height. Now from
the copy node, drag the Total Number parameter over and rename
it Number of Floors. Click Accept.

184
05 IN UNREAL – Press Rebuild in the building asset’s
Details panel. Scroll down to see that there are now
parameters for Floor Height and Number of Floors.
Set Floor Height to 1.5 and Number of Floors to 2.
You can Play the level to walk around the asset and review the
changes.

06 IN HOUDINI – Go back and set the Display flag on the

floor slab’s polyextrude node. Press 4 to get primitive
selection. Select the middle three primitives from the top then
tumble around and shift-select the three primitives from the
bottom of the slab. Press Delete. This adds a Blast node.
Press 3 to get edge selection. Double click on the bottom edge of
the hole to select the whole loop. Go tab > Polybridge. Press Enter.
Now double click on the top edge of the hole then press Enter. This
will add geometry to the inside of the hole.

07 Set the Display Flag on the matchsize node. You will see
the new hole with an overhang.
Press 4 to get primitive selection. Get the Select tool then select
the end face of the opening. It may not appear selected but it
is. Press tab > PolyExtrude. Under Extrusion, turn on Transform
Extruded Front. Start pulling the face out and down.

08 Copy the column node’s Size Y channel and Paste

Relative Reference to the new polyextrude node’s
Translate Y. Add a - sign to the expression then subtract 0.1. The
expression should read: -ch(“../column/sizey”)-0.1
Set Translate Z to 2.7.
Move the normal and group nodes to after this polyextrude node.
Otherwise, the new extrusion won’t be collision geometry and the
ramp won’t work properly.
Set the Display Flag on the output node. Choose Assets > Save
Import the Asset into Unreal
Asset > Building to save the changes to the HDA file on disk.

09 IN UNREAL – Press Rebuild then set the Floor Height to

1 and the Number of Floors to 6. Press Play and walk up
ramps and around the building.

185
 UVS Show UV Texture
Houdini has a number of nodes at the geometry level for setting up and manag-
ing UVs. In this lesson, you will use UV Unwrap and UV Transform to add UVs to
the building. These work well with the simple shapes used to model the building.
For more complex shapes you will use tools such as UV Flatten and UV Layout.
You can see a UV Grid on your geometry once UVs are in place. To hide the grid,
you can click on the Show UV Texture button on the Display Options bar to
toggle it on and off.

10 IN HOUDINI – You need to add some UVs to the

geometry for use in the game editor. Go to the part of the
network with the column node. Press tab> UV Unwrap and place
the node between the column node and the matchsize node.
Go to the part of the network with the second polyextrude
node. Press tab > UV Unwrap and place the node between the
polyextrude node and the matchsize node.

11 You will see a UV grid on the columns. They are a

different scale compared to the columns. Press tab > UV
Transform and place the node between the uvunwrap node and the
matchsize node. Set the Scale values to 5, 5, 5. Now the UVs appear
more alike.
Choose Assets > Save Asset > Building to save the changes to the
HDA file on disk.

12 IN UNREAL – Press Rebuild. You won’t see the UVs until

you add a Material.
In the Content Browser, navigate up to Content > Starter Content
> Materials. Make sure the Building asset is selected then add a
Material such as M_Concrete_Tiles to the building geometry in the
scene view. You may need to drag it to both the columns and the
slabs.

13 Press Play to walk through the level and explore the

HOUDINI FOUNDATIONS

textured building.
This asset could be used more than once on a level to create
different buildings with different floor heights and building heights.
You could also promote more parameters to the asset to add more
control. All the assets on your level that come from this digital asset
can all be updated at once when a change is made to the asset’s
network or its parameter interface.

186
PART THREE
Copy to Points
In this part of the lesson, you are going to copy some boxes to points of another grid. You will
then randomize those points to achieve a more organic look and add random attributes to
rotate and scale the boxes to create more variety in the distribution of the shapes. This creates
another procedural system that you will wrap up into a Houdini Digital Asset.

01 IN HOUDINI – Go to the Object level and hide the

building. In the viewport, press the c key to bring up a
radial menu and select Create > Geometry > Grid. Release and then
press Enter to place at the origin. You will start by using the points
on this grid surface to instance geometry.
Use the same radial menu to Create > Geometry > Box and again
press Enter to place at the origin. In the Operation Control bar
at the top of the viewport, set the Size to 0.1 0.1 0.1. This is the
geometry that you will be copying to the points on the grid.

02 With the box still selected, go to the Modify shelf and

click on the Copy to Points tool. Select the grid and press
Enter. Now the boxes have been copied to all of the grid points.
You can now adjust the look of the system by editing parameters.
Select the grid node and change the Size to 6, 6 and the Rows and
Columns to 12. The grid gets bigger and has more points but the
boxes aren’t being copied to all the points. Click on the copytopoints
node where you can see that Target Points are set to 0-99 to match
the points on the original grid. Remove the 0-99 to copy boxes to
the whole grid.

03 In the Network Editor, press tab > Scatter. Place the

scatter node in the Network editor between the grid
and the copytopoints nodes. It will wire itself into the network
automatically and the boxes will now be copied to these new
points.
Set Force Total Count to 250. Play with the Relax Iterations to
adjust how the points are being laid out. The scatter node gives you
a more organic layout compared to the original grid points.

SOURCE GROUP
When you model in the Scene view, your selected geometry will get
placed into the Source Group field as either numbered points or primi-
Copy to Points

tives depending on the tool. If the field is empty then the tool will work
on ALL of the points or primitives.
When you use tools at the geometry level this field will be left empty
when you choose Select All [n] before using the tools. Because you set
up copytopoints at the object level, it filled in the points as 0-99.

187
 04 In the Network view press tab and start typing Match
Size then select Match Size . Place this node between the
box node and the copytopoints node. In the Parameter pane, under
Matching, set Justify Y to Min. .
On the box node, set Size Y to 0.3 to test the expression. You can
see that all of the boxes are sitting on the ground.

05 In the Network Editor, press tab and begin to type rand...

then choose the Attribute Randomize tool. Place the
attributerandomize node between the grid and the scatter nodes. At
first, you will see random colors on your boxes because the default
attribute is color (Cd).
Set the Attribute Name to N. This changes the attribute to the
normal direction and all the boxes are now pointed in different
directions. Set Max Value Y to 0 which will limit the randomness to
the X and Z directions. Rename it attrrandomize_rotate.

06 Press the Alt key and drag on the attrrandomize_

rotate node to make a copy of it. Drop it between
the attrrandomize_rotate and the scatter nodes. Rename it
attrrandomize_scale.
Set the following:
ƒ Attribute Name to pscale.
ƒ Min Value to 0.5
ƒ Max Value to 2.
This gives you a nice variety of sizes for the boxes on the grid.

07 Select the grid node and increase the Rows and Columns
to 30 – this creates more points on the grid which
creates a wider variety of values on the scatter points.
Save your work.
HOUDINI FOUNDATIONS

HOW ATTRIBUTES WORK

Attributes that are assigned to geometry are passed down the
network chain to different nodes. In the case of this network, the
attributes being assigned to the grid geometry are passed on to
the scattered points which in turn affects the copied geometry.
This is an important way to control the flow of data in Houdini.
The attributes are initially assigned to the points on the grid there-
fore more points lead to more randomization in the attribute values.

188
PART FOUR
Create another Houdini Digital Asset
In this part of the lesson, you are going to create a digital asset and test the system in Unreal. Just like the build-
ing this means wrapping up the network and saving the results to disk as an HDA file. You will promote some
parameters to allow you to control the grid size, number of points and the relaxation. The parameters will then be
available to you in Unreal.

01 Select all the nodes in the Network editor. From the

Asset menu, choose New Digital Asset from Selection.
This will collapse the network into a subnetwork then use that
subnet node to create the Digital Asset.
The nodes you used to build the asset will remain a part of the
asset even after you save it. This will allow you to continue making
changes even after you have started using it in your game levels.

02 Set the Operator Name to populate which will change the

Operator Label to Populate. Click on the button to the
far right of Save to Library. In the Locations sidebar, click on $HIP
and then double-click on the hda directory. Press Accept and then
Accept again to save the asset to disk.
This creates a new Houdini Digital Asset file (.hda) on disk that is
being referenced by this scene. It can also be referenced into other
Houdini scenes or into other applications such as Unreal using the
Houdini Engine.

03 The Edit Type Properties window opens up. This panel is

where you will build the user interface for your asset. You
will revisit this window later. Click Accept to close this window.
In order for you to maintain the procedural nature of the Houdini
Digital Asset, you can build a high level interface that can be

Create another Houdini Digital Asset

used to access the nodes inside the network. You will add to the
interface for this asset later on in the lesson.

04 IN UNREAL – In the Content Browser, go back to the

Content directory. Click Import to and find the populate.
hda asset file in the current project directory. Drag the asset from
the Content Browser to the 3D workspace.
Press Play and walk around the asset. It is there but there is nothing
special about it. Press Esc to return to the asset UI.

189
 VERTEX NORMALS
By default, Houdini objects like the box have point normals but
don’t have vertex normals. To make sure that you have proper vertex
normals, you can add the Normal node which uses a cusp value to
decide which edges should appear hard and which ones appear soft.
Game editors such as Unreal need vertex normals to display properly
which can be set up easily in the existing network.
With Normals Without Normals

05 IN HOUDINI – One thing that you will notice about the

asset in Unreal is that the boxes don’t have sharp corners.
To fix this you should go back to Houdini and in the Network editor,
press tab > “norm...” then choose the Normal tool.
Add the Normal node right after the box node. From the Asset
menu, choose Save Asset > Populate. This saves the change to the
.hda file which makes the updated asset available to anyone who
is has it loaded. In this case it means that you can update the asset
definition inside Unreal and the correct normals will display.

06 IN UNREAL – In the details panel, under Houdini Asset,

open the Cooking Actions section. Click the Rebuild
Asset button to accept the changes. This makes sure that the
changes you made inside Houdini are properly updated in the
Unreal scene.
The asset now has proper normals but you don’t have any control
over the procedural network. To create an interface that you can
use in Houdini, you are now going to promote some parameters
from inside the asset to the top level.

07 IN HOUDINI – Choose Assets > Edit Asset Properties >

Populate. Click on the Parameters tab. In the Network
editor, click on the grid node. From the Parameter pane, drag the
Size parameter onto root in the Existing parameters list.
In the Network editor, click on the scatter node. From the
Parameter pane, drag the Force Total Count parameter onto root.
In the Parameter description, change the Label to Number of
Instances. Drag the Relax Iterations and Max Relax Radius to add
these parameters to the asset. Click Accept which saves these new
parameters to the asset.

08 IN UNREAL – Click the Rebuild Asset button to accept

HOUDINI FOUNDATIONS

the changes. The promoted parameters show up in the

Parameter pane. Change the Size and Number of Instances to see
how they affect the look of the resulting grid of boxes.
Now the procedural nature of the asset has been exposed and you
can create unique versions of the asset in different levels. You can
add more than one of these populate assets in this level and each
of them can have unique settings while referencing the same asset
in the .hda file. You can also use this asset in multiple levels being
worked on by multiple artists.

190
PART FIVE
Set up Instancing
When you set up instancing in Houdini properly, you can replace the default shape you have
copied to the points with other Unreal props. You can add more than one prop that will be ran-
domly distributed in place of your default shape. You are now going to make sure that instanc-
ing is in fact set up properly and then you will use this to add some other props into the system.

01 IN UNREAL – In the Details panel, go to the Houdini

Outputs section. You can see that the whole collection of
boxes is being imported as a single mesh. This means that you are
not yet using instancing. Houdini is copying the boxes to the points
then outputting the resulting geometry for Unreal.
This is not very efficient for gameplay. Therefore you need to set
things up just a little differently to get the instancing that you need
for this tool to work properly.

02 IN HOUDINI – Select the copytopoints node and turn on

Pack and Instance. From the Asset menu, choose Save
Asset > Populate.
By using packed primitives, the box geometry feeding into the
copytopoints node is treated as a single primitive. This sets up
instancing in Houdini and in turn triggers instancing in Unreal when
this asset is loaded into the editor using the Houdini Engine plug-in.

03 IN UNREAL – Click the Rebuild Asset button to accept

the changes. Go to the Houdini Outputs section. You can
see that only a single box is being imported and instanced to the
points. It is then rotated and scaled based on the attributes you set
up earlier.
Now you can replace this default box with other geometry in your
Unreal environment. This is the ideal way of populating a series
of points because you have lots of flexibility in the editor to add
different objects to the system.

PACKED PRIMITIVES IN HOUDINI

In Houdini, packed primitives provide an efficient way of managing
Set up Instancing

instances for viewport display and rendering. The geometry feeding

the copy node is packed into a single primitive and then treated like
an instance. For the copied boxes shown in this lesson, this takes you
from 1,500 primitives down to 250 once packing is in place. With the
Houdini Engine plug-in for Unreal, the packed primitive is recognized
as a Unreal instance which will allow for more efficient gameplay.

191
 INSTANCES IN UNREAL
When the Houdini Engine plug-in detects packed primitives, a
Houdini Output Instancer is created in the Details tab that contains
the input geometry and some parameters for Rotating and Scaling
the instance. You can replace this with geometry from your Unreal
content window and size it accordingly. The Plus sign lets you add
more instanced inputs to create more variations in the same system.

04 Go to the Houdini Outputs section and expand

HoudiniOutput1 instancer. This is the box instance which
you can replace with content from within Unreal.
In the Content Browser, open Starter Content > Props. Drag the
SM_Bush prop over to the Houdini Outputs. Set the Scale Offset to
0.25 in all three axes. The geometry is instanced to the points in the
populate asset and rotated and scaled just like the boxes.

05 Now click the Plus [+] sign next to the instance object.
This adds a second one. Drag the SM_Rock prop over to
the new Houdini Instanced Input. Set the Scale Offset to 0.1 in all
three axes..
In the Details panel, scroll to the Houdini Engine section and click
the Bake button. In the Outliner, scroll to the HoudiniAssetActor.
Click the eye icon to hide it so you can focus on the digital asset.
Press Play and walk around the scene to see the instances in
action.

06 Now click the Plus sign next to the bush object. This adds
a third one. Navigate to the Content > LevelPrototyping
> Meshes folder. Drag the SM_ChamferCube over to the new
Houdini Instanced Input. Set the Scale Offset to 0.4 0.4 0.2.
You can continue to add more instanced objects to create more
variety and maybe change the size. You can use the Relax Iterations
parameter if you need to separate the instances from each other a
bit more.

07 In the Generate section of the building asset’s Details

HOUDINI FOUNDATIONS

panel, press Rebuild. In the Bake section, turn on Replace

Preview Bake then click on the Bake button.
Press Play to walk around the scene to see the instanced geometry
in action. Now you can see that you are colliding with the cubes
which already had collision geometry set up properly. You should
always make sure your instanced objects have collision geometry
set up properly.

192
PART SIX
Use Geometry to Drive Asset
So far this asset supplies the input geometry for the populate asset. Another option is to use existing geometry
in the Unreal scene to instance the geometry. You will now add an input node to your asset which will accept this
Unreal geometry. The ability to work with objects on our level creates better integration between your procedural
assets and existing game art.

01 IN HOUDINI – Choose Assets > Edit Asset Properties >

Populate. Stay on the Basic tab and set Maximum Inputs
to 1. Click Accept on the Type Properties window to save the
changes. This will create an input node inside the asset that you can
wire into the network. This input will let you grab geometry from
the Unreal scene later.
Make sure you leave Minimum Inputs set to 0. If it is set higher
than that then the asset will ONLY work if the minimum input
requirement is met. Otherwise the asset won’t cook and you won’t
see anything.

02 In the Network editor, add a Switch node after the grid

and wire in the new Input. Add a Subdivide node then
wire it after the input node to make sure there is enough detail for
the randomizing of attribute values. Set Depth to 5.
You are getting an Error on the subdivide node because in Houdini
there is nothing feeding into the third input. Go back up one level
and feed a box into the asset’s input to help preview how this is
going to work. The box will subdivide into a sphere in Houdini
which is not how it will work in the final scene.

03 Choose Assets > Edit Asset Properties > Populate. From

the switch node, drag the Select Input parameter to
the Parameter list. Select the Select Input parameter. Click on the
Menu tab then turn on Use Menu.
Add 0, Grid then 1, Input Geometry. Click Accept on the Type
Properties window to save the changes.
This will add a menu to the UI of the asset which you can use in
Unreal.
Use Geometry to Drive Asset

04 IN UNREAL – Click the Rebuild Asset button to

accept the changes. Drag a new populate asset into the
foreground. Go to the Houdini parameters section and from the
Select Input menu choose Input Geometry.
In the Details panel, go to the Houdini Inputs section and from the
menu, choose World Outliner Input. Click on the Start Selection
button and in the 3D scene, select all the parts of the ramp and
platform. Click Use Current Selection and now the instances are
being scattered inside the subdivide platform. This is not exactly
what we need.

193
 USING THE INPUT NODE IN UNREAL
When you set up an input node in your asset, there are a number of
options for accessing input geometry. You can use Geometry from
the content browser.
You can set up Curve Input where you draw a curve in Unreal. You
can also grab content from the World outliner or input Unreal land-
scapes at heightfields for use with Houdini’s new terrain toolset.

05 IN HOUDINI –You are now going to isolate the top faces

of the box to limit where the points are being copied.
Insert a Group node after the input node in the network editor. Set
the Group Name to upfacing. Under Base Group, turn Enable to
Off. Under Keep by Normals, turn Enable to On. Set the Direction
to 0, 1, 0 and Spread Angle to 0.
Instead of scattering points to all the faces of the geometry in
Unreal, you are going to isolate the top faces and use them instead.

06 Insert a Blast node after the group node and set the
Group to upfacing and turn Delete Non Selected to On.
Now only those primitives facing up will be kept - the others will be
deleted. Choose Assets > Save Asset > Populate.
The key here is that because you used a group to identify the top
faces, it doesn’t matter what geometry is input into the asset the
solution will work. This is how a procedural asset works because
it provides a generalized solution instead of a specific one-off
solution.

07 IN UNREAL – Click the Rebuild Asset button to accept

the changes. Now only the top faces of the geometry
have boxes on them.
This asset can now use a default grid or geometry sourced from
the level to work. There are always lots of options available when
building Houdini Digital Assets for use in Unreal.

08 Set the Number of Instances to 40. Go to the Houdini

HOUDINI FOUNDATIONS

Outputs section and expand the instanced box.

In the Content Browser, open StarterContent > Particles. Drag the
P_Fire prop over to the Houdini Instanced Input. Set the Rotate Y
to 90. Press Play to test the level.
The instanced points in this asset can be used for more than just
geometry. They are now a part of the Unreal level and can be used
to solve different problems. You may want to hide populate2 with
the fire while you add more to your scene. It will still appear when
you play the level.

194
PART SEVEN
Import RBD Simulation into Unreal
In this part of the lesson, you are going to create a wall then smash it using rigid body dynam-
ics. Export this system to FBX then import into Unreal for use in your game. This is a simple
example of bringing visual effects from Houdini to Unreal.

01 IN HOUDINI - Create a Box. Go to the geometry level

and set its Size to 0.5, 4, 8.
Add a Match Size node and in the Parameter pane, under
Matching, set Justify Y to Min. This raises the box up so that it sits
on the ground.

02 Go to the Object level. Rename the node wall. Select the

wall node and from the Model shelf, select Shatter.
Dive down to the geometry level, select the chunkcenters node and
set the Force Total Count to 100.
Add an Assemble node at the end of the chain. Turn off Connect
Inside Edges and turn on Create Packed Primitives and Path
Attribute. Change the Path Attribute to:
op:`opfullpath(‘.’)/path’

03
Plane.
Add an RBD Bullet Solver node to the end of the chain.
On the Ground tab, set Add Ground Plane to Ground

Press Play to run the simulation. Not much happens because the
pieces of the wall are just falling down.

Import RBD Simulation into Unreal

04 Add a sphere node to the network. Set it in front the wall

with a Center X of 2 and a Center Y of 1. Press the Alt
key and click on Center X to keyframe it.
Go to frame 9. Set Center X to -3 and Alt-click on Center X to set a
second keyframe.
Wire the sphere into the fourth collision input of the rbdbulletsolver
node. Under Collisions set Collision Type to Deforming.
Press Play to run the simulation. The wall is now being smashed by
the sphere which will be hidden in the simulation.

195
 05 Add a Transform node after the rbdbulletsolver node then
set Uniform Scale to 100. This scales the simulation to
the size needed in Unreal.
Add an FBX Output node. Set Valid Frame Range to Render Frame
Range then set Output File to $HIP/wall_destruction.fbx.
Turn on Build Hierarchy from Path Attribute.
Click Save to Disk to save to disk.

06 IN UNREAL – In the Content Browser, go back to the

Content directory. Press Import and grab the wall_
destruction.fbx file. Check Skeletal Mesh, Import Mesh, Import
Animations.
Expand the Animation section and make sure Import Meshes in
Bone is checked off. Expand the Mesh section and set Normal
Import Method to import Normals and Tangents. Click Import.
A panel will come up saying that UVs haven’t been set up yet. Close
this. Four new items appear in the content list.
Drag the wall_destruction_Anim asset into your workspace.

07 Press Play to see the simulation working in game. Right

now the animation just loops and doesn’t interact with
anything. You could set up a Blueprint to trigger the animation
based on some action on the character’s part.

CONCLUSION
You have now created a variety of game assets for use in
Unreal. From Houdini Digital Assets to FBX files containing
rigid body simulations, you now have a good foundation to
build from. These Digital Assets let you integrate Houdini’s
node based workflow inside host applications such as Un-
real. The same workflow works with other applications such
as Unity, Autodesk Maya and Autodesk 3DS Max.
Go to SideFX.com/unreal for more information. This will point
you to a starter kit of assets you can use in Unreal right away
along with a link to more tutorials.
HOUDINI FOUNDATIONS

Be sure to take a look at Project Titan an in-house tech demo

designed to explore the creation of a 3D environment that lever-
ages the latest technologies in Unreal. The tools and techniques
created for Project Titan have been shared with the community as
learning materials and downloadable content.

196
HOUDINI FOUNDATIONS

BUILD A CITY WITH PDG

To manage pipeline workflows, you can turn to Task Operators (TOPs) which are built using a technology
called the Procedural Dependency Graph (PDG). Workflows created using PDG use TOP nodes that
generate work items that distribute tasks to either your local computer or a larger compute farm.
The TOP network lets you determine the dependencies between different work items and how they are
contributing to the final output. This information is easy to visualize in the node graph which can be used
to define how you want data to flow through your network. TOPs lets you build workflows that can be used
to automate, analyze and scale your pipeline.
In this lesson, you will use TOP nodes to take a city map, create buildings for each city block and then grow
this system to handle more complex buildings and larger city maps. Houdini artists will probably know how
to do this in SOPs, but by using TOPs, you can learn the PDG workflow while creating a system that can be
easily scaled to distribute multiple tasks to an external compute farm to process in parallel.
NOTE: This lesson uses Image Magick - make sure this app is installed on your computer.

LESSON GOAL
ƒ To create a TOP (Task Operator) Network to build up a procedural city and then render it out.

WHAT YOU WILL LEARN

ƒ How to convert a city map image into geometry
ƒ How to set up a TOP network to save out the city blocks
ƒ How to create Geometry in TOPS to build buildings on each city block.
ƒ How to create a city core to make some buildings higher than others
ƒ How to wedge the cityscape to try different locations for the city core
ƒ How to wedge the use of different map images
ƒ How to use TOPs to render out the city and compare wedges using an image mosaic

LESSON
COMPATIBILITY
Written for the features in Houdini 19.5+

The steps in this lesson can be

completed using the
following Houdini
Products:

Houdini Core ✔
Houdini FX ✔
Houdini Indie ✔
Houdini Apprentice ✔
Houdini Education ✔
Document Version 2.0 | Oct 2022
© SideFX Software

197
 PART ONE PROJECT FILES
Create a City Grid Go to the Foundations tutorial page on
SideFX.com, where you got this tutorial,
To build a procedural city, you will start with a city grid. You are going to and download the tops_lesson directory.
create the geometry by tracing an image file that includes a black and white Put it into the Houdini Projects directory
image of a city map. This will be the input geometry for the network you will which you can find in either the home
be building in TOPs. directory or the documents directory.

01 Select File > Set Project. Find the tops_lesson directory

that you downloaded and press Accept. This makes
this project directory and its sub folders the place for all the files
associated with this shot.
Select File > Save As... You should be looking into the new
tops_lesson directory. If not then click on $JOB in the left hand
column. Set the file name to city_01.hip and click Accept to save.

02 In the viewport, press tab to bring up a menu and start

typing TRACE. Choose Trace and your cursor now shows
the outline of a square waiting to be placed in the scene. Press
Enter to place it at the origin. Right now it is a traced circle.
Double-click on the node in the Network view to dive down to the
geometry level. Select the trace node and click on the File chooser
next to the Image Input parameter. Click on $HIP then navigate
into the tex directory. Select the citygrid.jpg image and then click
Accept. Next turn ON the Scale to Size option and set it to 500,
500. This will give you more accuracy.

03 In the upper section of the trace node, set the following:

ƒ Rotate X to -90
ƒ Scale to 100, 100
In the Network view, press tab and start typing Reverse. Choose
the Reverse node and place it down then wire the trace node into
it. Set its Display Flag. This node will point the normals up. In the
viewport, press Spacebar-H to view the whole city grid.
Turn on Display Points in the Display option bar to see that there
are lots of trace points around each city block.

04 In the Network view, press tab > Resample and click to

HOUDINI FOUNDATIONS

place it under the reverse node. Don’t wire it in yet. Set

Length to 1 then connect the output of the reverse node into the
resample node’s input.
RMB-click on the output of the resample node and choose Null.
Click to place the null node at the end of the chain. Set its Display
flag and rename it CITYBLOCKS_OUT.
Save your work.

198
PART TWO
Generate and Display Work Items
Now that you have a city grid, you can break out the different city blocks into separate work items.
That will allow you to use each block to generate buildings. You will set up a TOP network to save
out each of the blocks. At the same time, you will create an object for visualizing selected work
items to verify that you are getting the results you want.

01 Go back to the object level and rename the object to

street_grid. Go to the Network view and press Tab and
start typing TOP... and select TOP Network. Click to place the node
in the network.
Click on the little arrow in the top right of the Network View. From
the menu choose Split Pane Left/Right. Click on the Pin icon on
the Network view to the left. The other Network view is already
pinned. In the Network view on the right, double-click on the
topnet to dive into it. You can now work with both networks as you
develop the city.

02 In the TOP network press tab > Geometry Import. Click

to place the node. Set Geometry Source to SOP Node
and then click on the Choose Operator icon next to SOP Path. Use
the floating window to navigate to and select the CITYBLOCKS_
OUT null node.
Under Storage, leave Store Geometry As set to External File.
Under Data Extraction, set Copy from Class to Primitive. This will
store the files to disk so that they can be retrieved by the next TOP
node. This is especially important if you set up a TOP network to
distribute tasks to a compute farm.

03 Rename the geometryimport node to cityblocks and from

the Task bar at the top, click on the Cook Selected Node
button. You could also select it and hit Shift-G to cook it. You will
see dots representing work items appear as the node is cooked.
Ctrl-MMB-click on the dots to view the work item attributes. Some
attributes like Index are basic tops attributes that every work item
will have. Others are specific to the type of work item. You will

Generate and Display Work Items

see that each work item is associated with an Output file. You can
RMB-click on a work item and choose View Work Item Output to
see the geometry it contains in a separate geometry viewer.

WORK ITEMS
TOP nodes create work items for each task that you ask it to perform. These are
represented on the TOP node as dots and you can use them to visualize their sta-
tus and can select them individually to evaluate how that work item is progressing.
When all the work items are fully cooked, the node gets a checkmark and the com-
pleted work items are colored green.

199
 WHAT IS THE SCHEDULER NODE?
When you first create a TOP network, a local scheduler node is created to orga-
nize tasks as they are cooked. The local scheduler points to your local computer
and will use a percentage of available cores to cook. You can set Total Slots to
Equal to CPU Count Less One option to match available cores.
You can also set up scheduler nodes for HQueue, Deadline, Tractor and Python
to send tasks to a larger compute farm. This will allow more work items to be
processed in parallel to make your graph more efficient.

04 If you want to speed up processing on your network,

select the localscheduler node and set Total Slots to Equal
to CPU Count Less One. Now future cooks will be faster because
more processors are being used.
To see the work items in the viewport, you will need to set up a
work item viewer. In the Network view on the left, press tab > File.
Click to place the node and rename it work_item_viewer.
Normally you would point a File node to a file on disk. You will do
this at first to find the files then use a different method to acquire
the work item directly from TOPS.

05 Double-click to dive inside the work_item_viewer node

and in the Parameter pane, click on the Geometry File
file selector button. In the File selector, click on $HIP and then
double click on the geo folder.
Select the city_01_cityblocks file sequence. This imports the saved
out geometry as a numbered sequence. Add a color node after
the geometry sequence and set its color to red (1, 0, 0). Scrub in
the timeline to see the different pieces of geometry loading in
sequence. There are 73 pieces that have been saved out to disk.

06 Instead of linking the display of the city blocks to the

frame number, link it directly to the top network. Click
on the File node and change Geometry File to the following
expression:
`@pdg_output`
This means that instead of loading the files from disk in sequence,
you will load whichever work item is active in the topnet. At first,
you get an error because there aren’t any work items being output
from the PDG network.

07 From the Active Work Item menu in the Geometry

HOUDINI FOUNDATIONS

Network view, select different work items to select them

in the Scene view. The red color moves to the selected geometry.
The geometry related to that work item will appear in the viewport.
In the topnet, you will see that dot on the cityblocks TOP node
highlighted in yellow.
Save your work.

200
PART THREE
Add Attributes
To create buildings, you want to set a fixed base height and a random height variation so that your
buildings are not all the same size. You could set up these attributes at the geometry level of the city
map but you can also assign them here in TOPs. This will make it easier to make changes at the TOP
level if needed down the line.

01 In the TOP net, add an Attribute Create TOP. Set

Generate When to Each Upstream Item is Cooked.
Click on the Plus sign under Float Attributes to create a new
attribute then enter the following:
ƒ Name to base_height
ƒ Value to 10
This will ensure that all the buildings will have a minimum value
of 10. You can adjust this number later if you want the lowest
buildings to be a different height.

02 Click on the Plus sign again and enter the following:

ƒ Name to height_variation
ƒ Value to rand(@pdg_index)*20
This will create a random number from 0 to 20 using the work
item attribute Index as the seed. When you build the network for
creating a building, you will add this value to the base height value
to determine how tall each building will be.

03 Select the attributecreate node and press Shift-G to cook

it. In the 3D View, you will still see footprints because
you haven’t created buildings yet.
If you MMB-click on the work items, you can see that each one
has a base_height of 10 and a height_variation that is a number
somewhere between 0 and 20. The attributes have been generated
in this TOP node but they aren’t being used yet.

ADDING ATTRIBUTES IN TOPS

It would have been possible to add attributes in the geometry
network of the city grid but adding them in TOPs makes it easier to
Add Attributes

make changes to them within the context of TOPs.

Attributes are an important way of feeding important information
through your pipeline and setting them up properly is important.

201
 PART FOUR
Create Buildings for the City Grid
You are now going to use a ROP Geometry node to create a simple building. To see what you are
doing in the viewport, you need to generate the work items and select one of them. You can then
use that work item to design the building at the geometry level.

01 Add a ROP Geometry Output TOP. Rename it create_

buildings. Turn OFF the Use External SOP option.
RMB-click on it and choose Generate Node to create the work
items. These are grey work items that have not been processed
yet. In effect they are place holders for the tasks that will take place
once you set up this node.
IMPORTANT: Click on one of the work items. This will produce
an error in the work_item_viewer network because this node isn’t
producing any geometry yet.

02 Double-click on this node to dive to the geometry level.

Create a PolyExtrude node and place it between the
incoming and output nodes. Rename it block_offset. Set the Inset to
1 and under Extrusion, turn OFF the Output Side option. This will
inset the buildings to allow for sidewalks.
Next, add a Fuse node after the polyextrude node and set Snap
Distance to 0.2 to remove any lines that are small. Note that you
are seeing something in the viewport even though the work_item_
viewer nodes are showing errors. You will fix this later.

03 Create a second PolyExtrude node after the fuse.

Rename it to building_height and set the Distance to:
@base_height + @height_variation
Make sure the Distance parameter is highlighted. RMB-click on
the parameter and choose Copy Parameter. RMB-click on the
Divisions parameter and choose Paste Relative References. Edit
the resulting channel reference by dividing it by 2:
ch(“dist”)/2
Now you can see a grid of floors and windows.

04 In the Extrusion tab, turn ON the Front Group and

HOUDINI FOUNDATIONS

name it building_top and then turn ON the Side Group

and name it building_side. You will use these to add detail to the
building.

202
 05 Add a new Poly Extrude node into the chain and rename
it window_frames. Set the following:
ƒ Group to building_side
ƒ Inset to 0.05
ƒ Divide Into to Individual Elements
ƒ In the Extrusion tab, turn ON the Front Group and
name it window_out.
Add another new Poly Extrude node and rename it windows. Set
Group to window_out and Distance to -0.05.

06 Add a new Poly Extrude and rename it extend_roof. Set

ƒ Group to building_top
ƒ Distance to 0.5
Add another Poly Extrude node and rename it roof_edge. Set
ƒ Group to building_top
ƒ Inset to 0.3
ƒ In the Extrusion tab, turn ON the Front Group and
name it roof_out.
Add another new Poly Extrude node into the chain and rename it
roof. Set Group to roof_out and Distance to -0.25.

07 Add a color node after this and set the following:

ƒ Group to windows_out
ƒ Color to a dark grey (0.2, 0.2, 0.2).
Make sure that these nodes are wired into the output node and set
the Display flag on that node. Now the frames will be a light color
and the windows will be darker. This will help you read them in the
viewport.

08 Go back up one level and on the create_buildings TOP

change $F to `@pdg_index` to set the Output File to:
$HIP/geo/$HIPNAME.$OS.`@pdg_index`.bgeo.sc
This will use the index file of each building instead of the frame.
Click on the ROP Fetch tab and set the Cache Mode to Write Files.
Save your scene file then select the create_buildings TOP and press

Create Buildings for the City Grid

Shift-G to cook it.
Go back inside the work_item_viewer and remove the color node so
you no longer see red. Now on the create_buildings TOP node, click
on the work items to see the buildings with different heights.

WORK ITEM PROGRESS

Up until now the work items have processed very quickly but this
node takes a little longer. You can see the progress wheel which
shows the completed tasks in dark green, the in-progress tasks in
yellow and the queued work items in grey. This lets you observe
the node as it completes the assigned tasks.
At the Object level, the TOP Network also displays work item
progress for all the tasks in the network.

203
 PART FIVE
Combine the Buildings
Once the buildings have been completed, you will want to bring them back together to create the
city. This involves a partition node called Wait for All and a Geometry Import. You will also add in a
city core where the buildings will be taller than elsewhere in the city.

01 To speed up the processing of the tasks, you will use

something called PDG Services. In the TOP Network
view select Tasks > PDG Services. Here you can see the service set
up and ready for you to use.
Close the Services window then go to the ROP Fetch tab of the
create_buildings TOP and set Cook Type to Services. This will speed
up the processing of this TOP node.
Save your scene file then select the create_buildings TOP and press
Shift-D to dirty it then Shift-G to cook it. You will see that it is much
faster than it was the last time.

02 Add a Wait for All TOP node. This will just take all the
work items that the ROP Geometry node generates
and combine them into a single work item. It will also prevent any
descendent tops from cooking until all the work items above have
completed cooking.

03 Add a Geometry Import TOP. Name it building_merge.

Set Generate When to Each Upstream Item is Cooked
and set Merge Operation to Import and Merge All Geometry.
Save your scene file then select the building_merge node and press
Shift-G to cook it. When the network finished cooking, click on the
work item on the final building_merge node. You will see the whole
city in the viewport.

04 Go back to the Attribute Create TOP and edit the

HOUDINI FOUNDATIONS

height_variation attribute to the following:

ƒ Value to rand(@pdg_index) * 15
This will make the buildings a bit shorter overall.
RMB-click on the attributecreate TOP and choose Dirty This Node.
This will clear out all the work items going down the chain.

204
 05 Save your scene file. Now select the building_merge TOP
node and press Shift-G to cook it. When the network
finishes cooking, click on the work item on the final building_merge
node to see the change.

06 Navigate to the street_grid object and dive into it. Create

a color node and wire it into the chain after the resample
node. Set the Class to Primitive and the Color to black (0, 0, 0).
Rename this node base_color.
Create a sphere in the street_grid network and place it off to the
side and set:
ƒ Primitive Type to Polygon
ƒ Uniform Scale to 5
Add a Color node to the output of the sphere and set Class to
Primitive and Color to green (0, 1, 0). Name this node transfer_color.

07 Add an Attribute Transfer node after the base_color

node. Then connect the transfer_color node into its
second input.
Uncheck the Points checkbox, and set the Primitives field to Cd.
Then in the Conditions tab, set:
ƒ Distance Threshold down to 0
ƒ Blend Width to about 30.
With your cursor over the Scene view, press Spacebar-Y to google
to Hide Other Objects. You can now see the green color gradually
permeate the tiles as they approach the center of the sphere.

08 Go back to the TOP network and select the

attributecreate node and edit the height_variation
attribute to the following:
ƒ Value to rand(@pdg_index) * 15 + @Cd.g*30
This will add height to those buildings in the city core. The value of
30 will be the maximum height added. You can change this value if
you want taller or shorter buildings.

Combine the Buildings

DIRTY & CLEAN WORK ITEMS

One of the main purposes of the PDG technology is the dependen-
cies that arise in a complex system. As you make changes to the
downtown core, you will see that some nodes become dirty and have
to be recooked while others are fine the way they are. The way TOPs
handles these dependencies is one of its strengths.

205
 09 RMB-click on the building_merge node and from the
menu select Dirty This Node. You could also select it and
hit Shift-D to dirty it.
Select the building_merge node and press Shift-G to cook it. You will
be prompted to Save. When it is ready, click on the work item on
the final building_merge node.
In the other Network view, navigate back up to the object level. You
will see the city updated with the city core in the viewport.

10 Navigate to the street_grid object and with the

CITYBLOCK_OUT node displayed, you can select the
sphere node and use the handle tool to move it around to see it
affect a different part of the city grid.

11 Go back to the TOP net and RMB-click on the create_

buildings node and select Generate Node. You will see
that some of the work items have been “dirtied” automatically
because of the change at the geometry level. The other work items
are still considered clean and will not need to be recooked.

12 Now when you RMB-click on the building_merge node

and select Cook Selected Node, only the dirty work items
will be updated. Click on the work item to see the new city core.
This is one of the ways that the dependency graph in TOPs works
efficiently as you develop your workflow.
HOUDINI FOUNDATIONS

UPDATED DEPENDENCIES
When you select on a work item, you get lines that show how that
work item is connected to other work items in the system.
This is a mapping of the dependencies and this helps your graph work
efficiently because only dirty work items will be processed unless you
explicitly dirty all the work items on a node with Shift-D.

206
PART SIX
Isolate a Building
Up until now the building heights are determined by the randomness of the height_variation. If there is one building that you
want to fix to a certain height, a second attribute create node can be used to choose a primitive and set specific values. This
makes it easier to art direct a system when you want to override the randomness.

01 Press Spacebar-Y to google to Ghost Other Objects.

Go back to the street_grid object and turn on Display
Primitive Numbers in the Display Options. You can see the block
numbers. You can identify the buildings and decide if you want to
set one of them explicitly. Here block 24 is the chosen block.
You are now going to set specific parameters for this city block to
get the exact height you want instead of relying on the randomness
of our current setup. This will give you some artistic control
within the TOP network instead of relying on the randomness to
determine everything.

02 Now Alt-click-drag on the attributecreate TOP node

to make a copy of it. Don’t wire it in yet. Turn on the
checkbox next to Create When and add the following expression:
@primitiveoffset==24
RMB-click on height_variation value parameter and choose Delete
Channels then set its value to 0 and then set base_height value to
50. This will explicitly make the building on block 24 exactly 50
units high.

03 Insert this node in between the first attributecreate and

the create_buildings node. Save your work then recook
the building_merge node.
Click on the work item at index 24 on the create_buildings node to
see the building on block 24 at 50 units.

04 Click on the work item on the building_merge node to see

the new cityscape with the new building rising up. Now
anytime you recook the network, the building on block 24 will not
Isolate a Building

update but will remain set specifically.

If you want to change that building, you can use the second
attribute create node to explicitly set its height.

207
 PART SEVEN
Wedge the City Core Location
In order to wedge the position of the sphere within the street_grid object, the network will need to be turned into a Houdini
Digital Asset [HDA] and run through an HDA Processor TOP node. The ability to load these assets and add them to the system
lets you create a complex system out of discreet tools that can be updated and adapted as needed.

01 In the street_grid network, select all the nodes except

CITYBLOCKS_OUT. From the Assets menu, choose New
Digital Asset from Selection. Name the asset streetgrid_maker, label
it Street Grid Maker, and save it to the $HIP/hda/ directory.
In the Operator Type Properties panel, click on the Parameters tab.
Dive into the asset. Drag Image Input from the trace node to the
parameter list. Drag Scale X to Scale Y and choose Relative Channel
Reference. Now drag Scale X to the parameter list and name it Scale.
Drag Center from the sphere node and name it Sphere Center. Drag
Blend Width from the attributetransfer node then click Accept.

02 Go back to the TOP network. Create an HDA Processor

TOP and wire it into the Geometry Import node. On the
geometryimport node, change Generate When to Each Upstream
Item is Cooked and Geometry Source to Upstream Output File.
Select the HDA Processor node and set HDA File to $HIP/hda/
streetgrid_maker.hda. Click on the HDA Parameters tab to see the
Image Input, Center, Scale and Blend Width parameters listed.
Rename this TOP make_citygrid then press shift-V to Dirty and Cook
the node to make sure that it is still generating the city grid the way
it was before.

03 Create a Wedge TOP node and wire it into the make_

citygrid. Set the following:
ƒ Wedge Count to 4
Click the plus sign next to Wedge Attributes and set the Attribute
Name of the first one to center_wedge then set:
ƒ Type to Float Vector,
ƒ Start range to -50, 0, -50, 0
ƒ End range to 50, 0, 50, 0
ƒ Turn ON the Random Samples
HOUDINI FOUNDATIONS

WEDGING
Wedging is an idea that comes from photography where a shot is taken
using a variety of settings so that the best one can be chosen later in the
lab. Here the idea is the same except you will use random values to influ-
ence the system and get four unique results that can be compared.
Later you will render out a mosaic with all the options labelled with the
random wedge values displayed.

208
04 Select the HDA Processor node and on the HDA
Parameters tab, set the Center parameter to:
@center_wedge.0, @center_wedge.1, @center_wedge.2
Select the make_citygrid HDA processor node and press shift-V
to Dirty and Cook the node - now there are four citygrids being
created.
Hide the street_grid object. Click on the work items to visualize the
different grids. Each of them show a different position for the green
color which is where the city core is located.

05 Now add some wedging to the height variation so that

you get four different looks. Select the wedge TOP and
click the plus sign next to Wedge Attributes and set the Attribute
Name of the second one to seed_wedge, leave Type set to Float,
and set its Start/End to 0, 1000.
Turn ON the Random Samples parameter.
On the first attributecreate TOP, change the height_variation
attribute’s Value to:
rand(@pdg_index*@seed_wedge) * 15 + @Cd.g*30

06 You will have to update the output file names for all
the TOPS that create an output file to include the
wedgenumber, so that the files can be differentiated for the
different wedges. Include a .`@wedgenum` in the output file
parameters for the following TOP nodes just before .bgeo:
ƒ make_citygrid HDA Processor
ƒ geometryimport Geometry import
ƒ create_buildings ROP Geometry
ƒ building_merge Geometry Import

07 Replace the Wait for All node with a Partition by

Attribute TOP node. Make sure that Partition By is set
to Distinct Attribute Values.
Click the plus sign next to Attributes and set the Name to
wedgenum. This will create a partition for each wedge.

Wedge the City Core Location

08 Save your work then cook the building_merge TOP node.

This will create the four city maps which you can visualize
by clicking on the work items on the building_merge node.
Since this is using the local scheduler this will take a little while
longer to process. This is where using a scheduler that distributes
your tasks to a compute farm can speed things up considerably.

209
 PART EIGHT
Create Geometry for the Streets
To create context for the buildings, you will now create city blocks and streets to use in the final rendering. While the four
wedges currently all use the same map, you will set up this geometry in TOPs to allow for the use of different maps down the
line. It is always a good idea to build in flexibility to create the most robust system.

01 In the TOP network, add a ROP Geometry Output node

and branch it off from the make_citygrid HDA processor.
Rename this new node build_streets.
Turn OFF the Use External SOP option and in the Output File add
.`@wedgenum` to the expression in place of $F just before .bgeo.
On the ROP Fetch tab, set Cache Mode to Write Files.

02 RMB-click on this node and choose Generate Node to

get the four work items. Click on one of them to highlight
the work item and then double click on the node to dive into it.
This will bring you to the geometry level where you will create the
streets.

03 Add a PolyExtrude node between incoming and output

nodes. Set Distance to -0.05. Turn OFF the Output Front
checkbox and turn ON the Output Back option.
Add a Reverse node after to fix the normals and then add a Color
node to turn all the city blocks white.

04 Add a Box node into the network off to the side. Wire
HOUDINI FOUNDATIONS

polyextrude into the box node to match the bounding box

to the city grid geometry.
Add a Blast node after the box and set Group to 2 and turn ON the
Delete Non Selected option.
Add a Transform node after the blast and set:
ƒ Translate Y to -0.05
ƒ Scale X to 1.05
ƒ Scale Z to 1.05

210
 05 Add a Color node with a medium grey (0.33, 0.33, 0.33).
Now create a Merge node and connect it to the two
color nodes. Make sure the display flag is set to output to see the
street geometry in place.

06
the grid.
Save the Scene File. Go back to the TOP Network and
Cook the node, you can click on the work items to see

At the moment all four of them are the same. That will be different
later when you introduce more city grids into the equation.

07 Add a Partition by Index to the end of the chain. Feed

the building_merge node into the first input and the
build_streets into the second.
Cook the new node. When you are finished, click on the work items
and you will see that only the buildings are being displayed. If you
middle click on a work item you will see that there are two output
files being created by the Partition. You need to display the new
one to see it in the viewport.

08 Go to the Object level. Alt drag on the work_item_viewer

to create a second one called work_item_viewer1. Dive
into the geometry level and set Geometry File to:
`@pdg_output.1`
Go back up one level to the object level then in the TOP network
select one of the work items on the partitionbyindex node. Now you
will see both of the outputs being displayed in the viewport. This is

Create Geometry for the Streets

important for the next step when you will render out images of the
city.

ANOTHER PDG OUTPUT

Earlier you used pdg_output to create the work_item_viewer node to help you visualize what is com-
ing out of each node. Now that you have this partition in place, there are actually two outputs which
you can see by Ctrl-MMB clicking on a work item. This second output therefore needs its own viewer
to be properly displayed.

211
 PART NINE
Wedge Four City Maps
Let’s add one extra variable into the wedging. You are going to access four different maps and create one rendering for each
one of them. Each of the different black and while images will be traced and used to feed the system. This shows another way
to add more content into the pipeline.

01 Go back to the wedge TOP node. Click the plus sign next
to Wedge Attributes to add a new wedge parameter.
Set the following:
ƒ Attribute Name to file_wedge
ƒ Attribute Type to Integer
ƒ Start/End to 1,4
This will create integer values of 1, 2, 3 and 4 which will be used to
find the different city maps.

02 On the streetgrid_maker HDA processor node, click on

the File button next to Image Input and navigate to the
more_grids folder. Select the file from there then in the Parameter
pane, change $F in the expression to `@file_wedge`
There are four files in this folder and the wedge attribute will run
through them all in order when you cook the network.

03 Dirty and Cook (Shift-V) the streetgrid_maker HDA

processor node. You will see the four different maps
being used and the city core being moved around like before.

04 Bypass the second attributecreate TOP node. This

HOUDINI FOUNDATIONS

was designed for a static map and won’t work for four
different maps.
Cook the Final partition node (Shift-G) and view the four city plans
turned into cities.

212
PART TEN
Render a Mosaic
Now you are going to add a camera and a skylight then render out the cities. The resulting images will then be combined into
a single mosaic with the wedge attributes on display so that creative decisions can be made based on the results. The mosaic
node will use the Image Magick application which needs to be installed on your computer for these steps to work.

01 Tumble your view until you are looking down at the

city from an angle. From the camera menu, select New
Camera. Check the four city grids to make sure they all fit in the
camera. Adjust the camera view if necessary. You will use this to
render out the mosaic.
Add a Skylight. Set the Environment Light Intensity to 1.3.

02 At the object level , Alt-drag on work_item_viewer to

make a copy. Rename the new node render_item_viewer.
Dive into this node and change the Geometry File parameter to:
`@pdg_input`
Repeat these steps to create render_item_viewer1 with its Geometry
File parameter set to:
`@pdg_input.1`
Set the display flag on these new render_item_viewer nodes and
turn off the display on the two work_item_viewer nodes.

03 Append a ROP Mantra Render top after the

partitionbyindex. On the ROP Fetch tab, set the Cache
Mode to Write Files.
Make sure the camera parameter is set to the camera you made.
Turn On the Override Camera Resolution option and set it to ½.
Change the Output Picture parm to:
$HIP/render/city.$OS.`@wedgenum`.exr
Note: If you are running this tutorial using Houdini Apprentice then
you should use .pic files instead of .exr to avoid the Apprentice
watermarks.

PDG INPUT
While PDG OUTPUT lets you take the result of a node and display it, PDG
Render a Mosaic

INPUT will take the result of the node before and display that instead.
This way the Mantra node can use the results from the preceding node
as the input for the geometry and then render that geometry. These two
options are similar but understanding the difference is important.

213
 04 Append an Overlay Text top and set Output Picture to:
$HIP/render/$HIPNAME.$OS.`@wedgenum`.exr
On the ROP Fetch tab, set Cache Mode to Write Files.
In the Overlay tab type:
Map: `@file_wedge`
City Core: `@center_wedge.0`, `@center_wedge.2`
Variation Seed: `@seed_wedge`

05 Double-click on the Overlaytext node to dive into it.

Select the Font node and change the Text Size to 18. This
will create smaller text and leave more room to see the city.

06 Go back up and add a Wait for All partition node. This

will bring together all of the elements. Now add an
ImageMagick node.
Cook this node to process the TOP network. This will create four
renderings and then overlay the text and combine them into a
mosaic.

07 When you are finished, RMB click on the work item and
select View Task Output. This will show you the final
image in the Mplay image viewer.
This contact sheet lets you review the different wedges and
decide which one you want to move forward with. The parameters
displayed on the images are the ones you would need to lock down
to get the city you want.
HOUDINI FOUNDATIONS

PDG FOR MORE THAN HOUDINI

With PDG’s ability to work with Python, you can use it for tasks that are not
directly connected to Houdini. It offers a visual tool for organizing your pipeline
tools with TOP nodes organizing your workflow.
These networks can be processed through Houdini and sent to your chosen
scheduler. This will allow you to send your work to a compute cloud for
faster processing.

214
PART ELEVEN
Scale Up to Create More Content
In the beginning you pointed out that this city building example could probably be created in SOPs without too much difficulty.
The issue is that as the system gets more complex there is a bottleneck since all the processing happens within a single net-
work. With TOPs, you can distribute work items to a compute farm and as you scale up things don’t have to slow down.

01 On the wedge TOP, set the Wedge Count to 1. This will

give you only one iteration for now while you scale up
the city using a new map.
Select create_buildings ROP Geometry TOP node and RMB-click
and choose Generate Node. Now you are going to dive in and use
a digital asset to create a different type of building with floors and
windows and columns.

02 Double-click to dive down. From the Assets menu, select

Install Asset Library. Navigate to the hda directory and
select the make_buildings2.hda. Click Accept then click Install (not
Install and Create) Press tab > Make Buildings Hi Res to place the
node into the graph.
Wire the fuse node into the new asset.

03 On the make_buildings_hi_res digital asset, set the

following:
ƒ Floor Height to 2.
ƒ Base Building Height to @base_height
ƒ Height Variation to @height_variation
Add a switch node between the color node and the output node
then wire the make_buildings_hi_res node into the second input and

Scale Up to Create More Content

set Select Input to 1.

04 Select the Partition by Index node and Cook the network

to see the new buildings. Now they have floor slabs,
columns and windows.
A lot more geometry has been produced and you are starting to see
that increasing the number of available work items using a compute
farm would be a good idea. You could do this using one of the other
scheduler nodes.

215
 05 You could also make the city bigger using a different city
grid image. You would need to increase the Uniform
Scale on the streetgrid_maker HDA processor to 500, and change
the Image input to the citygrid_large map. You may also want to
increase the Blend Width to create more falloff in the city core.
On the Wedge node, you could set the center_wedge attribute’s to
something like -110, 0, 100.

06 Dirty and Cook the streetgrid_maker HDA processor

node. You will see the new city map being generated.
When this is finished, you will have a lot more city blocks,
compared to the original map and a lot more buildings are being
generated.
Going to a compute farm would definitely make sense at this point
because a single computer is not taking advantage of the parallel
processing capabilities of TOPS.

07 Recook the final node to render out an image of the city.

You can see how the simple system you built has the
potential to grow to whatever size you need.
Save your scene.

CONCLUSION
In this lesson, you have used TOP nodes to take a city map, create
buildings for each city block and then grow this system to handle more
complex buildings and larger city maps. This project has introduced you to
typical nodes used in a TOP-based workflow and how they can be used
to create and process work item tasks.
There are many things you can create using TOPs and this is only the
beginning. Not only can you use this network type to automate and
process typical Houdini workflows, you can use it to process
workflows that don’t involve Houdini at all.
This will make it a great tool for managing
HOUDINI FOUNDATIONS

dependencies throughout your pipeline,

whether you are a small studio looking
for efficiencies or a big studio
managing lots of data.

216
NOTES

217
 SCENE VIEW SHORTCUTS
TOOLS SELECTION MODES VIEWPORTS
Select S Objects 1 Expand Viewport Space + b
Move T Points 2 Select Viewport Space + n
Rotate R Edges 3 Perspective View Space + 1
Scale E Primitives (Faces) 4 Top View Space + 2
Pose Ctrl-R Vertices 5 Front View Space + 3
Handle Enter Right View Space + 4
View Esc Select Groups/Connected Geometry 9 UV View Space + 5
Tool Menu Tab Toggle Objects/Geometry F8
Custom Radial Menu C
Toggle Wireframe/Shaded W
Repeat Last Tool G SELECTING
Display Options D
Select LMB
VIEW
Add to Selection Shift + LMB VIEWPORT LAYOUT
Tumble Space/Alt + LMB Remove from Selection Ctrl + LMB
Track Space/Alt + MMB Single View Ctrl + 1
Toggle Selection Ctrl + Shift + LMB
Dolly Space/Alt + RMB Four Views Ctrl + 2
Home Grid Space + H Two Views Stacked Ctrl + 3
Select All N
Home All Space + A Two Views Side by Side Ctrl + 4
Select Nothing Shift-N
Home Selected Space + G Three Views Split Bottom Ctrl + 5
SNAPPING RADIAL MENU Three Views Split Left Ctrl + 6
VIEW RADIAL MENU Four Views Split Bottom Ctrl + 7
Grid Snap X 

Selection Tools V  Four Views Split Left Ctrl + 8
Primitive (Curve) Snap X 

Selection Options V 
Point Snap X 

Viewport V  FINDING THINGS
Multi-Snapping Snap X
Shading V Dashbox Ctrl + D

NETWORK VIEW SHORTCUTS

VIEW DOTS NAVIGATION
Pan Space + LMB or MMB Add Dot Alt + LMB on wire Enter a Node Double-click or Enter
Zoom Space + RMB or Scroll Wheel Pin/Unpin Dot Alt + LMB on dot Go up a level U
Show all Nodes H Radial Menu N
Show Selected Nodes G TOOLS Create a Quickmark Ctrl + <# 1-5>
Go to a Quickmark Shift + <# 1-5>
Toggle Parameter Pane P
CREATE Go to Previous View ` (Backtick)
Toggle Tree View Shift + W
Select the Node Upstream PgUp
Node Menu Tab Toggle Network Overview O
Select the Node Downstream PgDn
Add File Node = Toggle Color Palette C
Select Previous Sibling Shift + PgUp
Create Subnet Shift + C oggle Shape Palette S
Select Next Sibling Shift + PgDn
Add Background Image Shift - I
CLICKS AND DRAGS
ORGANIZE NODES
NOTES AND NETWORK BOXES Select LMB
Layout all L
Add Network Box Shift + O Add to Selection Shift + LMB
Align A + LMB-Drag Down/Across
Add Sticky Shift - P Remove from Selection Ctrl + LMB
Minimize Selected Notes/Boxes Shift - J Start Wiring from Node Alt + LMB DISPLAY FLAGS | SOP LEVEL
Expand Selected Notes/Boxes Shift - K Select Node + Inputs Alt + Shift + LMB
Render T + LMB
Shrink box to fit contents Shift - M Select Node + Output Alt + Ctrl + LMB
Display R + LMB
Select Inputs + Outputs Alt + Shift + Ctrl + LMB
Template E + LMB
WIRING Move Node LMB-Drag
Footprint W + LMB
Connect Nodes LMB on Connector Move Node + Inputs Shift + LMB-Drag
Bypass G or B + LMB
Connect Multiple Nodes J drag over nodes  Move Node + Outputs Ctrl + LMB-Drag
Insert Node RMB on Connector Copy Selected Nodes Alt + LMB-Drag
Branch MMB on Connector Copy Node + Inputs Alt + Shift + LMB-Drag
Connector List Alt + MMB on Node Copy Node + Output Alt + Ctrl + LMB-Drag
Cut Wire Y drag across wire Reference Copy Alt + Shift + Ctrl + LMB-Drag
Disconnect from Wires Shake Node

50