Ray Tracing (Week 4a)

Komputer Grafik 2 (AK045206)
Ray Tracing
1/38
Ray Tracing
Ray Tracing
2/38
Outline
Review Illumination
Render dengan ray tracing
Pembentukan sinar
Interseksi sinar-obyek
Interseksi world space
Vektor normal pada titik interseksi
Transformasi vektor normal
Ray tracing rekursif
Bayangan
Permukaan tembus (transparan)

Ray Tracing
3/38
Review Illumination
Global, physically-based illumination models
describe energy transport and radiation
subject to the properties of light and materials
subject to the geometry of light, objects and
viewer

For each surface, there is a distribution that
characterizes its absorption and reflection at
each wavelength

All illumination models are, by definition,
approximate
based on sampling geometry, light distribution
and material properties, and taking rendering
shortcuts

Ray Tracing
4/38
Why Trace Rays?
More elegant than polygon
rasterization, especially for
sophisticated physics
Testbed for techniques and
effects
modeling (reflectance, transport)
rendering (e.g. Monte Carlo)
texturing (e.g. hypertexture)
Easiest photorealistic renderer
to implement
Ray Tracing
5/38
Ray Tracing
Ray tracing allows to add more
realism to the rendered scene
by allowing effects such as
shadows
transparency
reflections

Ray Tracing
6/38
Raytracing, a simple idea: instead of forward
mapping an infinite number of rays from light
source to object to viewer, backmap finite number
of rays from viewer through each pixel to object to
light source (or other object)

Raytracing: shoot rays from eye through sample
point (e.g., a pixel center) of a virtual photo
compute what color/intensity you see at the
corresponding sample point on the object out there;
paint this on photo at the ray-photo intersection
point
What you see out there
combination of the color of the material of the
surface element you hit first and the colors of light
that hit it and reflect towards you

Ray Tracing
Ray Tracing
7/38
Rendering with Raytracing
Subproblems to solve
Generate primary (eye) ray
Find closest object along ray path
find the first intersection of a ray that goes out from
the eye through a pixel center (or any other sample
point on image plane) into the scene
Light sample
use illumination model to determine light at closest
surface element
generate secondary rays from object

Ray Tracing
8/38
Ray origin
Lets look at the geometry of the problem in
untransformed world-space
We start a ray from an eye point: P
We send it out in some direction d from your eye
toward a typical point on the film plane whose color
we want to know
Points along the ray have the form P + td where
- P is the base point of the ray: the cameras eye
point
- d is the unit vector direction of the ray
- t is a nonnegative real number
The eye point is the center of projection in the
perspective view volume (we use perspective
coords to avoid having to deal with the inverse of a
perspective transformation later)
Generating Rays (1/3)

Ray Tracing
9/38
We start raycasting with screen-space points (pixels).
Need to find a point in 3-D that lies on the
corresponding ray
we will use ray to intersect with the original objects in the
original, untransformed world coordinate system
Transform 2D screen-space points into points on the
cameras 3D film plane
Any planes orthogonal to the look vector is a
convenient film plane: constant z in canonical view
volume

Choose a plane to be the film plane and then create a
function that maps screen-space points onto it
whats a convenient plane? Try the far plane :-)
to convert, we just have to scale integer screen-space
coordinates into floating point values between -1 and 1


Ray direction
Ray Tracing
10/38
Transform film plane point into world-space point
- we can make the direction vector between the
eye and this world-space point
- we need direction to be in world space so that we
can intersect with the original object in the original,
untransformed world coordinate system

Normalizing transformation takes world-space
points to points in the canonical view volume
- translate to the origin, orient with the axes, scale x
and y to adjust viewing angles, scale so that z: [-1,
0]; x,y: [-1, 1]
Apply the inverse of the normalizing
transformation: the Viewing Transformation

Ray direction (cont.)

Ray Tracing
11/38
If an object is defined implicitly by a function f
such that f(Q) = 0 if and only if Q is a point on the
surface of the object, then ray-object intersection
is comparatively easy
we can define many objects implicitly
implicit functions give you potentially infinite
resolution
tessellating implicit functions is more difficult than
using them directly
For example, a circle of radius R is an implicit
object in the plane, and its equation is

points where f(x, y) = 0 are points on the circle
An infinite plane is defined by the function:

A sphere of radius R in 3-space:

Implicit objects
Ray Object Intersection (1/5)

(A,B,C)
(0,0,0)
-D
2 2 2
) , ( R y x y x f + =
D Cz By Ax z y x f + + + = ) , , (
2 2 2 2
) , , ( R z y x z y x f + + =
Ray Tracing
12/38

At what points (if any) does the ray intersect the
object?
Points on the ray have the form P + td
t is any nonnegative real
A point Q lying on the object has the property that
f(Q) = 0
Combining, we want to know For which values of
t is f(P + td) = 0 ? (if any)

We are solving a system of simultaneous
equations in x, y (in 2D) or x, y, z (in 3D)

Implicit objects(cont.)
Ray Tracing
13/38
An Explicit Example (1/3)
Consider the eye-point P = (-3, 1), the
direction vector d = (.8, -.6) and the unit
circle given by:

A typical point of the ray is:

Plugging this into the equation of the
circle:

Expanding, we get:

Setting this to zero, we get:

1 ) , (
2 2
+ = y x y x f
) 6 . 1 , 8 . 3 ( ) 6 . , 8 (. ) 1 , 3 ( t t t t P Q + = + = + = d
1 ) 6 . 1 ( ) 8 . 3 ( ) 6 . 1 , 8 . 3 ( ) (
2 2
+ + = + = t t t t f Q f
1 36 . 2 . 1 1 64 . 8 . 4 9
2 2
+ + + t t t t
0 9 6
2
= + t t
2D ray-circle intersection example
Ray Tracing
14/38
2D ray-circle intersection example (cont.)
Using the quadratic formula:

We get:

Because we have a root of multiplicity 2, our
ray intersects the circle at one point (i.e., its
tangent to the circle)
We can use the discriminant D = b
2
- 4ac to
quickly determine if a ray intersects a curve or
not
- if D < 0, imaginary roots; no intersection
- if D = 0, double root; ray is tangent
- if D > 0, two real roots; ray intersects circle at two
points
Smallest non-negative real t represents
intersection nearest to eye-point
a
ac b b
roots
2
4
2

=
3 , 3 ,
2
36 36 6
=

= t t
Ray Tracing
15/38
2D ray-circle intersection example (cont.)
Generalizing: our approach will be to take an
arbitrary implicit surface with an equation f(Q)
= 0 and a ray P + td, and plug the latter into
the former:

This results, after some algebra, in an
equation with t as unknown
We then solve for t, analytically or numerically
0 ) ( = + d t P f
Ray Tracing
16/38

Implicit objects-multiple conditions
For objects like cylinders, the equation

in 3-space defines an infinite cylinder of unit
radius, running along the y-axis Usually, its more
useful to work with finite objects, e.g. such a unit
cylinder truncated with the limits

But how do we do the caps?
The cap is the inside of the cylinder at the y
extrema of the cylinder

0 1
2 2
= + z x
1
1
>
s
y
y
1 , 0 1
2 2
= < + y z x
Ray Tracing
17/38

Multiple conditions (cont.)
Picture

We want intersections satisfying the cylinder:

or top cap:

or bottom cap:

1 1
0 1
2 2
s s
= +
y
z x
1
0 1
2 2
=
s +
y
z x
1
0 1
2 2
=
s +
y
z x
Ray Tracing
18/38
Turn an implicit surface equation into an equation
for t and solve
The thing you see first from eye point is at the
smallest non-negative t-value

For complicated objects (not defined by a single
equation), write out a set of equalities and
inequalities and then code as cases

Latter approach can be generalized cleverly to
handle all sorts of complex combinations of
objects
- Constructive Solid Geometry (CSG), where objects
are stored as a hierarchy of primitives and 3-D set
operations (union, intersection, difference)
- blobby objects, which are implicit surfaces defined
by sums of implicit equations

Implicit surface strategy summary
Ray Tracing
19/38
World-Space Intersection

World space - a global view
One way to do intersection is to compute the
equation of your object in world-space
Example: a sphere translated to (3, 4, 5) after it
was scaled by 2 in the x-direction has equation

One can then take the ray P+td and plug it into
the equation

Solve the resulting equation for t
Always start with the untransformed object
definition in its own coordinate system
We then try to derive what the transformed
version should be, given the CTM (call it M)
this is not easy for general transformations, and
furthermore, the transformed version of the
equation is always more complicated and thus
more expensive to compute
can we just work with the object in its own
coordinate system? Yes, as follows:
1 ) 5 ( ) 4 (
2
) 3 (
) , , (
2 2
2
2
+ +
= z y
x
z y x f
0 ) ( = + d t P f
Ray Tracing
20/38
Object-Space Intersection

Transform ray into object-space

Express world-space point of intersection as MQ,
where Q is some point in object-space:

If is the equation of the untransformed
object, we just have to solve

- note d is probably not a unit vector
- the parameter t along this vector and its world
space counterpart always have the same value.
Normalizing d would alter this relationship.
Do NOT normalize d .
d d
d
d
d
1 1
1 1
1
~
,
~
Let
) (

= =
= +
= +
= +
M P M P
Q tM P M
Q t P M
MQ t P
) , , (
~
z y x f
0 )
~ ~
(
~
= + d t P f
(0,0,0)
MQ
d t P+ d d
1 1
~ ~

+ = + tM P M t P
0 ) , , ( = z y x f
Q
0 ) , , (
~
= z y x f
Ray Tracing
21/38
World-Space vs. Object-Space

Use object-space if objects defined there
To compute world-space intersections, you have to
transform the implicit equation of your canonical object
defined in object space - often difficult
But to compute intersections in object-space, you need
only apply a matrix (M-1) to P and d, which is much
simpler.
does M-1 exist?
M was composed from two parts: the cumulative modeling
transformation that positions the object in world-space,
and the cameras normalizing transformation
the modeling transformations are just translations,
rotations, and scales (all invertible)
the normalizing transformation also consists of
translations, rotations and scales (also invertible); no
perspective! (Now you see why we used the hinged
perspective view volume)
When youre done, you get a t-value
This t can be used in two ways:
P + td is the world-space location of the intersection of the
ray with the transformed object
P + td is the corresponding point on the
untransformed object (in object space)

Ray Tracing
22/38
Normal Vectors at
Intersection Points (1/4)

Normal vector to implicit surfaces
For illumination we need to find a way, given a
point on the object in object-space, to determine
the normal vector to the object at that point so that
we can calculate angles between the normal and
other vectors
If a surface bounds a solid whose interior is given
by

then we can find the normal vector at a point(x, y,
z) via the gradient at that point:

Recall that the gradient is a vector with three
components:

0 ) , , ( < z y x f
) , , ( z y x f n V =
|
|
.
|
\
|
c
c
c
c
c
c
= V ) , , ( ), , , ( ), , , ( ) , , ( z y x
z
f
z y x
y
f
z y x
x
f
z y x f
Ray Tracing
23/38
Normal Vectors at

Sphere normal vector example
For the sphere, the equation is

The partial derivatives are

So the gradient is

Normalize n before using in dot products!
In some degenerate cases, the gradient may be
zero, and this method failsuse nearby gradient
as a cheap hack
1 ) , , (
2 2 2
+ + = z y x z y x f
z z y x
z
f
y z y x
y
f
x z y x
x
f
2 ) , , (
2 ) , , (
2 ) , , (
=
c
c
=
c
c
=
c
c
) 2 , 2 , 2 ( ) , , ( z y x z y x f n = V =
Ray Tracing
24/38
Normal Vectors at
Transforming back to world-space
We have an object-space normal vector
We want a world-space normal vector
To transform the object to world coordinates, we
just multiplied its vertices by M, the objects CTM
Can we treat the normal vector the same way?
Answer: NO

Example: say M scales in x by .5 and y by 2

See the normal scaling applets in
AppletsLighting and Shading

object world
Mn n =
n
object
Mn
object
Wrong!
Normal must be
perpendicular
Ray Tracing
25/38
Normal Vectors at

Why doesnt multiplying the normal work?
For translation and rotation, which are rigid body
transformations, it actually does work
Scale, however, distorts the normal in exactly the
opposite sense of the scale applied to the surface,
not surprising given that the normal is
perpendicular
scaling y by 2 causes normal to scale by .5:

Well see this algebraically in the next slides

x y
2
1
: normal =
x y 2 =
x y
4
1
: normal =
x y 4 =
Ray Tracing
26/38
Transforming Normals (1/4)

Object-space to world-space
Take the polygonal case, for example
Lets compute the relationship between the
object-space normal nobj to a polygon H and the
world-space normal nworld to a transformed
version of H, say MH
If v is a vector in world space that lies in the
polygon (e.g., one of the edge vectors), then the
normal is perpendicular to the vector v:

But vworld is just a transformed version of some
vector in object space, vobj. So we could write

But wed be wrong, because v is a tangent vector
its transformed by the derivative
which includes a 3D->4D->3D mapping

0 = -
world world
v n
0 = -
obj world
Mv n
0
3
= -
=
obj world
M v n
M M of 3x3 left upper
3
Ray Tracing
27/38

Object-space to world-space(cont.)
So we want a vector nworld such that

We will show in a moment that this equation can
be rewritten as

We also already have

Therefore

Left-multiplying by (Mt)-1,

0
3
= -
obj world
M v n
obj world
t
M n n =
3
0
3
= -
obj world
t
M v n
0 = -
obj obj
v n
( )
obj
t
world
M n n
-1
3
=
Ray Tracing
28/38

Object-space to world-space(cont.)
So how did we rewrite this:

As this:

Recall that if we think of vector as nx1 matrices,
then switching notation,

Rewriting the formula above, we have

Writing M = Mtt, we get

Recalling that (AB)t = BtAt, we can write this:

Switching back to dot product notation:

0 ) (
3
=
obj
t
world
t
M v n
0
3
= -
obj world
M v n
b a b a
t
= -
0
3
=
obj
t
world
M v n
0
3
=
obj
tt
t
world
M v n
0 ) (
3
= -
obj world
t
M v n
0 ) (
3
= -
obj world
t
M v n
Ray Tracing
29/38

Applying inverse-transpose of M to normals
So we ended up with

Fortunately, invert and transpose can be swapped,
to get our final form:

Why do we do this? Its easier!
A hand-waving interpretation of (M3-1)t
M3 is a composition of rotations and scales, R and S (why
no translates?). Therefore

so were applying the transformed (inverted, transposed)
versions of each individual matrix in their original order
for a rotation matrix, this transformed version is equal to
the original rotation, i.e., normal rotates with object
(R-1)t=R; the inverse reverses the rotation, and the transpose
reverses it back
for a scale matrix, the inverse reverses the scale, while
the transpose does nothing:
(S(x,y,z)-1)t = S(x,y,z)-1=S(1/x,1/y,1/z)

obj
t
world
M n n
1
3
) (

=
obj
t
world
M n n ) (
1
3
=
...) ) ( ) (( ) (... ) ...) ((
1 1 1 1 1 t t t t
S R R S RS

= =
Ray Tracing
30/38
Summary :
Non-Recursive Ray Tracer

P = eyePt

For each sample of image

Compute d

for each object
Intersect ray P+td with object

Select object with smallest non-negative t-value(this is
the visible object)

For this object, find the object-space intersection point

Compute the normal at that point
Transform the normal to world space

Use the world-space normal for lighting computations

Ray Tracing
31/38
Recursive Raytracing

Simulating global lighting effects (Whitted, 1979)
By recursively casting new rays into
the scene, we can look for more
information
Start from the point of intersection
Wed like to send rays in all
directions
Send rays in directions we predict
will contribute the most:
toward lights (shadow)
bounce off objects (specular reflection)
through the object (transparency)

Ray Tracing
32/38
Shadows

Each light in scene makes contribution to the
color and intensity of a surface element

Iff it reaches the object!
- could be occluded by other objects in the scene
- could be occluded by another part of the same
object
Construct a ray from the surface to each light
Check to see if ray intersects other objects before
it gets to the light
- if ray does NOT intersect that same object or
another object on its way to the light, then the full
contribution of the light can be counted
- if ray does intersect an object on its way to the
light, then no contribution of that light should be
counted
- what about transparent objects?

=
+
+ =
numlights
light

specular] [dif f use sity lightInten n attenuatio

ambient nsity objectInte
1
Ray Tracing
33/38
Transparent Surfaces (1/2)

Nonrefractive transparency
For a partially transparent polygon

2 polygon for calculated intensity
1 polygon for calculated intensity
1 polygon of nce transmitta
=
=
=
+ =
2
1
1
2 1 1 1
) 1 (

I
I
k
I k I k I
t
t t
I
2
I
1
polygon 1
polygon 2
Ray Tracing
34/38
Transparent Surfaces (2/2)

Refractive transparency
We model the way light bends at interfaces with
Snells Law

2 medium of refraction of index
1 medium of refraction of index
=
=
=
q
q
q
q u
u
t
i
t
i i
t
sin
sin
Ray Tracing
35/38
Recursive Raytracing (1/2)

Trace secondary rays at intersections:
- light: trace ray to each light source. If light source
blocked by opaque object, it does not contribute to
lighting, and object in shadow of blocked light at that
point
- specular reflection: trace ray in mirror direction by
reflecting about N
- refraction/transparency: trace ray in refraction
direction by Snells law
- recursively spawn new light, reflection, refraction
rays at each intersection until contribution negligible
Your new lighting equation

note: intensity from recursed rays calcd with same eqn
light sources contribute specular and diffuse lighting
Limitations
- recursive interobject reflection is strictly specular:
along ray reflected about N
- diffuse interobject reflection handled by radiosity
Informative videos: silent movie on raytracing, A
Long Rays Journey into Light

recursive
refracted
t t t
reflected
s s
m
specular
n
s s
diffuse
d d p att
ambient
d ka a
I O k O k V R O k L N O k I f O I I

+ + - + - + =

] ) ( [
Ray Tracing
36/38
Recursive Raytracing (2/2)

Light-ray Trees
Ray Tracing
37/38
Raytracing Pipeline

Raytracer produces visible samples from
model
samples convolved with filter to form pixel image
Additional pre-processing
pre-process database to speed up per-sample
calculations (done once, sometimes must be
redone if objects transformed (resize, translate))

Ray Tracing
38/38
Referensi
F.S.Hill, Jr., COMPUTER GRAPHICS
Using Open GL, Second Edition,
Prentice Hall, 2001

Andries van Dam, Introduction to
Computer Graphics, Slide-
Presentation, Brown University, 2003,
(folder : brownUni)

______, Interactive Computer Graphic,
Slide-Presentation, (folder :
Lect_IC_AC_UK)

Ray Tracing (Week 4a)

Загружено:

Сведения о документе

Исходное описание:

Оригинальное название

Авторское право

Доступные форматы

Поделиться этим документом

Поделиться или встроить документ

Параметры публикации

Этот документ был вам полезен?

Это неприемлемый материал?

Авторское право:

Доступные форматы

Ray Tracing (Week 4a)

Загружено:

Авторское право:

Доступные форматы

Komputer Grafik 2 (AK045206)

specular] [dif f use sity lightInten n attenuatio

Вам также может понравиться