May 2010 1

Computer Graphics Viewing

Shmuel Wimer
Bar Ilan Univ., School of Engineering
May 2010 2
min
xw
max
xw
min
yw
max
yw
Clipping Window
min
xv
max
xv
min
yv
max
yv
Viewport
Viewport Coordinates
The clipping window is
mapped into a viewport.
Viewing world has its own
coordinates, which may be
a non-uniform scaling of
world coordinates.
World Coordinates
May 2010 3
2D viewing transformation pipeline
Construct World-
Coordinate Scene
From Modeling-
Coordinate
Transformations
World
Coordinates
Modeling
Coordinates
Convert World-
Coordinates to
Viewing-
Coordinates
Viewing Coordinates
Transform Viewing-
Coordinates to
Normalized-
Coordinates
Normalized
Coordinates
Map Normalized-
Coordinates to
Device-Coordinates
Device
Coordinates
May 2010 4
Normalization and Viewport Transformations
First approach:
Normalization and window-to-viewport transformations are
combined into one operation.
Viewport range can be in [0,1] x [0,1].
Clipping takes place in [0,1] x [0,1].
Viewport is then mapped to display device.
Second approach:
Normalization and clipping take place before viewport
transformation.
Viewport coordinates are specified in screen coordinates.
May 2010 5
( )
, xw yw
( )
, xv yv
Maintain relative size and position between clipping window and viewport.
min min
max min max min
xv xv xw xw
xv xv xw xw

=

min min
max min max min
yv yv yw yw
yv yv yw yw

=

min
xw
max
xw
min
yw
max
yw
Clipping Window
min
xv
max
xv
min
yv
max
yv
Normalized Viewport
1
1 0
May 2010 6
Scaling factors:
max min max min
max min max min
,
x y
xv xv yv yv
s s
xw xw yw yw

= =

Translation factors:
max min min max max min min max
max min max min
,
x y
xw xv xw xv yw yv yw yv
t t
xw xw yw yw

= =

( )
Solving for , obtains:
, , where
x x y y
xv yv
xv s xw t yv s yw t = + = +
( ) ( ) ( )
window,
norm_viewport
min min min min
0
, , , 0
0 0 1
x x
x y y y
s t
xv yv s s xw yw s t
=
(
(
=
(
(

M
T S T
This can also be obtained by composing transformations:
May 2010 7
World clipping window can first be mapped to normalized square between -1
and +1, where clipping algorithm takes place, and then transform the scene
into viewport given in display coordinates.
max min
max min max min
max min
window,
norm_square
max min max min
2
0
2
0
0 0 1
xw xw
xw xw xw xw
yw yw
yw yw yw yw
+
(

(

(
+ (
=
(

(
(
(

M
( ) ( )
( ) ( )
max min max min
norm_square, max min max min
viewport
2 0 2
0 2 2
0 0 1
xv xv xv xv
yv yv yv yv
+ (
(
= +
(
(

M
May 2010 8
Clipping Algorithms
7
p
1
p
2
p
3
p
4
p
5
p
6
p
8
p
9
p
10
p
Before Clipping
1
p
2
p
5
' p
6
p
7
' p
8
' p
After Clipping
( ) ( )
( ) ( )
0 0 end end
0 end 0 0 end 0
Parametric equations of line segment from , to ,
, , 0 1.
x y x y
x x u x x y y u y y u = + = + s s
Used to determine the parts contained in clipping window.
May 2010 9
Cohen-Sutherland Line Clipping Algorithm
Intersection calculations are expensive. Find first lines
completely inside or certainly outside clipping window.
Apply intersection only to undecided lines.
Perform cheaper tests before proceeding to expensive
intersection calculations.
May 2010 10
Cohen-Sutherland Line Clipping Algorithm
Assign code to every endpoint of line segment.
Borderlines of clipping window divide the plane into two halves.
A point can be characterized by a 4-bit code according to its
location in half planes.
Location bit is 0 if the point is in the positive half plane, 1
otherwise.
Code assignment involves comparisons or subtractions.
Completely inside / certainly outside tests involve only
logic operations of bits.
May 2010 11
max
xw x
0 1 0
0 1 0
0 1 0
0 0 0
0 0 0
1 1 1
min
y yw
1 1 1
0 0 0
0 0 0
max
yw y
0 0 1
0 0 1
0 0 1
min
x xw
If endpoint codes has 1 in same bit, line is certainly outside.
Endpoint codes are 0000 for both iff line is completely inside.
Top bit Bottom bit Right bit Left bit
May 2010 12
Lines that cannot be decided are intersected with window
border lines.
Each test clips the line and the remaining is tested again
for full inclusion or certain exclusion, until remaining is
either empty or fully contained.
Endpoints of lines are examined against left, right, bottom
and top borders (can be any order).
May 2010 13
0 0 0 0
0 0 1 0 0 0 1 1
0 0 0 1
0 1 0 0 0 1 0 1
2
p
1
p
2
' p
1
' p
3
p
4
p
2
'' p
3
' p
May 2010 14
( )
0 0
, x y
( )
end end
, x y
Liang-Barsky Line Clipping Algorithm
Define clipping window by intersections of four half-planes.
Treat undecided lines in Cohen-Sutherland more efficiently.
min
xw
+
min
yw
+
max
xw
+
max
yw
+
May 2010 15
( ) ( )
0 end 0 0 end 0
Parametric presentation:
, , 0 1. x x u x x y y u y y u = + = + s s
( )
( )
min 0 end 0 max
min 0 end 0 max
A point on the line is cotained in the clipping window iff:
,
.
xw x u x x xw
yw y u y y yw
s + s
s + s
1 0 end 1 0 min
2 end 0 2 max 0
3 0 end 3 0 min
4 end 0 4 max 0
It can be expressed by: , 1, 2, 3, 4, where
, ;
, .
, ;
, .
k k
up q k
p x x q x xw
p x x q xw x
p y y q y yw
p y y q yw y
s =
= =
= =
= =
= =
May 2010 16
( ) ( )
0 0 end end
In the inequality if 0 ( 0), the
traversal from , to , by increasing
from to + proceeds the line from the ( )
half-plane to ( ) one (with respect to the -th
border).
k k k k
up q p p
x y x y
u
k
s < >
+
+
( )
Intersection of , extension with -th border
occurs at .
k k
k
u q p
+
=
May 2010 17
0 end
0
end
We calculate and update and progressively for
1, 2, 3, 4 borders (left, right, bottom, top).
If 0 is calculated since progression is from
to half planes. Similarly, if 0 is c
k
k
u u
k
p u
p u
=
<
+ > alculated.
0 end
end 0
end 0
is the maximum among 0 and all . is the
minimum among 1 and all . The feasibility
condition is progressively checked. The line
is completely outside if .
k k
k k
u q p u
q p
u u
u u
>
<
May 2010 18
( )
Notice that doesn't need actual division since
comparison of with can be done by
comparison of with , and the quotient
can be stored as a pair ,
k k
q p
q p q p
q p q p q p
q p
' ' '' ''
' '' '' '
( ) ( )
end 0 0 0 end end
Only if , given by , and , ,
the actual ends of the clipped portion are calculated.
u u q p q p >
May 2010 19
This is more efficient than Cohen-Sutherland Alg,
which computes intersection with clipping window
borders for each undecided line, as a part of the
feasibility tests.
May 2010 20
Nicholl-Lee-Nicholl Line Clipping Algorithm
Creates more regions around clipping window to avoid
multiple line intersection calculations.
Performs fewer comparisons and divisions than
Cohen-Sutherland and Liang-Barsky, but cannot be
extended to 3D, while they can.
For complete inclusion in clipping window or certain
exclusion well use Cohen-Sutherland.
May 2010 21
0
P
0
P
0
P
0
Examine first where the starting point is located.
Only three regionsare considered. Location in any
of the other six regions can be handled by symmety
transformation.
P
May 2010 22
0
P
T
L
B
R
| |
end
0 end
The location of in each
region defines what edge the
line , is intersecting.
P
P P
May 2010 23
LT
L
LB
LR
L
L
0
P
| |
end
0 end
Else, is detected for being positioned in any of
LB, LR or LT, case where , is clipped with
left border and bottom, right or top border, resp.
P
P P
end
Detecting whether is in
any of regions L is immediate.
P
May 2010 24
Once one of LT, LR or LB regions is found,
intersection point with appropriate border is calculated.
| |
0 end end
, is entierely clipped if is positioned
outside the regions.
P P P
LT
L
LB
LR
L
L
0
P
end
P
end
P
| |
| |
0 end
0 corner
end
The slope of , is
compared to ,
for each corner to find the
region of .
P P
P P
P
May 2010 25
0
There are two cases, depending on whether is
closer to left or top borders.
P
TB
0
P
T
TR
T
L
LB
TR
0
P
T
LR
L
L
LB
May 2010 26
TB
0
P
T
TR
T
L
LB
TR
0
P
T
LR
L
L
LB
end
Notice that simple inclusion test of P in
clipping rectangle is not enough since there
are both T and L labels for the regions inside.
Testing of the angular regions is a must.
May 2010 27
Sutherland-Hodgman Polygon Clipping
2
1
3
1
3
1
3
2 2
Efficient algorithm for clipping convex polygons.
Edges are clipped against every border line of clipping window. Edges
are processed successively.
Allows pipelining of edge clipping of polygons, as well as pipelining of
different polygons.
1
3
1
3
2 2
Clipping
Clipping Window
May 2010 28
The four possible outputs generated by the left clipper, depending on
the relative position of pair of edge endpoints.
1
v
1
'
v
2
v
1
v
2
v
1
v
1
'
v
2
v
1
v
2
v
out in
1
'
v
2
v
output:
in in
2
v
output:
in out
1
'
v
output:
out out
output: none
May 2010 29
Input Left Clipper Right Clipper Bottom Clipper Top Clipper
[1,2]: (in-in)>{2}
[2,3]: (in-out)>{2} [2,2]:(in-in)>{2}
[3,1]: (out-in)>{3,1} [2,3]:(in-in)>{3} [2,3]:(in-out)>{2}
[3,1]:(in-in)>{1} [3,1]:(out-out)>{}
[1,2]:(in-in)>{2} [1,2]:(out-in)>{1,2} [2,1]:(in-in)>1}
[2,2]:(in-in)>{2} [1,2]:(in-in)>{2}
[2,2]:(in-in)>{2}
[2,2]:(in-in)>{2}
1
1
2
2
2
3
3
May 2010 30
The four clippers can work in parallel.
Once a pair of endpoints it output by the first clipper,
the second clipper can start working.
The more edges in a polygon, the more effective
parallelism is.
Processing of a new polygon can start once first
clipper finished processing.
No need to wait for polygon completion.
May 2010 31
Convex Polygon Intersection
P
Q
PQ
Theorem: The intersection of an L-vertex convex polygon and an M-vertex
convex polygon is a convex polygon of L+M vertices at most.
Proof: PQ is an intersection of L+M interior half planes determined by the
two polygons.
Intersection of convex polygons can answer the question of whether two
sets of points can be separated by a line.
May 2010 32
Theorem: The intersection of an L-vertex convex polygon and an M-vertex
convex polygon can be found in (L+M) time.
Proof: Polygons are given in cyclic order of vertices. We start from the
leftmost vertex of both polygons and progress along the border of both
in a left-to-right fashion, defining O(L+M) slabs.
Inside a slab each polygon forms a trapezoid. The intersection of two
trapezoids can be calculated in constant time.
May 2010 33
Another solution: The non empty intersection of two convex polygons
forms a sequence of sickles enclosing the intersection. The border of a
sickle comprises internal and external sequence of vertices originating
from P and Q, which are alternating in every sickle.
R=PQ
P
Q
sickle
Let P: (p
1
,p
2
,,p
L
) and Q: (q
1
,q
2
,,q
M
) be counterclockwise cyclically
ordered. The algorithm advances such that P and Q are chasing one
another, adjusting their speeds so that they meet at every intersection.
May 2010 34
( )
( )
( )
( )
1
1
Let and be the vertices where the traversal is.
and are the current edges, defining the half planes
and , containing and , respectively.
Clearly .
i j i i
j j
i j
i j
p q p p
q q
h p h q P Q
R h p h q

c
i
p
j
q
There are four possible situations of
and with respect to . For each
an advancing rule is in order. The
idea is to progress along boundary of
the "lagging" polygon.
i j
p q R
May 2010 35
1 1
Advance is from since there's no
chance for a future intersection point
in the edge . Edge may
have a future intersection with the
boundary of .
i
i i j j
p
p p q q
P

i
p
j
q
(a)
1
1
Edge may still be intersected
later by , while edge have
already exahusted its intersections.
Hence advance is from .
i i
j j
j
p p
Q q q
q

i
p
j
q
(b)
May 2010 36
Arbitrary choice. Advancing from
yields case (a), while advancing
from yields case (b).
j
i
q
p
1
1
Edge may still be intersected
later by , while edge have
already exahusted its intersections.
Hence advance is from .
j j
i i
i
q q
P p p
p

i
p
j
q
(c)
i
p
j
q
(d)
May 2010 37
( )
1 1
CONVEX_POLYGON_INTERSECT ( , ) {
1; 1; 1;
( 2 ) {
( and intersect) { report intersection }

// initialization
// apply one ADVANCE;
i i j j
void polygon P polygon Q
i j k
while k L M
if p p q q

= = =
< +
;
}
(no intersection reported) {
( ) { }
( ) { }

of (a), (b), (c)
{ }

or (d) case
}
}
s

i
j
k
if
if p Q P Q
elseif q P Q P
else Q P
+ +
e _
e _
= C
May 2010 38
May 2010 39
If and belong to the
same sickle, the intersection point terminating the sickle
must be found since algorithm never ADVANCE along
the polygon whose current edge may contain
i j
p q Correcness of algorithm:
a sought
intersection point.
This in turn guarantees that once an intersection point
of a sickle is found, all the others will be constructed
successively.
May 2010 40
To complete the proof we must show that an
intersection point must be found if . P Q = C
( )
1
1
Let the current edge on , contain an
intersection point with edge on , so
.
i i
r r
i
p p P
v q q Q
v h p

e
May 2010 41
s
C
r
C
v
i
p
1 i
p

r
q
1 r
q

s
q
farthest vertex
May 2010 42
v
s
C
r
C
i
p
1 i
p

r
q
1 r
q

s
q
farthest vertex
case (a)
case (c)
case (d)
case (b)
1
1
Let be current edge and . Sequence of
cases (a), (c), (d) and (b) (possibly empty) occurs, while
stays stationary,

j j j r
i i
q q q C
p p

e
until is found. v
May 2010 43
1
Let , be a line determined by and
supporting at , the first
support reached when traversing
from along .
j s j j
m
i
q C l q q
l l P p
p P

e
'
1 j
q

j
q
m
p
1 r
q

r
q
i
p
1 i
p

l
l
'
1
A sequence of cases (a)
starts at and will stay
so until crossing ,
where case (c)
holds up to (may be
empty). stays stationary.
i
m
j j
p
l
p
q q

May 2010 44
1 j
q

j
q
m
p
1 r
q

r
q
i
p
1 i
p

l
l
'
( )
( )
If ADVANCE
continues marching
along with cases (d)
and (b) until first
point
is found.
m j
t j
p h q
P
p h q
e
e
May 2010 45
1
If an edge like exists, steps must reach
it since the boundary of at least one polygon must be
consumed.
i i
p p L M

+
( )
If no intersection was found then the relations
, or can be resolved
in time.
P Q Q P P Q
O L M
_ _ = C
+
( )
Additional ADVANCE steps suffice to
obtain all the intersection points in cyclic order,
yielding 2 total steps of ADVANCE.
L M
L M
+
+
May 2010 46
3D Viewing Concepts
World Coordinate System
Viewing Coordinate System
May 2010 47
Choose viewing position, direction and orientation of the
camera in the world.
A clipping window is defined by the size of the aperture
and the lens.
Viewing by computing offers many more options which
camera cannot, e.g., parallel or perspective projections,
hiding parts of the scene, viewing behind obstacles, etc.
2D Reminder
May 2010 48
Clipping window: Selects what we want to see.
Viewport: Indicates where it is to be viewed on the output
device (still in world coordinates).
Display window: Setting into screen coordinates.
In 3D the clipping is displayed on the view plane, but
clipping of the scene takes place in the space by a clipping
volume.
3D transformation pipeline is similar to 2D with addition of
projection transformation.
May 2010 49
3D Viewing Transformation Pipeline
Transform Projection-
Coordinates to
Normalized-
Coordinates
Normalized
Coordinates
Map Normalized-
Coordinates to
Device-Coordinates
Device
Coordinates
Construct World-
Coordinate Scene
From Modeling-
Coordinate
Transformations
World
Coordinates
Modeling
Coordinates
Convert World-
Coordinates to
Viewing-
Coordinates
Viewing
Coordinates
Projection
Transformation
Projection Coordinates
May 2010 50
Model is given in model (self) coordinates.
Conversion to world coordinates takes place.
Viewing coordinate system which defines the position and
orientation of the projection plane (film plane in camera) is
selected, to which scene is converted.
2D clipping window (lens of camera) is defined on the
projection plane (film plane) and a 3D clipping, called view
volume, is established.
May 2010 51
The shape and size of view volume is defined by the
dimensions of clipping window, the type of projection and
the limiting positions along the viewing direction.
Objects are mapped to normalized coordinated and all
parts of the scene out of the view volume are clipped off.
The clipping is applied after all device independent
transformation are completed, so efficient transformation
concatenation is possible.
Few other tasks such as hidden surface removal and
surface rendering take place along the pipeline.
May 2010 52
World to Viewing 3D Transformation
world
x
world
y
world
z
view
x
view
y
view
z
v
n
u
0
p
view point
view-up vector
May 2010 53
( )
( )
( )
( )
0 0 0 0
: viewing directionon
: viewing plane
, ,
, ,
, ,
, ,
x y z
x y z
x y z
u u u
v v v
n n n
x y z

=
=
=
=
n
u v
u
v
n
p
view
x
view
y
view
z
v
n
u
0
p
view point
view-up vector
0
0
WC,VC
0
0 1 0 0
0 0 1 0
0 0 0 1
0 0 0 1 0 0 0 1
x y z
x y z
x y z
u u u x
v v v y
n n n z

( (
( (

( (
=
( (
( (

M
translation rotation
world to viewing
transformation
May 2010 54
Projection Transformations
Projection can be perpendicular or oblique to viewing plane.
Preserves relative size of objects portions.
Next step in 3D viewing pipeline is projection of object to
viewing plane
Parallel Projection
Coordinate are transferred
to viewing plane along
parallel lines.
View Plane
May 2010 55
Perspective Projection
Projection lines converge
in a point behind viewing
plane.
Doesnt preserve relative size but looks more realistic.
May 2010 56
Plane View
Side
Elevation
View
Front
Elevation
View
Used in engineering and architecture. Length and angles can be
measured directly from drawings.
Orthogonal (orthographic) projections
Projection lines are parallel to normal.
May 2010 57
Clipping Window and View Volume
Orthogonal Projection
View Volume
Far
Clipping
Plane
Near
Clipping
Plane
View
Plane
Clipping window
view
x
view
y
view
z
May 2010 58
Normalizing Orthogonal Projection
( )
1, 1, 1
( )
1,1,1
norm
z
norm
y
norm
x
Normalized View Volume
Display coordinate system is usually left-handed.
Orthogonal Projection View Volume
view
x
view
y
view
z
( )
min min near
, , xw yw z
( )
max max far
, , xw yw zw
May 2010 59
max min
max min max min
max min
max min max min
near far
near far near far
ortho,norm
2
0 0
2
0 0
2
0 0
0 0 0 1
xw xw
xw xw xw xw
yw yw
yw yw yw yw
z z
z z z z
=
+
(

(

(
+ (

(

(
(
+
(

(
(

M
WC,VC ortho,norm
The complete transformation from world coordinated
to normalized orthogonal-projection coordinates is
obtained by . M M
May 2010 60
Oblique Parallel Projections
Projection is defined by a viewing vector .
p
V
( )
, , x y z
( )
, ,
p p vp
x y z
( )
, ,
vp
x y z
L
p
V
Projection lines are
parallel to vector .
p
V
May 2010 61
The above is called shear transformation, where the
displacement in x and y linearly increases with z.
( )
, , x y z
( )
, ,
p p vp
x y z
( )
, ,
vp
x y z
L
p
V
( )( ) ( )( )
,
p vp px pz p vp py pz
x x z z V V y y z z V V = + = +
p py
vp pz
y y V
z z V

p px
vp pz
x x V
z z V

Not orthogonal
as viewed
May 2010 62
View
Volume
(Top)
View
Volume
(Side)
Clipping
Window
View Plane
Near
Plane
Far
Plane
p
V
p
V
View Volume of Parallel Oblique Projection
May 2010 63
oblique
1 0
0 1
0 0 1 0
0 0 0 1
px px
vp
pz pz
py py
vp
pz pz
V V
z
V V
V V
z
V V
(

(
(
(
( =
(
(
(
(

M
This is a 3D shear
transformation.
x and y are displaced by
amount proportional to z.
Normalization oblique projection is similar to orthogonal projection.
The composite transformation is obtained by the product of the two.
oblique,norm ortho,norm oblique
= M M M
May 2010 64
Perspective Projections
Projection
Reference
Point
View Plane
Closer objects look larger.
( )
, , x y z = p
( )
, ,
prp prp prp prp
x y z = p
( )
, ,
p p vp
x y z
view
x
view
y
view
z
View Plane
Projection Point
May 2010 65
( )
( )
( ) ( )
( )
Parametric representation of a point on the line
connecting , , with , , :
, ,
, 0 1.
prp prp prp prp
prp prp
prp
x y z x y z
x x x x u y y y y u
z z z z u u
= =
' '
= =
'
= s s
P P
( ) ( )
At viewing plane: .
vp prp
u z z z z =
( ) ( ) ( ) ( )
( ) ( ) ( ) ( )
Substitution for the point on viewing plane:
p prp vp prp prp vp prp
p prp vp prp prp vp prp
x x z z z z x z z z z
y y z z z z y z z z z
= +
= +
May 2010 66
The problem with the above representation is that Z appears in
denominator, so matrix multiplication representation of X and Y
on view plane as a function of Z is not straight forward.
Z is point specific, hence division will be computation killer.
Different representation is in order, so transformations can
be concatenated
May 2010 67
Vanishing Points
Vanishing points occur when the viewing plane intersects with the
axes of viewing coordinate system.
Vanishing
Point
One-Point perspective Projection
x
y
z
Principle axes for cube
Parallel to Z lines of XZ plane and parallel lines to Z in YZ plane will
vanish. Vanishing point on viewing plane correspond to infinity in world.
May 2010 68
Vanishing points of all three axes occur when viewing plane
intersects all three axes.
x-axis vanishing point
z-axis vanishing point
Viewing plane is parallel to y-axis,
intersecting both x-axis and z-axis
May 2010 69
Transformation Matrix
( )
To get rid of in denominator we define parameter
and new homogeneous coordinates
, , , .
prp
h h h
z
h z z
x y z h
=
, , yielding:
p h p h
x x h y y h = =
( ) ( )
( ) ( )
,
.
h prp vp prp vp
h prp vp prp vp
x x z z x z z
y y z z y z z
= +
= +
May 2010 70
( ) ( )
pers
, , ,1 , , , , , ,
h h h h h
x y z x y z h = = = P P P M P
is using a scaling factor and translation .
and can be set arbitrarily. We'll set those to satisfy
desired normalization.
h z z z z
z z
z zs t s t
s t
= +
Matrix Representation:
pers
0
0
0 0
0 0 1
prp vp prp prp prp
prp vp prp prp prp
z z
prp
z z x x z
z z y y z
s t
z

(
(

(
=
(
(

(

M
h z z
z zs t = +
May 2010 71
Transformed coordinate is very useful for deciding
later on the hidden parts of the scene.
z
Notice that drawing the perspective point requires a
division by . depends on coordinate, hence evey
point must be divided by a different number, w
very e
hic
xpe
h is
ve! nsi
h h z
May 2010 72
Perspective-Projection View Volume
Rectangular Frustum
View Volume
Far
Clipping
Plane
Near
Clipping
Plane
View
Plane
Clipping Window
view
x
view
y
view
z
Projection
Reference Point
u
Field-of-view Angle
May 2010 73
Perspective
Mapping
Clipping Window
View Plane
Near Plane
Far Plane
Symmetric Frustum
View Volume
Projection
Reference Point
Parallel Piped
View Volume
May 2010 74
Perspective
Mapping
Parallel Piped
View Volume
Oblique Frustum
View Volume
min
xw
max
xw
Projection
Reference Point
May 2010 75
Oblique Perspective-Projection Frustum
( ) ( )
We'll take the projection reference point to be the origin of viewing
coordinate system , , 0, 0, 0 and viewing plane at near
clipping plane. Oblique projection results shear transformation
prp prp prp
x y z =
( )
min max min max
near near
. It
transforms the intersection point of center-line with clipping window
, , to 0, 0, .
2 2
xw xw yw yw
z z
+ +
| |
|
\ .
z shear
1 0 sh 0
0 1 sh 0
0 0 1 0
0 0 0 1
zx
zy
(
(
(
=
(
(

M
( )
( )
min max
min max
z shear
naer naer
0 2
0 2
1 1
xw xw
yw yw
Z Z
+ ( (
( (
+
( (
=
( (
( (

M
May 2010 76
min max min max
which solves to sh and sh .
2 2
zx zy
xw xw yw yw + +
= =
( ) ( )
near
near
pers
Substituting , , 0, 0, 0 in the perspective-projection
matrix and positioning the viewing near clipping plane at the view
plane simplifies it to
0 0 0
0 0 0
0 0
0 0 1 0
prp prp prp
z z
x y z
z
z
z
s t
=

(
(

(
=
(
(

M
The coordinate scaling factor and the translation will be
determined by the normalization requirements.
z z
z s t
May 2010 77
( )
( )
near min max
near min max
obliquepers pers z shear
The complete oblique perspective-projection is obtained by
concatenating the perspective and shear matrices.
0 2 0
0 2 0
0 0
0 0 1 0
z z
z xw xw
z yw yw
s t
+ (
(
+
(
= =
(
(

M M M
( )
min max min max
near
near
symmetricpers pers z shear
For symmetric viewing volume , .
Transformation is simplified to:
0 0 0
0 0 0
0 0
0 0 1 0
z z
xw xw yw yw
z
z
s t
= =

(
(

(
= =
(
(

M M M
May 2010 78
Normalized Perspective-Projection
( )
( )
normpers scale obliquepers
near min max
near min max
Normalization is obtained by multiplying with standard scaling matrix.
0 2 0
0 2 0
0 0
0 0 1 0
xy
x x
y y
z z
z s s xw xw
z s s yw yw
s t
=
+ (
(
+
(
=
(
(

M M M
normpers
Homogenous coordinates

are obtained as follows:
1
h
h
h
x x
y y
z z
h
( (
( (
( (
=
( (
( (

M
May 2010 79
( )
( )
( )
near min max
near min max
Projection coordinates are:
2 ,
2 ,
.
p h x x
p h y y
p h z z
x x h z s x s xw xw z z
y y h z s y s yw yw z z
z z h s z t z
= = + + (

( = = + +

= = +
( ) ( ) ( ) ( )
( ) ( )
( ) ( )
min min near max max far
max min max min
near far near far
We'd like normalization transform to result:
, , 1, 1, 1 and , , 1,1,1 .
Substitution in above equations yields:
2 , 2 ,
, t 2
x y
z z
xw yw z xw yw z
s xw xw s yw yw
s z z z z

= =
= + =
( )
near far near far
. z z z z
May 2010 80
near min max
max min max min
near min max
max min max min normpers
near far near far
near far ne
Back substitution in the normalized perspective transformation yields:
2
0 0
2
0 0
2
0 0
z xw xv
xw xw xw xw
z yw yw
yw yw yw yw
z z z z
z z z
+

+
=
+

M
ar far
0 0 1 0
z
(
(
(
(
(
(
(
(

(
(

No division by z!
May 2010 81
3D Viewport Transformation
( )
( ) ( )
( )
min min
max max
The normalized view volume cube extending from 1, 1, 1 to
1,1,1 is mapped to a screen viewport, extending from ,
to , . information is stored for depth calculations. is
often
xv yv
xv yv z z

max min max min
max min max min
normviewvol,
3D screen
renormalized to the range from 0 to 1.0, yielding:
0 0
2 2
0 0
2 2
1 1
0 0
2 2
0 0 0 1
xv xv xv xv
yv yv yv yv
+
(
(
(
+
(
(
=
(
(
(
(

M
May 2010 82
Settings of Perspective Projection
Perspective projection point
Where the viewer (camera, eye) is positioned in the world.
Positioning viewing plane with respect to viewing
coordinates
Results vanishing points, one, two or three.
Clipping window on viewing plane
Defines the infinite pyramid view volume.
Near and far clipping planes (parallel to view plane)
Define the rectangular frustum view volume.
Scale and translation parameters of perspective matrix
Define the normalization range.
May 2010 83
3D Clipping
min min min max max max
Clipping can take place on normalized cube:
1, 1, 1, 1, 1, 1. xw yw zw xw yw zw = = = = = =
Top bit Bottom bit Right bit Left bit Far bit Near bit
Similar to 2D, we add two bits to code the far and near planes.
( )
Recall that in 3D point is represented in homogeneous form. Hence
, , , is inside the normalized cube iff
, , if 0
, , if 0
h h h
h h h
h h h
x y z h
h x h h y h h z h h
h x h h y h h z h h
=
s s s s s s >
s s s s s s <
P
May 2010 84
Left
Right
Top
Bottom
Far
Near
x
z
y
011001 011000 011010
010001 010000 010010
010101 010100 010110
001001 001000 001010
000001 000000 000010
000101 000100 000110
101001 101000 101010
100001 100000 100010
100101 100100 100110
May 2010 85
A line is completely accepted if the codes of its ends are both 000000, or
equivalently, if the logical OR between codes is zero.
A line is rejected if the codes of its ends has at least one 1 in same bit, or
equivalently, if the logical AND between codes is nonzero.
Otherwise, the line is tested against each of the planes and the 2D Liang-
Barsky algorithm can be extended to 3D.
( )
( )
( )
1 1 1
2 2 2
1 2 1 1
2 2 1 2 1
A point of a line segment extending from , , ,
to , , , is given by , 0 1.
h h h
h h h
x y z h
x y z h u u
=
= = + s s
P PP P
P P P P P
May 2010 86
( )
( )
1 2 1
1 2
max
1 2 1
If for instance the codes of the two end points of w.r.t the right
clipping plane 1 are different, the intersection point is derived
from 1.
Solving for
p h h h h
x
x x h x x x u h h h u
u
=
(
= = + + = (


PP
( ) ( )
1 1 2
1 1 2
yields .
h h h
u x h x h x h
(
(
=


0 1 1 0
1 0
Such calculation proceeds for each , 1 6, while
updating and . If at some iteration , line segment is
completely clipped. If upon termination , the end points
of the clipped line s
k k
up q k
u u u u
u u
s s s
s
>
egment are obtained.