MODULE - 8

e-PG Pathshala
Subject: Computer Science
Paper: Computer Graphics and Visualization
Module: 2D Viewing
Module No: CS/CGV/8
Quadrant 1– e-text
Objectives:
 Understand 2D viewing pipeline
 Learn how to perform, window-viewport transformations
 Understand the theory behind clipping in 2D with focus on Point Clipping.

Discussion:

2D Viewing Pipeline:

Let’s start our discussion with 2D viewing pipeline. This involves a series of
transformations performed, right from creation, till the object is finally seen or displayed at a
position or a viewing region on a display device like monitor. Each transformation step
performed converts data from one coordinate system to the other. From the figure below
(Recollect discussion from the first module),

The sequence of steps in the 2D viewing Pipeline are

MC  WC VC CCNVC DC

where
MC – Modeling Coordinates (Modeling or Local Coordinate system)
WC – World coordinates (World Coordinate system)
VC – viewing coordinates (Viewing coordinate system)
CC – Clipping coordinates (Viewing coordinate system)
 NVC – Normalized viewing or device independent coordinates (Viewing coordinate
system, but normalized)
DC – Device coordinates (Device coordinate system specific to every device)

For conversion from MC to WC we perform a transformation called Modeling

Transformation (MT), from WC to VC we perform Viewing Transformation (VT), From VC to
NVC we perform window-viewport transformation (WV), From NVC to DC we perform Device
coordinate system transformation (DT). When the object is first created it assumes MC and a
Modeling Transformation is applied to convert to WC, which is in the world coordinate
system. Then a viewing coordinate system is set up and data is now converted from WC to
VC (recall discussion from the previous module on transformation between coordinate
systems). Later VC are sent for clipping, depending on the window or region of interest, to
get CC. In many graphics packages, Viewing and Clipping steps are combined to get clipped
coordinates straight from world coordinates. Next the CC are converted to NVC, where we
simply normalize data in a step called window-Viewport transformation. This step is
performed to ensure that the data should be independent of the end display device so that
the data can be conveniently mapped to any display device of choice. If this is not done we
need to perform a separate transformation for each display device, which is a cumbersome
task. Each kind of coordinates belong to a specific coordinate system or reference frame.

Window-Viewport Transformation:

In Computer Graphics terminology, a window is a rectangular region of interest, and

a viewport is a display area chosen on the monitor screen for viewing the contents of the
window. In short, window is ‘what we want to see’ and viewport is ‘where we want to see’. As
shown in the figure below, we can define a rectangular region in the world coordinate
system, and call that as window. The contents of the window are to be mapped to a region
or a display area chosen as viewport. In a way this is, mapping contents between two
rectangular regions of different dimensions. Window is generally defined in the world
coordinate system, while viewport is defined in the end device coordinate system. Since we
wish to convert data to device independent or normalized coordinates, we perform window-
viewport transformation. So in window-viewport transformation contents of the window are
mapped to normalized coordinates (coordinates between 0 and 1).

For simplicity, both window and viewport are chosen to be rectangular in shape
whose boundaries are aligned with the coordinate axes. Because the dimensions of window
and viewport are different, we will be performing a scaling like operation. If the window is
bigger in size than the viewport, we perform a scale down operation and vice-versa. As you
can notice from the figure above, the boundaries of the window and viewport are defined by
four infinitely extending lines parallel to the coordinate axes. The window boundaries are
defined by the four lines, xwmin, xwmax, ywmin, yWmax. Where xwmin, xwmax define the two
vertical boundaries, while, ywmin and ywmax define the two horizontal boundaries of the
window. Similarly the viewport boundaries are defined by four infinitely extending lines, xvmin,
xvmax, yvmin, yvmax.
Now for a point (xw,yw) in the window, we have to identify an equivalent point in the
viewport, say (xv,yv). i.e., for every point in the window, there is an equivalent point in the
viewport. The four corner vertices of the window have got their equivalent vertex points in
the viewport, which are nothing but the four corner points of the viewport. For the point
(xwmin, ywmin) in the window, the equivalent point in the viewport is (xvmin, yvmin) and so on.

For any general point in the window, how do we find an equivalent point in the viewport? As
we can see, from the figures above, the percentage horizontal distance of the point in the
window from its left boundary, should be exactly equal to the percentage horizontal distance
of the equivalent point in the viewport from its left boundary. We can apply the same analogy
to the vertical distance of the point from its bottom boundary.

We can represent the percentage horizontal distance of the point from the left boundary as

𝑥𝑤 − 𝑥𝑤𝑚𝑖𝑛 𝑥𝑣 − 𝑥𝑣𝑚𝑖𝑛
× 100 = × 100
𝑥𝑤𝑚𝑎𝑥 − 𝑥𝑤𝑚𝑖𝑛 𝑥𝑣𝑚𝑎𝑥 − 𝑥𝑣𝑚𝑖𝑛
After cancelling 100 on both sides we have

𝑥𝑤 − 𝑥𝑤𝑚𝑖𝑛 𝑥𝑣 − 𝑥𝑣𝑚𝑖𝑛
=
𝑥𝑤𝑚𝑎𝑥 − 𝑥𝑤𝑚𝑖𝑛 𝑥𝑣𝑚𝑎𝑥 − 𝑥𝑣𝑚𝑖𝑛

Our aim to compute the values of xv and yv in terms of other variables in the equation. We
get

𝑥𝑣𝑚𝑎𝑥 − 𝑥𝑣𝑚𝑖𝑛
𝑥𝑣 = 𝑥𝑣𝑚𝑖𝑛 + × (𝑥𝑤 − 𝑥𝑤𝑚𝑖𝑛 )
𝑥𝑤𝑚𝑎𝑥 − 𝑥𝑤𝑚𝑖𝑛
We can rewrite the above equation as

𝑥𝑣 = 𝑥𝑣𝑚𝑖𝑛 + (𝑥𝑤 − 𝑥𝑤𝑚𝑖𝑛 )𝑠𝑥

𝑥𝑣𝑚𝑎𝑥 −𝑥𝑣𝑚𝑖𝑛
Where 𝑠𝑥 =
𝑥𝑤𝑚𝑎𝑥 −𝑥𝑤𝑚𝑖𝑛

Here sx is a ratio and is also the scaling factor along x-direction. If the horizontal
boundary of the window is larger than that of the viewport then sx <1 (down scaling),
otherwise sx >1 (up scaling). If both the boundaries are of same lengths then sx=1 (no
scaling).
Similarly we can represent the percentage vertical distance of the point from the
bottom boundary as

𝑦𝑤 − 𝑦𝑤𝑚𝑖𝑛 𝑦𝑣 − 𝑦𝑣𝑚𝑖𝑛
× 100 = × 100
𝑦𝑤𝑚𝑎𝑥 − 𝑦𝑤𝑚𝑖𝑛 𝑦𝑣𝑚𝑎𝑥 − 𝑦𝑣𝑚𝑖𝑛

After cancelling 100 on both sides we have

𝑦𝑤 − 𝑦𝑤𝑚𝑖𝑛 𝑦𝑣 − 𝑦𝑣𝑚𝑖𝑛
=
𝑦𝑤𝑚𝑎𝑥 − 𝑦𝑤𝑚𝑖𝑛 𝑦𝑣𝑚𝑎𝑥 − 𝑦𝑣𝑚𝑖𝑛
And rearranging the terms as we did before for yv we get,

𝑦𝑣 = 𝑦𝑣𝑚𝑖𝑛 + (𝑦𝑤 − 𝑦𝑤𝑚𝑖𝑛 )𝑠𝑦

𝑦𝑣𝑚𝑎𝑥 −𝑦𝑣𝑚𝑖𝑛
Where 𝑠𝑦 =
𝑦𝑤𝑚𝑎𝑥 −𝑦𝑤𝑚𝑖𝑛

Here sy is a ratio and is also the scaling factor along y-direction. If the vertical
boundary of the window is larger than that of the viewport then sy <1 (down scaling),
otherwise sy >1 (up scaling). If both the boundaries are of same lengths then sy=1 (no
scaling).
It is important to note that the transformation, converts real coordinates to normalized
coordinates, which are device independent. Later the device independent coordinates are
converted to appropriate screen coordinates as shown in the figure below.

*********************************************************************************************************
Example:

Problem: For a point (1, 2) in the window, find the equivalent point in the viewport. Assume
window at [(1,1) and (4,4)] and viewport at [(-1, -1) and (1,1)].

Sol:
The point (1, 2) is identified as (xw, yw) respectively. The aim is to compute the equivalent
point (xv,yv).

xwmin=1, xwmax=5, ywmin=1, ywmax=5,

xvmin=-1, xvmax=1, yvmin=-1, yvmax=1,
𝑥𝑣𝑚𝑎𝑥 −𝑥𝑣𝑚𝑖𝑛
Now compute 𝑠𝑥 using the relation derived above 𝑠𝑥 =
𝑥𝑤𝑚𝑎𝑥 −𝑥𝑤𝑚𝑖𝑛
 1−(−1)
𝑠𝑥 = =½
5−1
𝑦𝑣𝑚𝑎𝑥 −𝑦𝑣𝑚𝑖𝑛
Now compute 𝑠𝑦 using the relation derived above 𝑠𝑦 =
𝑦𝑤𝑚𝑎𝑥 −𝑦𝑤𝑚𝑖𝑛
1−(−1)
𝑠𝑦 = 5−1
=½
Now compute 𝑥𝑣 using the relation below
𝑥𝑣 = 𝑥𝑣𝑚𝑖𝑛 + (𝑥𝑤 − 𝑥𝑤𝑚𝑖𝑛 )𝑠𝑥
(1−1)1
𝑥𝑣 = −1 + 2 = -1
Now compute 𝑥𝑣 using the relation below
𝑦𝑣 = 𝑦𝑣𝑚𝑖𝑛 + (𝑦𝑤 − 𝑦𝑤𝑚𝑖𝑛 )𝑠𝑦
(2 − 1)1 1
𝑦𝑣 = −1 + = −
2 2
The equivalent point for (1,2) is (-1, -1/2)
********************************************************************************************************
2D Clipping Algorithm:

In 2D clipping, we choose a rectangular region of interest in the world coordinate

system and call that as window or clip window or clip rectangle. The contents inside the
window are of interest to us and so to be clipped. So here clipping means select the contents
that are inside the clip window region. It is cumbersome and less efficient to consider
performing transformations on the regions that are not of interest to us, instead we can
consider the inside of the clip window and perform any transformations on that. Thus clipping
is ‘selecting what is of interest to us’ and not ‘discarding what is not of interest to us’.
For better understanding and simplicity we choose the clip rectangle to be aligned
such that its sides are parallel to the coordinate axes. In the most general case we can
choose the clip rectangle to be oriented in any direction in 2D space. In the extreme case,
the shape of the clip window region need not be rectangular, instead, it can be of any shape
of choice.
The region against which the contents are clipped is called ‘clip window’. Any
procedure that identifies those portions of a picture that are either inside or outside of a
specified region of space is considered as the ‘clipping algorithm’. Usually a clipping region
is a rectangle in standard position. These boundary edges typically correspond to a
normalized square in which the x and y values range either from 0 to 1 or from -1 to 1.

Clipping transformation can be performed either in the world coordinate system or in

the viewport coordinate system. Clipping in the World coordinate system involves selecting
the contents inside the clip window region and apply transformations only on them. In
viewport clipping, the whole of the contents from the world are brought to the viewport and
then clipping against the viewport. It is thus more convenient to perform clipping in the world
coordinate system.

The types of clipping that we are going to learn are

 Point clipping
 Line clipping
 Polygon Clipping
 Text Clipping
Point Clipping:

In point clipping, points that are inside the rectangular region are selected for display
and points that are outside the region are discarded. Now the question is, how do we know
whether a point is inside or outside a given rectangular region? For this we should perform a
test as shown.

𝑥𝑤𝑚𝑖𝑛 ≤ 𝑥 ≤ 𝑥𝑤𝑚𝑎𝑥
𝑦𝑤𝑚𝑖𝑛 ≤ 𝑦 ≤ 𝑦𝑤𝑚𝑎𝑥
From the above two equations we can infer that if x and y satisfy the above
equations, the point is an inside point, so can be selected for display, otherwise it is an
outside point so can be discarded. In the above two equations, there are 4 inequalities,
𝑥𝑤𝑚𝑖𝑛 ≤ 𝑥, - (1)
𝑥 ≤ 𝑥𝑤𝑚𝑎𝑥 , - (2)
𝑦𝑤𝑚𝑖𝑛 ≤ 𝑦, - (3)
𝑦 ≤ 𝑦𝑤𝑚𝑎𝑥 - (4)

The x coordinate should satisfy two inequalities, 𝑥𝑤𝑚𝑖𝑛 ≤ 𝑥 , 𝑥 ≤ 𝑥𝑤𝑚𝑎𝑥 , which says
that the x coordinate should be between the two vertical boundaries xwmin and xwmax, while
the y coordinate should satisfy two other inequalities, 𝑦𝑤𝑚𝑖𝑛 ≤ 𝑦, 𝑦 ≤ 𝑦𝑤𝑚𝑎𝑥 which
means that the y-coordinate should lie between the two horizontal boundaries ywmin and
ywmax . A point (x,y) is said to be an inside point, only if all the 4 inequalities are satisfied by
the point. Let’s check how? Say for example, the first inequality, 𝑥𝑤𝑚𝑖𝑛 ≤ 𝑥 is not satisfied
by the x-coordinate, instead it satisfied the relation, 𝑥𝑤𝑚𝑖𝑛 > 𝑥 , it means that, x coordinate
is onto the left of the left boundary of the window, and so we can infer that the point is
outside the clip window region and so can be discarded. Even if one of the inequalities is not
satisfied by the point, the point can be safely ignored.
*********************************************************************************************************
Example:

Problem: Check the position of point (2,1) with respect to the clip window at [(2,2), (4,4)].
Sol: of course, the point (2,1) is outside the clip window and so can be discarded.
Discussion: Treating the given point as (x,y) and xwmin=2, xwmax=4, ywmin=2, ywmax=4,
Except the 3rd inequality, the other inequalities are satisfied by the point. So as per the
discussion above, the point is an outside point.
*********************************************************************************************************
Summary:
• Understood the steps in 2D viewing pipeline
• Learnt about window-viewport transformation
• Also learnt the idea behind clipping, and in particular point clipping