GPU编程与 CG语言中文版 PDF

GPU Programming And Cg Language Primer 1rd Edition
GPU CG
kang_yu_zhi@yahoo.cn
2009 9
GPU CG
GPU Programming And Cg Language Primer 1rd Edition

kang_yu_zhi@yahoo.cn
2009 9
3

26
4
7
17
IT
5

5

GPU
Cg NVIDIA CgUsersManual
7
1 ............................................................................................................13
1.1 Programmable Graphics Processing Unit ....................................13
1.2 GPU VS CPU..................................................................................................15
1.3 ...........................................................................................18
1.4 ...................................................................................18
2 GPU .......................................................................................21
2.1 .......................................................................................................22
2.1.1 object space world space............................................................23
2.1.2 world space eye space................................................................24
2.1.3 eye space project and clip space ................................................25
2.2 Primitive Assembly && Triangle setup .........................................................26
2.3 ...................................................................................................27
2.3.1 Rasterization .........................................................................................27
2.3.2 Pixel Operation .....................................................................................28
2.4 .......................................................................................................30
2.4.1 GPU .....................................................................................30
2.4.2 Z Buffer Z ...................................................................................31
2.4.3 Stencil Buffer ........................................................................................33
2.4.4 Frame Buffer .........................................................................................34
2.5 ........................................................................................................35
3 Shader Language............................................................................................36
3.1 Shader Language ..................................................................................37
3.2 Vertex Shader Program.................................................................................39
3.3 Fragment Shader Program ............................................................................40
3.4 CG VS GLSL VS HLSL.................................................................................41
3.5 .......................................................................................................43
4 Cg ..................................................................................................44
4.1 Cg ................................................................................................44
4.2 CG ........................................................................................................45
8
4.3 CG .......................................................................................................45
4.3.1 CG ........................................................................................45
4.3.1 CGC .....................................................................................47
4.4 CG Profiles....................................................................................................50
5 CG ................................................................................................53
5.1 ...............................................................................................53
5.2 .......................................................................................................55
5.3 .......................................................................................................56
5.4 Interfaces...............................................................................58
5.5 .......................................................................................................58
6 CG ................................................................................60
6.1 Comparison Operators ......................................................60
6.2 Logical Operators..............................................................61
6.3 Math Operators .................................................................62
6.4 ...................................................................................................62
6.5 Swizzle ............................................................................................63
6.6 Conditional Operators........................................................64
6.7 ...........................................................................................65
6.8 Control Flow Statement .....................................................66
7 \..................................................................................67
7.1 Cg .....................................................................................................67
7.2 uniform..........................................................................................................68
7.3 const ..............................................................................................................69
7.4 \in\out\inout..............................................................69
7.5 SemanticBinding Semantics ..........................70
7.5.1 ...............................................................71
7.5.2 ...................................................................72
7.5.3 ...................................................................73
7.5.4 ...................................................................75
9
7.5.5 .......................................................................................75
8 ............................................................................................79
8.1 ...............................................................................................................79
8.1.1 ...............................................................................................80
8.2 ........................................................................................................81
8.3 .......................................................................................................82
8.4 CG ............................................................................................83
8.4.1 Mathematical Functions................................................84
8.4.2 Geometric Functions.....................................................87
8.4.3 Texture Map Functions .........................................88
8.4.4 Derivative Functions.....................................................91
8.4.5 Debugging Function......................................................94
8.5 ................................................................................94
...............................................................................................................96
9 illumination model .........................................................97
9.1 ...............................................................................................................97
9.2 Lambert ...............................................................................98
9.2.1 ..........................................................................................99
9.3 Phong .............................................................................102
9.3.1 phong ..................................................................................103
9.4 Blinn-Phong ...............................................................................107
9.5 Rendering Equation ........................................................109
9.6 .....................................................................................................110
10 ............................................................................................ 111
10.1 Cook-Torrance .........................................................................112
10.1.1 Cook-Torrance .................................................116
10.2 BRDF .......................................................................................118
10.2.1 BRDF ...............................................................118
10.2.2 ............................................................................119
10
10.3 Bank BRDF ................................................................................120
10.4 ...................................................................................................123
11 ........................................................................124
11.1 Snell Fresnel ........................................................................125
11.1.1 Snell ......................................................................125
11.1.2 ................................................................................................126
11.1.3 Fresnel ...................................................................................127
11.2 ....................................................................................................128
11.3 ...................................................................................132
11.4 ...........................................................136
12 Projective Texture Mapping.......................................137
12.1 ...............................................................................137
12.2 Homogeneous Texture Coordinates..........................139
12.3 ........................................................................................139
12.4 ...................................................................................................143
13 Shadow Map ..............................................................................................144
13.1 depth map.....................................................................................144
13.2 Shadow map shadow texture ..................................................145
13.3 Shadow map .................................................................146
14 Volume Rendering .........................................................152
14.1 ...............................................................................153
14.2 .......................................................................................153
14.3 ...................................................................................154
14.4 Volume Data .........................................................................155
14.4.2 Voxel ................................................................................156
14.4.1 Volume Texture ............................................................157
14.5 ...............................................................................................159
15 Ray Casting................................................................160
15.1 ...................................................................................160
11
15.1.1 ...........................................................................................161
15.2 ...................................................................161
15.2.1 ......................................................................161
15.2.2 ..................................................................................163
15.2.3 ..............................................................................164
15.2.2 ..........................................................166
15.3 ...................................................................................................168
15.4 ...................................................................................169
15.5 ...................................................................................................172
A ................................................................................................173
B.............................................................................175
CStencil Shadow ............................................................178
....................................................................................................................182
Alice
Alice

Alice

-------Lewis Carroll Alices adventures in Wonderland

12
1

------Henry David Thoreau
1.1 Programmable Graphics Processing Unit
Programmable Graphics Processing UnitGPU
GPU 20 70 80
monolithic GPU
20 80 90
digital signal processor chip GPU
1991
S3 Graphics 2D 1995 PC
2D
fixed-function Windows accelerators
PC
1998 NVIDIA modern GPU GPU

20 70 1998
pre-GPU 1998 GPU modern GPU pre-GPU
SGIEvans & Sutherland GPU GPU
modern GPU transistorsmicrochip

GPU CPUIntel 2.4GHz Pentium IV
5 5 55 million NVIDIA GeForce FX GPU
1 2 5 125 million NVIDDIA 7800 GXT
13
3 2 302 million
Modern GPU 1998 4 NVIDIA

1998 Modern GPU Modern GPU
Modern GPU NVIDIA TNT2ATI Rage 3Dfx Voodoo3
GPU CPU
GPU
GPU
1999 2000 modern GPU GPU

3D Object Transformation and Lighting,
T&L OpenGL DirectX7
workstation GPU
cube mapNVIDIA GeForce256
GeForce MAXATI Radeon 7500
2001 modern GPU GPU vertex

programmability GeForce 3GeForce 4TiATI 8500
GPU GPU
GPU
2009

Direct8 OpenGL
vertex programmability
GPU fragment programmability
modern GPU vertex programmability
fragment programmability GPU
modern GPU 2002 2003 NVIDIA GeForce

FX ATI Radeon 9700 GPU vertex
14
programmability fragment programmability DirectX OpenGL
API vertex programmability fragment programmability 2003
DirectX OpenGL
GPU GeForce
ATI
DirectX OpenGL
1. vertex programmability fragment programmability
2. IEEE32
3. 4 4
4.
5. >27.1GB/s
6. 1D2D3D
7. Render to TextureRTT
GPU Feng Liu Platform Independent

Real-time X3D Shaders and Their Applications in Bioinformatics Visualization
1.2 GPU VS CPU
GPU CPU GPUGPU

vertex programmability fragment programmability
2 6
Modern GPUs implement a number of graphics primitive operations in a way
that make running them much faster than drawing directly to the screen with the host
CPU. They are efficient at manipulating and displaying computer graphics, and their
highly parallel structure makes them more effective than typical CPUs for a range of
15
complex algorithms.
GPU highly parallel structure
GPU CPU 1 GPU VS
CPU GPU CPU CPU
GPU ALUArithmetic Logic Unit
CPU
OS GPU
Control ALU ALU
ALU ALU
Cache
DRAM DRAM
GPU CPU
1 GPU VS CPU
GPU
1
2 2D
CPU C++ GPU
16
2 CPU GPU
GPU CPU
GPU CPU GPU CPU
GPU CPU
GPU
GPU CPU
GPU CPUGPU
CPU
OpenGL fp2.0fp3.0
DirectX ps_4_0 profile
GPU
GPU
4-6
2012 GPU
GPU API
GPU
GPU
Shader language GPU
17
1.3
GPU Visualization in Scientific Computing
CT CPU
GPU
CPU
GPU GPU
GPUGPGPU
General-Purpose Computing on Graphics Processing Units,GPGP
GP 2 GPU
linear algebra
kruger and westermann 2003
physical simulation
Harris et al. 2003ray tracer
Purcell et
al.2002 Carr et al. 2002GPU
GPU29GPGPU
http://gpgpu.org/
GPU GPU
2004 BROOKGPU
C
BROOK GPU BROOKGPU AMD
30
1.4
GPU GPU
GPU bank BRDF
ray-casting GPU
18

GPU
GPU
GPU GPU CPU

GPU GPU shader language
3 GPU Shader
LanguageGPU GPU GPU
Shader Language vertex program
fragment program GPU
Cg
NVIDIA The Cg Tutorial The Definitive Guide to
Programmable Real-Time Graphics Cg ToolKit Users Manual
The Cg Tutorial The Definitive Guide to
Programmable Real-Time Graphics Cg _
Cg ToolKit
Users Manual

GPU
Cg BRDF
BRDF
19

GPU
20
2 GPU
GPU
GPU
GPU
GPU
GPU
CC++JAVA
CPU
b/s
Accelerated Graphic Port,AGP
387

234 GPU
10

Pixel
color buffer frame buffer
21

2.1

T&LT&L Transform & Lighting
Because, your application supplies the geometric data as a collection of vertices,

but the resulting image typically represents what an observer or camera would see
from a particular vantage point.
As the geometric data flows through the pipeline, the GPUs vertex processor
transforms the continuant vertices into one or more different coordinate system, each
of which serves a particular purpose. CG vertex programs provide a way for you to
program these transformations yourself.

GPU
GPU
GPU

Object spaceWorld spaceEye space
Clip and Project space 3 GPU
22
3 GPU
2.1.1 object space world space
When an artist creates a 3D model of an object, the artist selects a convenient

orientation and position with which to place the models continent vertices.
The object space for one object may have no relationship to the object space of
another object.3
object space object space coordinate

3DMAX
.max object space
coordinateobject space coordinate
object space coordinate world space coordinate
world space coordinate
23
world space coordinate object space coordinate
world space coordinate world
matrix
world coordinate space

eye coordinate space
object space GPU

world space
object space world space
object space world space world matrix
3D
11
2.1.2 world space eye space
GPU
world space eye space
eye space camera

viewing frustum
4
Frustum Culling
24
4
2.1.3 eye space project and clip space
Once positions are in eye space, the next step is to determine what positions are
actually viewable in the image you eventually intend trend.3
eye space
viewing frustum
Frustum Culling
clip
viewing frustum
(-1,-1,-1)(1,1,1)Canonical view volume,

CVV 9 CVV
XY 00Z
CVV
screen coordinate space
25
1. CVV
2. CVV
3.
viewing frustum
CVV
Twinsen
12 3

CVV
12
5
1 Twinsen
2 GPU C\C++
CPU
GPU
GPU
GPU C++ OGRE
Object-Oriented Graphics Rendering Engine
2.2 Primitive Assembly && Triangle setup
Primitive Assembly primitive,
26

Back-face Culling
OpenGL
enable disable
2.2 2.3
View Frustum Culling

Back-Face Culling
Occlusing Culling
Triangle Setup
Rasterizing
2.3
2.3.1 Rasterization
Rasterization is the process of

determining the set of pixels covered by a geometric primitive
Screen coordinate

27

2 3

10.4820.511021
52

DDA Bresenham
3
(vertex)
(pixel)
2.3.2 Pixel Operation
Pixel operation Raster Operation2 Raster

Operation

Pixel operation
2Texture operation
3Blending
28
Alpha
alpha
RGB Z
alpha
alpha 1.0 0
RGBA over
cd = aica + (1 a )cs over
a alpha ca cs
cd Over
compositingOGRE
compositor

z buffer
4.5 59
4Filtering Filtering
5 21.2.3
29
5 OpenGL Direct3D Raster Operations
2.4
GPU
z bufferstencil buffer
frame buffer color buffer
2.4.1 GPU
cpu gpu
cpu gpu
Shader
GPU glsl hlsl
GPU
30
6 GPU
2.4.2 Z Buffer Z
Z buffer depth buffer

Z
Z 01
Z 0.0 Z
1.0 z buffer
depth-buffer method Z
Z
Z Euclidean distance
steve Baker Learning to love your

Z-buffer GPU Z
az +b
z _ buffer _ value = (1 << N )
z
f
a=
f n
31
f n
b=
n f
f n
z N Z
z buffer z
Z
3D
7 AE
1 5 234
BCD BCD
Z Z
Z

1000
32
7 Z
16 Z buffer 24 Z buffer
32 32 Z buffer Z
Z-precision
Z Z
flimmeringZ-fighting
2.4.3 Stencil Buffer
A stencil buffer is an extra buffer, in addition to the color buffer and depth buffer
found on modern computer graphics hardware. The buffer is per pixel, and works on
integer values, usually with a depth of one byte per pixel. The depth buffer and stencil
buffer often share the same area in the RAM of the graphics hardware.
Stencil buffer, buffer

z buffer 15 z buffer 1 stencil buffer( 2 )
24 z buffer 8 stencil buffer 4
33
stencil buffer( Z buffer) Z buffer stencil buffer
Stencil buffer
buffer stencil buffer 1
shadow volume
2.4.4 Frame Buffer
A framebuffer is a video output device that drives a video display from a

memory buffer containing a complete frame of data. The information in the buffer
typically consists of color values for every pixel on the screen. Color values are
commonly stored in 1-bit monochrome, 4-bit palettized, 8-bit palettized, 16-bit
highcolor and 24-bit truecolor formats. An additional alpha channel is sometimes used
to retain information about pixel transparency. The total amount of the memory
required to drive the framebuffer depends on the resolution of the output signal, and
on the color depth and palette size.
Frame buffer buffer

Frame buffer color buffer
z buffer 12 frame buffer
webMediaBrands frame buffer
frame buffer
general main memory
Frame buffer: The portion of memory reserved for holding the complete
bit-mapped image that is sent to the monitor. Typically the frame buffer is stored in
the memory chips on the video adapter. In some instances, however, the video chipset
is integrated into the motherboard design, and the frame buffer is stored in general
main memory.
34
2.5
GPU
GPU
openGPU
2 3 Blinn A Trip
Down the Graphics Pineline
35
3 Shader Language
In the last year I have never had to write a single HLSL/GLSL shader. Bottom
line, I cant think of any reason NOT to use CG.
shader languageshade
Wikipedia shader language The job of a surface shading procedure is
to choose a color for each pixel on a surface, incorporating any variations in color of
the surface itself and the effects of lights that shine on the surface(Marc Olano)
shader language
GPU shader
language
shader language
shader language GLSL High Level Shading

LanguageCg C for Graphic shader language
C
shader language shader language GPU
shader language
GPU shader language
shader language C++\JAVA
Shader language
3 OpenGL GLSL Direct3D HLSL NVIDIA
Cg
shader language
Programmable Vertex Processor
36
Programmable Fragment Processor
vertex program fragment program
GLSLHLSL Cg
3.1 Shader Language
shader language shader program

vertex shader program fragment shader program
GPU Programmable Vertex Processor

Programmable Fragment Processor
[2] 1.2.4
The vertex and Fragment processing broken out into programmable units. The
Programmable vertex processor is the hardware unit that runs your Cg Vertex
programs, whereas the programmable fragment processor is the unit that runs your Cg
fragment programs.

GPU 8
37
8
3 GPU
Vertex program Fragment program
9 \
buffer
38
9 \
3.2 Vertex Shader Program
Vertex shader program Fragment shader program

Programmable Vertex Processor
Programmable Fragment Processo
GPU
10
vertex buffer
texture buffer
39

phong
Gourand
phong
phong Gourand
GPU shader
10
3.3 Fragment Shader Program
40

GPU
3.4 CG VS GLSL VS HLSL
Shader language 3 OpenGL GLSLOpenGL

Shading Language GLslang Direct3D HLSLHigh Level Shading
Language
NVIDIA Cg C for Graphic
GLSL HLSL OpenGL Direct3D

OpenGL Direct3D
OpenGL Direct3D
ATI OpenGL OSG

Open Scene
Graphic OpenGL 3D
GLSL
HLSL Cg HLSL
Cg OpenGL API
OpenGL
GLSL GLSL
OpenGL unix
41

HLSL windows
HLSL HLSL
Shader language HLSL

NVIDIA AMD ATI
GPU GeForce
Cg NVIDIA
GPU
GPU Cg
NVIDIA Cg
Cg
Cg OpenGL Direct3D
Cg OpenGLDirectX OpenGL DirectX
Cg OpenGL DirectX
Cg Microsoft NVIDIA
[1] 1.3.1 Microsoft and NVIDIAs
Collaboration to Develop Cg and HLSLHLSL Cg
Cg _ 29
HLSL Cg cg
Microsoft NVIDIA cg
GPU CG

Microsoft NVIDIA Cg
42
GLSL
Cg C for graphics C
C
C++ Java C cg C
CC++Java Cg
Cg C
Cg cg
NVIDIA Cg
CgGLSLHLSL Cg
2007 12 OSG CgGLSLHLSL
Cg GLSL
3D Cg
2009 Cg
OGRE Cg GLSL
In the last year I have never had to write a single HLSL/GLSL shader. Bottom
line, I cant think of any reason NOT to use CG.
3.5

CgGLSLHLSL
Cg GLSL
43
4 Cg
CgC for Graphcis NVIDIA Microsoft
Cg C Cg
*. Cg
4.1 Cg
NVIDIA http://developer.nvidia.com/object/cg_toolkit.html Cg
Toolkit 2009 10 Cg 2.2
bin cgc.exe NVIDIA Cg
Cg NVIDIA Cg
IDE
.cg NShader
http://nshader.codeplex.com/ Visual Studio2008
11
44
11 NShader
4.2 CG
Cg OpenGL Direct3D API

Cg Cg
OpenGL Direct3D
4.3 CG
4.3.1 CG
01
C C++
.exe GPU
static
coompilation
dynamic compilation
Cg Cg ( Cg
45
Runtime library) Cg Cg
OGRE
[2] 1.4.2 Cg
Cg
Cg Cg APIOpenGL Direct3D
OpenGL Direct3D Cg
OpenGL Direct3D
NVIDIA Cg cgc.exe
Cg
Cg
GPU Cg
AIP Cg
Cg Profiles profile profile
profile
Cg Profiles Cg Cg
Cg Profile
46
4.3.1 CGC
12 cgc -h
Cg Toolkit NVIDIA Corporation\Cg\bin

cgc.exe cgc h
12
Cg
cgc [options] file
[options]file Cg
profile
47

cgc profile glslv entry main_v test.cg
-profile profile glslv profile

-entry
main_v
test.cg
main\
main_v ,main_f
1. profile profile
profile
profile
profile profile fp20
cgc profile fp20 glslv entry main_v test.cg

profile
2. profile
profile
tex2D(sampler2D tex , float3 sz ,float2 dsdx , float2 dsdy ) fp20

cgc profile fp20 entry main_f test.cg
error C3004: function tex2D not supported in this profile
48
fp30
cgc profile fp30 entry main_f test.cg
for,whilevs_2_x, vp30, vp40,fp40

profilesCgUsersManualIn other profiles, for and
while loops may only be used if the compiler can fully unroll them (that is, if the
compiler can determine the iteration count at compile time)
profilesforwhile
profilesfor,while
error c6003instruction limit of exceeded
profiles
3. .cg

cgc profile glslv entry main_v
fatal error C9999: Cant open file:test
4. cgc Cg
GLSL HLSL
test.cg main_v
glsl direct.glsl
cgc profile glslv o direct.glsl entry main_v test.cg
5.

49

GPU
C++\JAVA
4.4 CG Profiles
Profile [1]A Cg profile

defines a subset of the full Cg language that is supported on a particular hardware
platform or APICgUsersManual 21 Cg profile
API Cg
shader language
:Profile
Profile Profile profile profile OpenGL DirectX
OpenGL DirectX
Cg
Cg compiler profiles
z OpenGL ARB vertex programs
Runtime profile: CG_PROFILE_ARBVP1
Compiler option: _profile arbvp1
z OpenGL ARB fragment programs
Runtime profile: CG_PROFILE_ARBFP1
Compiler option: _profile arbfp1
50
z OpenGL NV40 vertex programs
Runtime profile: CG_PROFILE_VP40
Compiler option: _profile vp40
z OpenGL NV40 fragment programs
Runtime profile: CG_PROFILE_FP40
Compiler option: _profile fp40
z OpenGL NV30 vertex programs
z OpenGL NV30 fragment programs
z OpenGL NV2X vertex programs
z OpenGL NV2X fragment programs
z DirectX 9 vertex shaders
Runtime profiles: CG_PROFILE_VS_2_X
CG_PROFILE_VS_2_0
Compiler options:-profile vs_2_x
-profile vs_2_0
z DirectX 9 pixel shaders
Runtime profiles: CG_PROFILE_PS_2_X
CG_PROFILE_PS_2_0
Compiler options: -profile ps_2_x
-profile ps_2_0
z DirectX 8 vertex shaders
Runtime profiles: CG_PROFILE_VS_1_1
51
Compiler options:-profile vs_1_1
z DirectX 8 pixel shaders
Runtime profiles: CG_PROFILE_PS_1_3
CG_PROFILE_PS_1_2
CG_PROFILE_PS_1_1
Compiler options: -profile ps_1_3
-profile ps_1_2
-profile ps_1_2
-profile ps_1_1
2009 10 profile
DirectX 11 profile
52
5 CG
Cgbuilt inprimitive
5.1
Cg 7
1. float32 profile
DirectX8 pixel profiles
2. half16
3. int32 profile int float
4. fixed12 fragment profiles
5. bool, if ?:
profiles
6. sampler*,the handle to a texture object

6
sampler, sampler1D, sampler2D, sampler3D, samplerCUBE, samplerRECT
DirectX profiles samplerRECT pixel
profiles NV40 vertex program profile CgUsersManual 30

7. string profile
Cg Cg runtime API
Cg
53
6 Wikipedia Cg
6 string Cg
(built-in vector data types)
float4 float 4 bool4 bool
4
4 Cg float1float2float3
float4 4
float5 array;//
float4 array = float4(1.0, 2.0, 3.0, 4.0);
float2 a = float2(1.0, 1.0);

float4 b = float4(a, 0.0, 0.0);
Cg 4*4
float1x1 matrix1;// float matirx1; x

float2x3 matrix2;// 2*3 6 float
float4x2 matrix3;// 4*2 8 float
float4x4 matrix4;// 4*4
float2x3 matrix5 = {1.0, 2.0, 3.0, 4.0, 5.0, 6.0};
54
Cg
C\C++ C\C++
Cg
5.2
General-purpose arrays can only be used as uniform parameters to a vertex

program. The intent is to allow an application to pass arrays of skinning matrices and
arrays of light parameters to a vertex program(3 Array )

Cg
Cg
Cg C
float a[10];// 10 float

float4 b[10];// 10 float4
float a[4] = {1.0, 2.0, 3.0, 4.0}; //
.length
float a[10]; //
int length = a.length;//
55
float b[2][3] = {{0.0, 0.0, 0.0},{1.0, 1.0, 1.0}};
int length1 = b.length; // length1 2

int length2 = b[0].length; // length2 3
4*4
float M[4][4];4 float4x4 M 16
float 4 float4x4 M[4]
4 4

profiles
Unsized arrays may only be declared as function parameters-they may not be

declared as variables. Furthermore, in all current profiles, the actual array length and
address calculations implied by array indexing must be known at compile time

3
8 8.1
5.3
Cg structure Cg C++
Cg
C++
public private shader
Cg
56

struct
myAdd
struct myAdd
{
float val;
float add(float x)
{
return val + x;
}
};
myAdd s;
float a = s.value;
float b = s.add(a);
profile
profile [3] structures and Member functions
,Cg
C
( Binding Semantics)
57
Cg POSTION )COLOR()NORMAL(
)Texcoord()
5.4 Interfaces
Cg
Cg
GPU
3
Interfaces
5.5
Cg C C
Cg
float a = 1.0;
half b = 2.0;
float c = a+b; // float c = a + (float)b;
float a = 1.0;
float b = a + 2.0; //2.0 float
Cg
58
float a = 1.0;
float b = a + 2.0h; //2.0h half
type suffix 3
z f float;
z h: half;
z x: fixed
59
6 CG
Cg7
Cg
operatoroperand
CgCC
CgSwizzle
CCg
>Cg
C
6.1 Comparison Operators

Cg6
1, bool
< expr < expr
<= expr <= expr
!= expr != expr
== expr == expr
>= expr >= expr
> expr > expr

1
60
Cgbool
boolean operator
Cgboolean operator
bool
float3 a = float4(0.5, 0.0, 1.0);

float3 b = float4(0.6, -0.1, 0.9);
bool3 c = a<b;
cfloat3(true, false, true);
6.2 Logical Operators
Cg3(boolean Operators) 2
bool
&& expr && expr
|| expr || expr
! !expr
2

bool
Cg&&||C
short-circuitingbool
bool
61
6.3 Math Operators
CgCg*
/-+%++*=/=+=-=
23.3Math Expressions
++
Cg
%int
error C1021: operands to % must be integral.
void function()
{
float2 a = float2(1.0, 1.0);
float b = 2.0;
f *= d;
f *= 2.0;
}
6.4
CgC
int
62
int2 a = int2(0.0,0.0);
int2 b = a>>1;
error C1021:operands to shr

must be integral.
float2 a = int2(0.0,0.0);
float2 b = a>>1;
6.5 Swizzle
Cgswizzle.
swizzleGPUswizzlexyzw
swizzle
rgba
swizzlergba
float4(a, b, c, d).xyz float3(a, b, c)

float4(a, b, c, d).xyy float3(a, b, b)
float4(a, b, c, d).wzyx float4(d, c, b, a)
float4(a, b, c, d).w float d
Cgfloat a float1 a
floatboolhalfswizzle
float a = 1.0;
float4 b = a.xxxx;
swizzle
swizzleerror C1010: expression left of .x is
63
not a struct or arrayarrayvector
[]
float a[3] = {1.0,1.0,0.0};

float b = a[0]; //
float c = a.x; //
6.6 Conditional Operators
expr1 ? expr2 : expr3;
expr1trueflasetrue,expr2expr3
if-else
if(a < 0){b = a}

else{c = a}
a < 0?(b = a) :( c = a);
Cgexpr1
boolexpr2expr3expr1
64
float3 h = float3(-1.0,1.0,1.0);
float3 i = float3(1.0,0.0,0.0);
float3 g = float3(1.0,1.0,0.0);
float3 k;
k = (h < float3(0.0,0.0,0.0))?(i):(g);
hfloat3(0.0, 0.0, 0.0)true, false, false,

iKgk
K(1.0, 1.0, 0.0)
6.7
Cg 3
Cg_
3.3.1
() [] -> .

! ~ ++ - + - * &
(type) sizeof

*/%
+-
<< >>
< >= > >=
== !=
&
^
|
&&
||
?:
= += -= *= /= %=
&= ^= != <<= >>=
,
65
3
6.8 Control Flow Statement
statement
Cifif-else
switchwhiledo-whileforCg
Cifif-elsewhileforbreak
for
Cgbool
CC01
vs_2_x, vp30vp40profilebranch
instruction
forwhileprofile3
In other profiles, for and while loops may only be used if the compiler can
fully unroll them (that is, if the compiler can determine the iteration count at compile
time)
profilesforwhile
fp40ps_3_0
profilesfor, whileerror c6003instruction
limit of exceededprofiles
returnrecursion
CgSwitch casedefaultCg
profile
66
7 \
GPU
Cg
1. GPUGPU
GPU

2. GPU

3.
7.1 Cg
67
intfloat structCg
C\C++ Cg
Semanticsbinding semantics
Cg inoutinout
\
in out
Cg uniformconst
C\C++

Cg
7.2 uniform
Cg [3]Program inputs and Outputs
1. Varying inputs,
GPU
Cg

2. Uniform inputs

Uniform

68
uniform float brightness,
uniform float4x4 modleWorldProject
float 4

OpenGL DirectX
Uniform
uniform Cg
uniform
uniform
Error C5056:uniformnot allowed on local variable
7.3 const
Cg const C\C++ const

const
const float a = 1.0;

a = 2.0; //
float b = a++; //
error C1026: assignment to const variable
const uniform
const uniform
7.4 \in\out\inout
C\C++
69
pass-by-value
pass-by-reference
C\C++
Cg
GPU Cg
1. in:

2. out:

3. inout:
void myFunction(out float x); // x
void myFunction(inout float x); // x
void myFunction(in float x); // x
void myFunction(float x); / in float x C\C++
return out \
7.5 SemanticBinding Semantics
70
Varying inputs
binding semantics
7.5.1
GPU
C/C++
Cg
Cg
binding semantics\
OpenGL Cg profiles
DirectX-based Cg profiles
\
POSITION
Semantics attached to parameters to non-main

functions are ignored(261 )

Shader POSITION
71
POSITION
POSITION
POSITION
OpenGL Cg profiles \
DirectX Cg profiles 3
25
In the OpenGL Cg profiles, binding semantics implicitly specify the mapping of varying
inputs to particular hardware registers. However, in DirectX based Cg profiles there is no
such implied mapping.
3 260
Often, these predefined names correspond to the names of hardware registers or API
resources
7.5.2
13 Cg vertex profile ,
profile
POSITION BLENDWEIGHT
NORMAL TANGENT
BINORMAL PSIZE
BLENDINDICES TEXCOORD0---TEXCOORD7
13
POSITION0 POSITION
in float4 modelPos: POSITION

POSITION OpenGL
72

in float4 modelNormal: NORMAL

NORMAL OpenGL
A
1
0
7.5.3

POSITION
Cg vertex profilesCg fragment

profilesPOSITIONPSIZEFOG,COLOR0-COLOR1,
TEXCOORD0-TEXCOORD7
POSITION
POSITION
73
void main_v(float4 position: POSITION,
//out float4 oposition : POSITION,
uniform float4x4 modelViewProj)
{
//oposition = mul(modelViewProj,position);
}
vp20vp30error C6014: Required output

HPOS not writtenvs_2_0vs_3_0error C6014
Required output POSITION not written.

struct
struct VertexIn
{
float4 position : POSITION;
float4 normal : NORMAL;
};
struct VertexScreen
{
float4 oPosition : POSITION;
float4 objectPos : TEXCOORD0;
float4 objectNormal : TEXCOORD1;
};
struct
POSITION

TEXCOORD
TEXCOORD
74
struct
vertex programPOSITION
fragment program
7.5.4
COLOR
GPU out
COLOR
void main_f(out float4 color : COLOR)

{
color.xyz = float3(1.0,1.0,1.0);
color.w = 1.0;
}
fragment profileDEPTH
multiple
render targets , MRTs
COLOR
7.5.5
\ 4 3 260
75
1.
[const] [in | out | inout]<type><identifier> [ :

<binding-semantic>][=<initializer>]
const inoutinout
type identifier

3

void mian_v(float4 position_obj : POSITION,

float3 normal_obj : NORMAL,
out float4 oPosition : POSITION,

out float4 oColor : COLOR,

{

}
2. struct
struct <struct-tag>
{
<type><identifier> [:<binding-semantic >];
};
C2E1v_Outpu 2 POSITION
COLOR C2E1v_green C2E1v_Outpu
76
struct C2E1v_Output {
float3 color : COLOR;
};
C2E1v_Output C2E1v_green(float2 position : POSITION)

{
C2E1v_Output OUT;
OUT.position = float4(position,0,1);
OUT.color = float3(0,1,0);
return OUT;
}
3.
<type> <identifier> (<parameter-list>) [:<binding-semantic]

{
<body>
}
COLOR
float4 return
float4 main_v(float4 position: POSITION,

out float4 oposition : POSITION,
uniform float4x4 modelViewProj):COLOR
{
oposition = mul(modelViewProj,position);
float4 ocolor = float4(1.0,0,0,0);
return ocolor;
}
4.

77
<type> <identifier> [:<binding-semantic>][=<initializer>];
78
8
5 7 5
6 \ 7
Cg
\ C\C++ main() Cg
8.1
Cg
C\C++return typeparameter list
void
void myFunc(inout float val ) float myFunc(float vals[])

{ {
float sum = 0.0;
..
val += 10.0; return sum;
} }
void
error C0000: syntax error, unexpected(at token (

error C0501: type name expected at token (
error C1110: function main_v has no return statement
7.4 \
79
C\C++ Cg
8.1.1
C\C++
stephen C.Dewhurst C++
Cg
Cg [3]Array
The most important difference from C is that arrays are first-class types. That
means array assignments actually copy the entire array, and arrays that are passed as
parameters are passed by value, rather than by reference.
Cg first-class types

first-class
first-class Matthieu Sozeau and

Nicolas Oury First-Class Type Classes

Unsized Array
float myFunc( float vals[])

{
float sum = 0.0;
for(int i = 0; i< vals.length; i++)
{
sum += vals[i];
}
return sum;
}
80
myFunc
float
error C1102: incompatible

type for parameter
float myFunc( float vals[3])

{
float sum = 0.0;
for(int i = 0; i< vals.length; i++)
{
sum += vals[i];
}
return sum;
}
void main()
{
float a[2] = {0.0, 1.0};
float b[3] = {0.0, 1.0, 2.0};
myFunc(a); //
myFunc(b); //
}
8.2
Cg Functon Overlaoding C++
81
bool function(float a, float b) {return ( a == b);}
bool function(boo a, bool b) {return ( a == b);}
Cg
8.3
C\C++ main()

4.4 CG main
Cg
7.5
C2E1v_green
return CgNVIDIA
Corporation\Cg\examples\OpenGL\basic\ 01_vertex_program\C2E1v_green.cg
82

struct C2E1v_Output {
};
C2E1v_Output C2E1v_green(float2 position : POSITION)

{
C2E1v_Output OUT;
OUT.position = float4(position,0,1);
OUT.color = float3(0,1,0);
return OUT;
}
8.4 CG
C Cg
reflect
refract mul
GPU
Cg GPU
GPU
Cg
1. Mathematical Functions;
2. (Geometric Functions)
3. (Texture Map Functions)
4. (Derivative Functions)
83
5. (Debugging Function)
8.4.1 Mathematical Functions
4 Cg

abs(x)
acos(x) [-1,1]
[ 0, ]
all(x) 0 ture
flase&&
any(x) 0
true||
asin(x)
, [ 1,1]

,
2 2
atan(x)

,
2 2
atan2(y,x) y/x atan(x)
atan(x) =
atan2(x, float(1))
ceil(x) ceil(float(1.3))
2.0
clamp(x,a,b) x a a x b
b x
cos(x)
x [ 1,1]
cosh(x) hyperbolic cosine x

cross(A,B) (cross product)

degrees(x) (radians)
(degrees)
84
determinant(m)
dot(A,B) A B (dot product)
A B

exp(x)
e x e= 2.71828182845904523536
exp2(x)
2 x
floor(x) floor(float(1.3))
1.0 floor(float(-1.3))
-2.0 ceil(x)
fmod(x,y) x/y y 0
frac(x) Returns the fractional portion of a scalar or
each vector component
frexp(x, out x
exp) x = m* 2^exp m exp
x 0 0
isfinite(x)
true false;
(not-a-number NaN)
isinf(x)
true false;
isnan(x)
(not-a-number NaN) true
false;
ldexp(x, n)
x 2n
lerp(a, b, f)
(1 f ) a + b f a + f ( b a )
a b f
a b f

lit(NdotL, N L H
NdotH, m) m

4
X 1.0
Y N L < 0
0 N L
Z N L < 0
N H < 0 0 ( N H ) m
W 1.0
85
log(x)
ln ( x ) x 0
log2(x)
log (2x ) x 0
log10(x) ( x)
log10 x 0
max(a, b)
min(a,b)
modf(x, out ip) Cg Reference Manual
mul(M, N) M AxB
N BxC AxC

mul(M, v)
mul(v, M)
noise( x) 01

pow(x, y) y
x
radians(x)
round(x) Round-to-nearest closest integer to x

rsqrt(x) X x 0
saturate(x) x 0 0 x 1
1 x
sign(x) x 0 1 x 0
01 0
sin(x)
[ 1,1]
sincos(float x, x sin cos

out s, out c) s=sin(x)c=cos(x)
sin cos
!
sinh(x) hyperbolic sine
smoothstep(min, x minmax x=min
max, x) 0 x=max 1 x

x min 3 x min 2
2*( ) + 3*( )
max min max min
step(a, x) x<a 0 1
sqrt(x)
x x x 0
86
tan(x)
tanh(x)
transpose(M) M
4 Cg
8.4.2 Geometric Functions
5
Cg
3 normalize reflect
refract
1.

2. reflect refract

distance( pt1, pt2) Euclidean
distance
faceforward(N,I,Ng)
Ng I < 0 N-N
length(v) sqrt(dot(v,v))
normalize( v)
reflect(I, N) I
N I N

I

refract(I,N,eta) I N
eta I N
I N
000
I
87

5 Cg
8.4.3 Texture Map Functions
Cg ps_2_0ps_2_x
arbfp1fp30 fp40 profiles fully supported

tex1D(sampler1D tex, float s)

tex1D(sampler1D tex, float s, float dsdx, float dsdy)
derivatives
Tex1D(sampler1D tex, float2 sz)

Tex1D(sampler1D tex, float2 sz, float dsdx,float dsdy)
derivatives
Tex1Dproj(sampler1D tex, float2 sq)

Tex1Dproj(sampler1D tex, float3 szq)

Tex2D(sampler2D tex, float2 s)

Tex2D(sampler2D tex, float2 s, float2 dsdx, float2 dsdy)
derivatives
Tex2D(sampler2D tex, float3 sz)

Tex2D(sampler2D tex, float3 sz, float2 dsdx,float2 dsdy)
derivatives
Tex2Dproj(sampler2D tex, float3 sq)


texRECT(samplerRECT tex, float2 s)
texRECT (samplerRECT tex, float2 s, float2 dsdx, float2 dsdy)
texRECT (samplerRECT tex, float3 sz)

texRECT (samplerRECT tex, float3 sz, float2 dsdx,float2 dsdy)
88
texRECT proj(samplerRECT tex, float3 sq)
texRECT proj(samplerRECT tex, float3 szq)
Tex3D(sampler3D tex, float s)

Tex3D(sampler3D tex, float3 s, float3 dsdx, float3 dsdy)
derivatives

texCUBE(samplerCUBE tex, float3 s)

texCUBE (samplerCUBE tex, float3 s, float3 dsdx, float3 dsdy)
derivatives
texCUBEproj (samplerCUBE tex, float4 sq)

6 Cg
s z depth comparison
q perspective value,
S
0 1
1D
2D 3D TexREC
3D Volume Texture
14
15 16
89

14
proj
float4
uvproj
tex2Dproj(texture,uvproj);
float4 uvproj = uvproj/uvproj.q;

tex2D(texture,uvproj);
90
8.4.4 Derivative Functions

ddx(a) a
X
ddy(a) a
X
7 Cg

ddx x
ddy y
Cg Refernece
Manual ddx (ddy )returns
approximate partial derivative with respect to window-space X
\
1. ddx ddy
2. ddx ddy
3. ddx ddy
If you evaluate ddx (myVar), the GPU will give you the difference between the
value of myVar at the current pixel and its value at the pixel next door. It's a straight
linear difference, made efficient by the nature of GPU SIMD architectures
(neighboring pixels will be calculated simultaneously).The derivative of any uniform
value will always be zero.Because these derivatives are always linear, the second
derivativesfor example, ddx(ddx(myVar))will always be zero.
ddx myVar
p ( i, j ) ddx(myVar) p ( i + 1, j ) myVar
91
ddy(myVar) p ( i, j + 1) myVar ddx ddy
ddx\ddy myVar ddx(myVar)
p ( i + 1, j ) myVar
1. ddx ddy
2. ddx ddy
3. ddx ddy
GPU Gems: Programming Techniques, Tips and Tricks

for Real-Time Graphics 24.2. Understanding GPU Derivatives by Randima
Fernando.
Complex filtering depends on knowing just how much of the texture (or shading)
we need to filter. Modern GPUs such as the GeForce FX provide partial derivative
functions to help us. For any value used in shading, we can ask the GPU: "How much
does this value change from pixel to pixel, in either the screen-x or the screen-y
direction?"
These functions are ddx() and ddy(). Although they are little used, they can be
very helpful for filtering and antialiasing operations.
GPU partial derivative

DDX DDYCg ddx()
ddy()(glsl hlsl ) x y
TXD
92

ddx X ddy Y
ddx ddy ddx ddy
ddx ddy
A B
A B
GeForce GPU
15
15
tex2D(sampler2D tex , float3 sz ,float2 dsdx , float2 dsdy )

x,y0.010.010.020.020.0
0.015
93
16
8.4.5 Debugging Function

void debug(float4 x) DEBUG
x
COLOR

8 Cg
profiles shader language

8 debug
Cg
C++
8.5
94

IDE

GPU
95

913
1415
illumination model
ward BRDF
IMopenGPU
96
9 illumination model

---
CG
CG
9.1
(Ambient Light):
389
97

9.2 Lambert

diffuse reflection
Lambert
(9-1)
I ambdiff = kd I a (9-1)
I a kd 0< kd <1
I ambdiff
Lambert
98
Lambert Lambert
I ldiff = kd I l cos (9-2)
Il
(0 90) I ldiff
90
N L
cos N L (9-2)
(9-3)
I ldiff = kd I l ( N L) (9-3)
Lambert
I diff = I ambdiff + I ldiff = kd I a + kd I l ( N L) (9-4)
9.2.1
Hello Cg World
Cg
99
1
void main_v(float4 position : POSITION,

float4 normal : NORMAL,

out float4 color : COLOR,
uniform float4x4 modelViewProj,

uniform float4x4 worldMatrix,
uniform float4x4 worldMatrix_IT,
uniform float3 globalAmbient,
uniform float3 lightPosition,
uniform float3 lightColor,
uniform float3 Kd)
{
oPosition = mul(modelViewProj, position);
float3 worldPos = mul(worldMatrix, position).xyz;

float3 N = mul(worldMatrix_IT, normal).xyz;
N = normalize(N);
//
float3 L = lightPosition - worldPos;
L = normalize(L);
//
float3 diffuseColor = Kd*lightColor*max(dot(N, L), 0);
//
float3 ambientColor = Kd*globalAmbient;
color.xyz = diffuseColor+ambientColor;
color.w = 1;
}
17
17
100

2
struct VertexIn
{
};
struct VertexScreen
{
};
void main_v(VertexIn posIn,
out VertexScreen posOut,

uniform float3 Kd)
{
posOut.oPosition = mul(modelViewProj, posIn.position);
float3 worldPos = mul(worldMatrix, posIn.position).xyz;

float3 N = mul(worldMatrix_IT, posIn.normal).xyz;
N = normalize(N);
//
float3 L = lightPosition - worldPos;
L = normalize(L);
//
float3 diffuseColor = Kd*lightColor*max(dot(N, L), 0);
//
float3 ambientColor = Kd*globalAmbient;
posOut.color.xyz = diffuseColor+ambientColor;
posOut.color.w = 1;
}
101

9.3 Phong
Lambert
Phong Bui Tuong phong
(9-5)
I spec = k s I l (V R ) ns (9-5)
k s ns V
ns
ns
R L N
R + L = (2 N L) N (9-6)
102
R = (2 N L) N L (9-7)
Cg 8.3.2
VS
9.3.1 phong
Phong 3 18
phong
18 phong
18 phong
18

phong
103
3 phong
struct VertexIn
{
float4 position : POSITION; // Vertex in object-space
};
struct VertexScreen
{
};
void main_v( VertexIn posIn,
uniform float3 eyePosition,
uniform float3 Kd,
uniform float3 Ks,
uniform float shininess)
{
float3 worldPos = mul(worldMatrix, posIn.position).xyz;

float3 N = mul(worldMatrix_IT, posIn.normal).xyz;
N = normalize(N);
//
float3 L = normalize(lightPosition - worldPos);
float3 V = normalize(eyePosition - worldPos);
float3 R = 2*max(dot(N, L), 0)*N-L;
R = normalize(R);
//
float3 diffuseColor = Kd * globalAmbient+Kd*lightColor*max(dot(N, L), 0);
//
float3 specularColor = Ks * lightColor*pow(max(dot(V, R), 0), shininess);
posOut.color.xyz = diffuseColor + specularColor;

posOut.color.w = 1;
}
104
phong
4 3 VertexScreen
TEXCOORD objectPos objectNormal
4 phong
struct VertexIn
{
};
struct VertexScreen
{
};
TEXCOORD
5 phong

{
posOut.objectPos = posIn.position;
posOut.objectNormal = posIn.normal;
}
105
6 phong
6 phong
void main_f( VertexScreen posIn,


uniform float3 Kd,
uniform float3 Ks,
{
float3 worldPos = mul(worldMatrix, posIn.objectPos).xyz;
float3 N = mul(worldMatrix_IT, posIn.objectNormal).xyz;
N = normalize(N);
//
float3 R = 2*max(dot(N, L), 0)*N-L;
R = normalize(R);
//
//
float3 specularColor = Ks * lightColor*pow(max(dot(V, R), 0), shininess);
color.xyz = diffuseColor + specularColor;

color.w = 1;
}
19 phong
106
19 phong
9.4 Blinn-Phong
Blinn-Phong Blinn-phong BlinnPhong reflection

model phong modified Phong reflection model Jim Blinn
1977 Models of light reflection for computer synthesized pictures
phong phong
Blinn-phong Lambert
Phong CG
OpenGL Direct3D Blinn-Phong
phong V R R
V Blinn-phong N H
V R Blinn-phong
I spec = ks I l ( N H ) ns
9-8
N H L V
107
L +V
H= 9-9
| L +V |
Blinn-phong phong
phong blinn-phong
Blinn-phong Blinn-phong
Blinn-phong
phong
Blinn-phong
7 Blinn-phong
void main_f(VertexScreen posIn,

uniform float3 Kd,
uniform float3 Ks,
{
N = normalize(N);
//\\
float3 H = normalize(L + V);
//
//
float3 specularColor = Ks * lightColor*pow(max(dot(N, H), 0), shininess);
color.xyz = diffuseColor + specularColor;

color.w = 1;
}
108
Blinn-phong
20 Blinn-phong
19 20
phong blinn-phong
phong blinn-phong
9.5 Rendering Equation
Kajia 1986 rendering equation
Lo ( x, wo ) = Le ( x, wo ) + f r ( x, wi , wo ) Li ( x, wi )(n wi )dwi 9-10

x Lo ( x, wo ) x wo
Le ( x, wo ) x wo
f r ( x, wi , wo ) wi x wo
BRDF Li ( x, wi ) wi x n x

phong BRDF
109

Lo ( x, wo ) = f r ( x, wi , wo ) Li ( x, wi )(n wi ) 9-11

BRDF
112

Lo ( x, wo ) = I diff + f rs ( x, wi , wo ) Li ( x, wi )(n wi ) 9-12
I diff f rs ( x, wi , wo ) BRDF
phong rendering equation
Phong
ks (nih) ns
f rs ( x, wi , wo ) = 9-13
ni wi
10 BRDF f rs ( x, wi , wo )
9.6
GPU
110
10
21
21
BRDFBidirectional Reflectance Distribution Function
BRDF
111
Cook-TorranceGPU
BRDFBRDFBidirectional Reflectance Distribution
FunctionBRDF
10.1 Cook-Torrance
phong blinn-phong
1981 Robert L. Cook Kenneth E. Torrance A Reflectance

Model For Computer Graphics Cook-Torrance
Cook-Torrance rough surface
Cook-Torrance
10-1
I c t = I diff + I spec = I diff + k s I l Rs 10-1
I diff k s I l Rs
cook-Torrance phong
blinn-phong Rs cook-Torrancephong
blinn-phong Rs
Rs specular term
Cook-Torrance Rs Wikipedia
112
specular highlight 10-2
F D G
Rs = 10-2
V N
3D 117 D3DBook (Lighting)
Cook-Torrance 10-3
F D G
Rs = 10-3
( N iV ) ( N i L )
A Reflectance Model For Computer Graphics
10-3
F Fresnel Fresnel reflect term

D Beckmann distribution factor

G Geometric attenuation
term N V L
schlick Fresnel 11 1%
10-4
F = f 0 + (1 f 0 )(1 V i H )
5
10-4
f 0 0 Fresnel V
H
Backmann
tan 2
1
D= 2 e m2
10-5
m cos 4
m m m
113
N H
1 cos 2
tan 2
cos 2
cos 2 1 ( N i H )2 1
= = 2 = 10-6
m2 m2 m cos 2 m 2 ( N i H ) 2
Backmann 10-7
( N i H ) 2 1 ( N i H ) 2 1
1 2 2 1 2 2
D= 2 e m ( N i H ) = 2 e m ( N i H ) 10-7
m cos
4
m (N iH ) 4
G = min(1, G1 , G2 ) 10-8
2( N i H )( N i L)
G1 = 10-9
V iH
2( N i H )( N iV )
G2 = 10-10
V iH
Cook-Torrance specular term
Ict = Idiff + Ispec = kd Il ( NiL) +ksIl Rs

(NiH)2 1
= kd Il ( NiL) +ksIl
( f0 +(1 f0 )(1ViH)
5
) 1
m cos
2 4
em2(NiH)2
min(1,
2(NiH)(NiL) 2(NiH)(NiV)
ViH
,
ViH
)
ViN
10-11
119 beckmann

Backmann beckmann 10-12
114
( N i H ) 2 1
1 2 2
D= e m ( N i H ) 10-12
4 m ( N i H )
2 4
A Reflectance Model For Computer Graphics

D 10-7

Cook-Torrance
beckmann
1. Beckmann, Petr and Spizzichino, Andre, The Scattering of Electromagnetic

Waves from Rpugh Surfaces, MacMillan, pp.1-33, 70-98, 1963;
115
2. Computer graphics: principles and practice, James D. Foley - 1995 -
Computers - 1175
10.1.1 Cook-Torrance
Cook-Torrance 22
22 Cook-Torrance
Cook-Torrance
8 Cook-Torrance
void main_v( float4 position : POSITION,


out float3 worldPos : TEXCOORD0,
out float3 oNormal : TEXCOORD1,

uniform float4x4 worldViewProj)
{
oPosition = mul(worldViewProj, position);
worldPos = mul(worldMatrix, position).xyz;
oNormal = mul(worldMatrix_IT,normal).xyz;
oNormal = normalize(oNormal);
}
116
9 Cook-Torrance
void main_f(float3 position : TEXCOORD0,

float3 normal : TEXCOORD1,
uniform float3 Ka,
uniform float3 Kd,
uniform float3 Ks,
uniform float f,
uniform float m)
{
float3 P = position.xyz;
float3 N = normalize(normal);
float3 ambient = Ka * globalAmbient; //
float3 L = normalize(lightPosition - P);

float nl = max(dot(L, N), 0);
float3 diffuse = Kd * lightColor * nl; //
float3 V = normalize(eyePosition - P);

float3 specular = float3(0.0,0.0,0.0);
float nv = dot(N,V);
bool back = (nv>0) && (nl>0);
if(back)
{
float nh = dot(N,H);
float temp = (nh*nh-1)/(m*m*nh*nh);
float roughness = (exp(temp))/(pow(m,2)*pow(nh,4.0)); // beckmann
float vh = dot(V,H);
float a = (2*nh*nv)/vh;
float b = (2*nh*nl)/vh;
float geometric = min(a,b);
geometric = min(1,geometric); //
float fresnelCoe=f+(1- f)*pow(1- vh,5.0); //fresnel

float rs = (fresnelCoe*geometric*roughness)/(nv*nl);
specular = rs * lightColor * nl*Ks; //
}
color.xyz = ambient + diffuse + specular;
color.w = 1;
}
117

10.2 BRDF
10.2.1 BRDF
1965 Nicodemus, Fred Directional reflectance and emissivity of an

opaque surface BRDF BRDF Bidirectional Reflectance
Distribution Function
BRDF
111 23 wi wo
BRDF wi
wo
polynomial texture map
23
BRDF
dLr ( wo ) dLr ( wo )
f r ( wi , wo ) = = 10-13
dEi ( wi ) Li ( wi ) cos i dwi
118
Lr ( wo ) Ei ( wi )
wo Radiance
wi Irradiance
d dA
BRDF
BRDF rendering equation

9.4
10.2.2
(anisotropy)
anisotropic surface
24
24
119

25
1.
2.
3.
25
10.3 Bank BRDF
Bank BRDF
Bank BRDF
10-14
f = k s ( 1 ( L T ) 2 1 (V T ) 2 ( L T )(V T )) ns 10-14
ks ns L V
Bank BRDF 26 27 27
120
26 Bank BRDF 1
27 Bank BRDF 2
Bank BRDF
10 Bank BRDF


out float3 worldPos : TEXCOORD0,
out float3 worldNormal : TEXCOORD1,

uniform float4x4 worldMatrix_IT)
{
oPosition = mul(modelViewProj, position);
worldPos = mul(worldMatrix, position).xyz;
worldNormal = mul(worldMatrix_IT, normal).xyz;
}
121
11 Bank BRDF
void main_f(float4 position : TEXCOORD0,

float3 normal : TEXCOORD1,

uniform float3 Ka,
uniform float3 Kd,
uniform float3 Ks,
{
float3 P = position.xyz;
float3 N = normalize(normal);
float3 ambient = Ka * globalAmbient; //
float3 L = normalize(lightPosition - P);

float ln = max(dot(L, N), 0);
float3 diffuse = Kd * lightColor *ln; //
//
float3 V = normalize(eyePosition - P);
float3 specular = float3(0.0,0.0,0.0);
bool back = (dot(V,N)>0) && (dot(L,N));
if(back)
{
float3 T = normalize(cross(N,V)); //
float a = dot(L,T);
float b = dot(V,T);
float c = sqrt(1-pow(a,2.0))* sqrt(1-pow(b,2.0)) - a*b; // Bank BRDF
float brdf = Ks* pow(c, shininess);
specular = brdf * lightColor *ln;

}
color.xyz = ambient + diffuse + specular;
color.w = 1;
}
122
10.4
BRDF HTSG BRDF

BRDF Ward BRDF
BRDF
123
11
CPUGPU

CPU
GPU
GPU
GPU
SnellFresnel
GPUGPU
3D
124
11.1 Snell Fresnel
11.1.1 Snell
c v 11-1
1
1
c
n= 11-1
v

\ 1.0\ 1.0003
1.333
1.5-1.7
2.417
1.309
Snell
28 i
ni t nt
sin i ni = sin t nt 11-2
snell
125

i
28 Snell
fresnel
fresnel
11.1.2

1672
29
1936
29
126

11.1.3 Fresnel
Fresnel
Fresnel Fresnel fresnel

fresnel fresnel fresnel
139 fresnel fresnel
0.7 70% 30%
kr fresnel kt
fresnel
fresnel
3D
117 Fresnel
schlick Fresnel 1%
F = f 0 + (1 f 0 )(1 V i H )
5
11-3
f 0 0 Fresnel V
H 11-3
90 1
fresnel equations-Wikepedia 0 fresnel
127
(n n )
2
f0 = i t 2 11-4
( ni + nt )
fresnel 1
0 fresnel
4n1n2
ft 0 = 1 f 0 = 11-5
( ni + nt )
2
11-311-4
fresnel
(n n )
2
4n1n2
F i t 2+ (1 V H )5 11-6
( ni + nt ) ( ni + nt )
2
11-6 0
fresnel 11-7
fresnel
11-7
F (1 V H ) 4 11-7
4%
11.2
Environment Mapping, EMReflection Mapping

Blinn Newell 1976
10
Greece 1986 11
128
6
6
DirectX
.dds 6
29 OGRE 2
30 OGRE 2
2 5.7.4
31
129
31
1.
2.
3.
Cg smaplerCUBE
texCUBE
12
130
12
void main_v( float4 position: POSITION,

float4 normal: NORMAL,
out float4 oposition : POSITION,

out float3 R : TEXCOORD1,

uniform float4x4 modelToWorld)
{
oposition = mul( modelViewProj, position);
//
float3 positionW = mul( modelToWorld, position).xyz;
float3 normalW = mul( modelToWorld_IT, normal).xyz;
normalW = normalize( normalW);
float3 I = positionW - eyePosition;
// R.
R = reflect( I, normalW );
}
void main_f( float3 R : TEXCOORD1,

uniform samplerCUBE environmentMap)
{
color = texCUBE( environmentMap, R);
}
32
131
32
11.3
33
t A A
ia , ib A
i = (1 t ) ib + t ia 11-8
132
ib
ia
33
ia ib
ib
34
34
133

11-8
Cg lerp
8.3.1
11.2
13
struct VertexIn
{
};
struct VertexScreen
{
};
14

{
posOut.objectPos = posIn.position;
posOut.objectNormal = posIn.normal;
}
134
15
void main_f( VertexScreen posIn,


uniform float3 Kd,
uniform float3 Ks,
uniform float shininess,
uniform float etaRatio, //
uniform float transmittance, //
uniform samplerCUBE environmentMap //
)
{
N = normalize(N);
//\\
//
float3 specularColor = Ks * lightColor*pow(max(dot(N, H), 0), shininess);
float3 reflectColor = diffuseColor+specularColor;
//, refract
float3 I = normalize(worldPos - eyePosition);
float3 T = refract(I, N, etaRatio);
//
float3 refractedColor = texCUBE(environmentMap, T).xyz;
color.xyz = lerp(reflectColor, refractedColor, transmittance);

color.w = 1;
}
135
11.4
2005 Wyman ACM SIGGRAPH

GPU Interactive
image-space refraction of nearby geometry
GPU 16
136
12 Projective Texture Mapping
Projective Texture Mapping Segal Fast shadows

and lighting effects using texture maaping
(coordinates in projective space)

3D

35
35 (12)
12.1
137

13
15
36
gradient
36
138
37

(a) (b)
37
12.2 Homogeneous Texture Coordinates
homogeneous texture coordinates
s,t,r,q,(x, y, z,
w) s (s, t) r
q w 1
12.3
139

3D
NDC
POSTION
Projective texture mapping

38
38
38
140

01 normalMatrix ,
1/21/2
0.5 0 0 0.5
0 0.5 0 0.5
normalMatrix = 12-1
0 0 0.5 0.5

0 0 0 1
texView Pr ojMatrix = biasMatrix projectionMatrix viewMatrix worldMatrix 12-2
texView Pr ojCoordinate = texViw Pr ojMatrix mod elCoordinate 12-3
Cg
tex2DProj(sampler2D tex, float4 szq)
tex2DProjtex2D
q
141

API
16
void main_v(
float4 position : POSITION,
out float4 outPos : POSITION,

out float4 outShadowUV : TEXCOORD0,

uniform float4x4 worldViewProj,
uniform float4x4 texViewProj //
)
{
outPos = mul(worldViewProj, position);
//
float4 worldPos = mul(worldMatrix, position);
outShadowUV = mul(texViewProj, worldPos);
}
142
17
void main_f(
float4 shadowUV : TEXCOORD0,
out float4 result : COLOR,
uniform sampler2D projectiveMap //
)
{
shadowUV = shadowUV / shadowUV.w;
float4 mapColor ;
// 0-1
shadowUV.x = (shadowUV.x +float(1.0))/float(2.0);
shadowUV.y = (shadowUV.y +float(1.0))/float(2.0);
mapColor = tex2D(projectiveMap, shadowUV.xy);
result = mapColor;
}
12.4

Shadow Map 16
143
13 Shadow Map
Shadows are created by testing whether a pixel is visible from the light source,
by comparing it to a z-buffer or depth image of the light source's view, stored in the
form of a texture.
Shadow Map depth map Lance

Williams 1978 Casting curved shadows on curved surfaces
l enth1 l enth2 l enth1 l enth2
13.1 depth map
2D
Depth map l enth1
v l enth2 l enth1 l enth2
l enth2 > l enth1 v depth texure
v v v
depth map shdaow texture

2
144
Shadow texture Depth texture
39
wikipid shadow map depth map
shadow texture
39depth map shadow texture
Shadow texture
texture
projective
13.2 Shadow map shadow texture
Shadow map
Shadow map shadow
texture
map shadow map shadow

map shadow map
153 6.12.4 shadow map
shadow texture Shadow map depth

map
145
shadow texture
13.3 Shadow map
Shadow Map depth map()

depth map
depth map
1.

2. worldViewProjMatrix
Z Z
shadow
01Cg
0-1
depth map
depth map Z Buffer Z
Euclidean distanceZ
2.4.2 GPU
depth map
01
depth map
146

3. frame buffer depth

map depth map
depth map depth
map depth map
18 depth map

out float2 depth : TEXCOORD0,
uniform float4x4 worldViewProj )
{
oPosition = mul(worldViewProj, position);
//
depth.x = oPosition.z;
depth.y = oPosition.w;
}
19 depth map
void main_f(float2 depth : TEXCOORD0,

out float4 result : COLOR,
uniform float pNear ,
uniform float pFar,
uniform float depthOffset )
{
float depthNum = 0.0;
// 0-1
depthNum = (depth.x - pNear) / (pFar - pNear);
depthNum += depthOffset;
result.xyz = depthNum.xxx;
result.w = 1.0;
}
147
19 depthOffset
0.01
depth map
1.
worldViewProjMatrix
13
l enth2

2. depth map
l enth1
3. l enth2 l enth1
20 depth map

float2 tex : TEXCOORD,
out float4 outPos : POSITION,

out float4 outShadowUV : TEXCOORD0,

uniform float4x4 texViewProj)
{
outPos = mul(worldViewProj, position);

float4 worldPos = mul(worldMatrix, position);
//
outShadowUV = mul(texViewProj, worldPos);
}
148
21 depth map
void main_f(float4 position : POSITION,

float4 shadowUV : TEXCOORD0,
out float4 result : COLOR
uniform sampler2D shadowMap ,

uniform float pNear ,
uniform float pFar,
uniform float depthOffset,
uniform int pixelOffset)
{
//
float lightDistance = (shadowUV.z - pNear) / (pFar - pNear);
lightDistance = lightDistance - depthOffset;
shadowUV.xy = shadowUV.xy/ shadowUV.w;
//
float4 depths = float4(
tex2D(shadowMap, shadowUV.xy + float2(-pixelOffset, 0)).x,
tex2D(shadowMap, shadowUV.xy + float2(pixelOffset, 0)).x,
tex2D(shadowMap, shadowUV.xy + float2(0, -pixelOffset)).x,
tex2D(shadowMap, shadowUV.xy + float2(0, pixelOffset)).x);
float centerdepth = tex2D(shadowMap, shadowUV.xy).x;
//
float l_Lit = (lightDistance >= centerdepth? 0 : 1);
l_Lit += (lightDistance >= depths.x? 0 : 1);
l_Lit += (lightDistance >= depths.y? 0 : 1);
l_Lit += (lightDistance >= depths.z? 0 : 1);
l_Lit += (lightDistance >= depths.w? 0 : 1);
l_Lit *= 0.2f;
result = float4(l_Lit, l_Lit, l_Lit, 1.0);

}
40 shadow map
149
40 shadow map
Shadow map
Shadow map
depth map
01
z-fighting
Self-shadow Aliasing,
150
Spot light
Point light
cube textureDirectional light
, worldViewProjMatrix
stencil
shadow map C
151
14 Volume Rendering
1982 2
----1.1
20 80 Visualization in Scientific Computing
M.Levoy
Display of surfaces from volume data(14)volume rendering
describes a wide range of techniques for generating images from three-dimensional
scalar data
CT
,
CT CT
152

80
GPU
openGPU
server the people
14.1
14.2
153

server the people,
CT
CT
B
41
41
14.3

,
phong
154
cook-torrance

ray-cast
14.4 Volume Data
155

C
H20
XYZ
1.
2. CT MRI

VoxelVolume Texture
14.4.2 Voxel
Wikipediavoxel
A voxel (a portmanteau of the words volumetric and pixel) is a volume element,

representing a value on a regular grid in three dimensional space. This is analogous to
a pixel, which represents 2D image data in a bitmap
156

42
n*m*t
XYZnmt
256*256*256
256
mm
0.412mm0.412
42
14.4.1 Volume Texture
volume texture

2d texture3d texturevolume texture
3d texture
23
2d texture3d texture
157
3d texturevolume texture22
3D texture (Three Dimensional Texture), also known as "volume texture," is a

logical extension of the traditional (and better known) 2D texture. In this context, a
texture is simply a bitmap image that is used to provide surface coloring for a 3D
model. A 3D texture can be thought of as a number of thin pieces of texture used to
generate a three dimensional image map. 3D textures are typically represented by 3
coordinates.
2D
2D
1.
2.

x,y
XYvolume
texture
CT
DirectX.dds.raw
.raw
.raw.raw
158
.raw 43
University of TbingenGermanyViatronix Inc.(USA)Walter Reed Army Medical
Center (USA)
43
256 x 320 x 128 \0.66, 0.66, 0.66512 x 512

x 174\0.8398, 0.8398, 3.2512 x 512 x 463\0.625, 0.625, 1.0
http://wwwvis.informatik.uni-stuttgart.de/~engel/pre-integrated/data.html
http://www9.informatik.uni-erlangen.de/External/vollib/
http://www.volren.org/
14.5
Ray-casting-
Shear-warp Frequency Domain
Splatting

GPU
15
159
15 Ray Casting
15.1

CPU
GPU
16.1.2
160

15.1.1
Optical Models
Optical Model
15
1. Absorption only
2. Emission only
3. Absorption plus emission
4. Scattering and Shading/shadowing
5. Multiple Scattering
Absorption plus emission
15.2
15.2.1
161

OpenGL Direct3D
44
162
44
OpenGL 8
GPU 6 GPU
11
15.2.2
alpha
alpha over Alpha
c0 = as cs + (1 as )cd 15-1
as cs cd
c0
163

Ci = (1 Ai1 )Ci + Ci1

15-2
Ai = (1 Ai1 ) Ai + Ai1
Ci Ai
Ci Ai
Ci = (1 Ai )Ci+1 + Ci
15-3
Ai = (1 Ai ) Ai+1 + Ai
15.2.3
45
45 F L
164
m F nn<m
t = tstart + d * delta 15-4
tstart d delta
n t
n>m 1
OpenGL DirectX
profile
Cg OpenGL DirectX
165
15.2.2
m
n n>=m
n
CPU m
GPU
GPU
GPU
OpenGL Direct

1. frontDepth
14 frontDepth

2. backDepthbackDepth
166

3. m
14 shadow Map

\
46
46
frontDepth backDepth
rgba
a
01
167
15.3
Front face generation
Back face generation

front face
back face
47
47
168

48
48
15.4
169
22
struct VertexIn
{
float4 texCoord: TEXCOORD;
};
struct VertexScreen
{
float4 worldPos : TEXCOORD0;
float4 projPos : TEXCOORD1;
float4 texCoord : TEXCOORD2;
};
23
VertexScreen main_v(VertexIn posIn,

uniform float4x4 world,
uniform float4x4 texViewProj)
{
VertexScreen posOut;
posOut.position = mul(worldViewProj, posIn.position);

posOut.worldPos = mul(world,posIn.position);
posOut.projPos = mul(texViewProj, posOut.worldPos);
posOut.texCoord = posIn.texCoord;
return posOut;
}
170
24
void main_f(VertexScreen posIn,

uniform sampler3D volumeTex: register(s0),
uniform sampler2D frontDepthTex: register(s1) ,
uniform sampler2D backDepthTex: register(s2) ,
out float4 result : COLOR)
{
//
float3 dir = posIn.worldPos.xyz-eyePosition;
dir = normalize(dir);
float3 deltaDir = float3(0.0, 0.0, 0.0);
//
float3 tex = posIn.texCoord.xyz;
float2 uvDelta;
uvDelta.x = 0.0;//ddx( tex ).x;
uvDelta.y = 0.0;//ddy( tex ).y;
//,
float2 uv= posIn.projPos.xy/posIn.projPos.w;
float frontDis = tex2D(frontDepthTex,uv).x;
float backDis = tex2D(backDepthTex,uv).x;
float len = backDis-frontDis;
//
float3 norm_dir = normalize(dir);
float stepsize = 0.01;
float delta = stepsize;
float3 delta_dir = norm_dir * delta;
float delta_dir_len = length(delta_dir);
float3 vec = posIn.texCoord.xyz;
float4 col_acc = float4(0,0,0,0);
float alpha_acc = 0;
float length_acc = 0;
float4 color_sample;
float alpha_sample;
for(int i = 0; i < 800; i++){

color_sample = tex3D(volumeTex,vec);
alpha_sample = color_sample.a * stepsize;
col_acc += (1.0 - alpha_acc) * color_sample * alpha_sample * 3;
alpha_acc += alpha_sample;
vec += delta_dir;
length_acc += delta_dir_len;
if(length_acc >= len || alpha_acc > 1.0) break; //
}
result.xyz = col_acc.xyz*2.0+float3(0.2,0.2,0.2);
result.w = col_acc.w;
}
171
15.5
1415

Object-Order Empty Space
Skipping
Markus Hadwiger
Advanced Illumination Techniques for GPU-Based Volume
Raycasting SIGGRAPH ASIA2008
166
172
A
----F.S.Hill,JR
( a, b, c )

x ' cos sin 0 0 x

'
y = sin cos 0 0 y
. A-1
z' 0 0 1 0 z

1 0 0 0 1 1
Z 317
1
0

v v1 v 2 v3
A-2
v = a v1 + b v 2 + c v3 A-2
p p1 p 2 p3 po
A-3
173
p = a p1 + b p 2 + c p3 + po A-3
A-2A-3 2
A-4A-5
v = ( a b c 0 ) ( v1 v 2 v3 p0 ) A-4
p = ( a b c 1) ( p1 p 2 p3 p0 ) A-5

1 0 1
(100 100 100 1)
100 (100 100 100 0 )
p ( x p yp z p ) ( t x ty t z )
(x '
p y' p z' p ) = ( xp yp z p ) + (tx ty tz ) A-6
x ' p 1 0 0 tx x p
'
y p = 0 1 0 t y y p
A-7
z ' p 0 0 1 tz z p

1 0 0 0 1 1
174
B

1CT
X
XX
X
1963X
CT1967
XCT
19719
104
X
X
19724
CTCT
X
1979CT
CTXX
1895XX
X70
175
30CT
CT
CTXX
XX
x
CTXCTX
CT
X XCT2000
X20

2CT
CTX
CTXX
CT
X
176
CT
1~10

CT
CT
CT
3CT
CTX
X/
analog/digital converter
voxel
X
digital matrix/
digital/analog converter
pixelCTCT
X
I0 X , d X
X I
177
I = I 0 e ud B-1

()
, X
CT
CT X
X
123 d
I = I 0e (u1 +...+un ) d B-2
I ,

CT CT
CT
u uwater
CT = *1000 B-3
uwater
CStencil Shadow
Stencil Shadowstencil
buffer
shadow map
178

OGRE
shadow volume
stencil buffer
2.4.3
Frank Crow 1977 19

49, ABC
DEFABC DEF
shadow volume
B
A
C
D
E
49
1
2
179
3D
50
50
501 502
503
Z Pass Z Fail Z Pass
Z Pass stencil
stencil stencil
stencil
stencil
stencil stencil
2.4.3
180
Z Pass z buffer
stencil buffer
1Z Enable z buffer
2Z Disable z buffer
stencil
stencil
3
stencil
4
stencil stencil buffer
Z Pass stencil
shadow volume
Z Fail Z Pass z pass

Z Fail
1
2
stencil
3
181
stencil
4 stencil
stencil
Z Fail
51

Doom3
object
Outlining
1 Feng Liu, Platform Independent Real-time X3D Shaders and Their Applications
in Bioinformatics Visualization, Georgia State University, 2005, ISBN:
0-542-56173-5, Order Number: AAI3207182.
182
2 Randima Fernando, Mark J. Kilgard, The Cg Tutorial The Definitive Guide to
Programmable Real-Time Graphics, Addison-Wdsley Professional, March 8, 2003,
ISBN-10: 0321194969.
3 Cg Toolkit Users Manual A Developers Guide to Programmable Graphics.
Release 1.4, September 2005.
4 Kr uger, J., and Westermann, R. 2003. Linear algebra operators for GPU
implementation of numerical algorithms. ACM Trans. Graph. 22, 3, 908-916.
5 Harris, M. J., Baxter, W. V., Scheuermann, T., and Lastra, A. 2003. Simulation of
cloud dynamics on graphics hardware. In Proceedings of Graphics hardware,
Eurographics Association, 92-101.
6 Purcell, T. J., Buck, I., Mark, W. R., and Hanrahan, P. 2002. Ray tracing on
programmable graphics hardware. ACM Trans. Graph., 703-712.
7 Alexander J.Faaborg, A Psychophysical Study of BRDF-Based Lighting, May 10,
2001.
8 Wynn, C. (2000). An Introduction to BRDF-Based Lighting. nVIDIA
Corporation.
9 Michael Ashikhmin, Peter Shirley, An Anisotropic Phong BRDF Model. August
13, 2000.
10 Blinn, J.F., and M.E. Newell, Texture and reflection in computer generated
images, Communications of the ACM, vol.19, no. 10, pp.542-547, October 1976.
11 Greene, Ned, Environment Mapping and Other Applications of World
Projections, IEEE Computer Graphics and Application, vol, 6, no, 11, pp21-29,
November 1986.
12 Cass Everitt, Projective Texture Mapping, 04/24/2001
13 Lance Williams, Casting curved shadows on curved surfaces, 1978,
ISSN:0097-8930, page: 270-274.
14 Levoy,M. Display of surfaces from volume data. May, 1988, Computer Graphics
and Applications, IEEE, Volume: 8, Issue: 3.
15 Nelson Max. Optical models for direct volume rendering. IEEE
Transactions on Visualization and Computer Graphics, 1(2):99108, June 1995.

16 Matthieu Sozeau , Nicolas Qury, First-Class Type Classes.
17 steve Baker, Learning to love your Z-buffer.
18 Markus Hadwiger, Patric Ljung, Christof Rezk Salama, Timo Ropinski. Advanced
Illumination Techniques for GPU-Based Volume Raycasting. SIGGRAPH
ASIA2008.
183
19 Crow,Franklin C:Shadow Algorithms for Computer Graphics, Computer
Graphics (SIGGRAPH 77 Proceedings), vol. 11, no. 2, 242-248.
20 http://nshader.codeplex.com/
21 http://www.opengpu.org/
22 www.tech-faq.com3d-texture.shtml#
23 http://wwwvis.informatik.uni-stuttgart.de/~engel/pre-integrated/data.html
24 http://www9.informatik.uni-erlangen.de/External/vollib/
25 http://www.volren.org/
26 Randima Fernando, Mark J.Kilgard, Cg
_ 2004 9 ISBN:
711512430
27 Tomes Akenine-Moller, Eric-Haines,
2004-8-18 7301071051
28 Stephen C.DewhurstC++2005 12 16
7115141010
29 GPU 2006 4

30 2004 2
31 Frangois X.Sillion, James R.Arvo, A Global Illumination Solution for General
Reflectance Distributions. ACM O-89791-436-8/91/007/0187, 1991.
32 Robert L. Cook, A Reflectance Model For Computer Graphics, ACM
O-8971-045-I/81-0800-0370. 1981.
33 Xin Sun, Kun Zhou, Yanyun Chen, Interactive Relighting with Dynamic BRDFs.
34 Wolfgang Heidrich , Hans-Peter Seidel. Efficient Rendering of Anisotropic
Surfaces Using Computer Graphics Hardware.
35 David C.Banks. Illumination In Diverse Codimensions.
36 Zhao Dong, Jan Kautz, Christian Theobalt. Interactive Global Illumination Using
Implicit Visibility.
37 Daniel Reiter Horn, Jeremy Sugerman, Mike Houston. Interactive k-D Tree GPU
Raytracing.
38 Benthin,C., Wald,I., Scherbaum, M., And Friedrich,H. 2006. Ray Tracing on the
Cell Processor. In Proceedings of the 2006 IEEE Symposium on Interactive Ray
Tracing.
39 Buck, I., Foley, T., Horn, D., Sugerman, J., Fatahalian, K., Houston, M., And
Hanrahan, P. 2004. Brook for GPUs: Stream computing on graphics hardware. In
184
Proceedings of ACM SIGGRAPH 2004.
40 Xin Sun, Kun Zhou, Eric Stollnitz. Interactive Relighting of Dynamic Refractive
Objects.
41 Timothy J.Purcell, Ian Buck, William R.Mark. Ray Tracing on Programmable
Graphics Hardware.
42 Ziyad S.Hakura, John M.Snyder. Realistic Reflections and Refractions on
Graphics Hardware With Hybrid Rendering and Layered Environment Maps.
43 Turner Whitted, An Improved Illumination Model for Shaded Display. 1980 ACM
0001-0782/80/0600-0343.
44 Mark Segal, carl Korobkin, Rolf van Widenfelt. Fast Shadows and Lighting
Effects Using Texture Mapping. 1992 ACM-0-89791-479-1/92/007/0249.
45 Mark J.Kilgard. Improving Shadows and Reflections via the Stencil Buffer.
NVIDIA Corporation.
46 Randima Fernando. Percentage-Closer Soft Shadows. NVIDIA Corporation.
47 Fan Zhang, Hanqiu Sun, Leilei Xu. Parallel-Split Shadow Maps for Large-scale
Virtual Environments.
48 Uwe Behrens, Ralf Rastering. Adding shadows to a texture-based volume renderer.
In VVS98: Proceedings of the 1998 IEEE symposium on Volume visualization,
pages 39-46. ACM Press, 1998.
49 Johanna Beyer, Markus Hadwiger. Smooth Mixed-Resolution GPU Volume
Rendering. In IEEE/EG International Symposium on Volume and Point-Based
Graphics, pages 163-170, 2008.
50 Praveen Bhaniramka, Yves Demange. OpenGL Volumizer: A Toolkit for High
Quality Volume Rendering of Large Data Sets. In Proceedings IEEE Visualization
2002, pages 45-53, 2002.
51 Imma Boada, Isabel Navazo, Roberto Scopigno. Multiresolution volume
visualization with a texture-based octree. The Visual Computer, 17:185-197, 2001.
52 N. Carr, J.Hall, J.Hart. GPU Algorithms for Radiosity and Subsurface Scattering.
In proc. Graphics Hardware, 2003.
53 Nathan A. Carr, Jesse D.hall, and John C.Hart. GPU algorithms for radiosity and
subsurface scattering. In HWWS03: Proceedings of the conference on Graphics
Hardware 03, pages 51-59. Eurographics Association, 2003.
54 R.A.Drebin, L.Carpenter, and P. Hanrahan. Volume rendering. In Proceedings of
SIGGRAPH 88, pages 65-74, 1988.
185

GPU编程与 CG语言 中文版 PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

GPU编程与 CG语言 中文版 PDF

Uploaded by

Copyright:

Available Formats

GPU Programming And Cg Language Primer 1rd Edition

GPU Programming And Cg Language Primer 1rd Edition

-------Lewis Carroll Alices adventures in Wonderland

------Henry David Thoreau

1.1 Programmable Graphics Processing Unit

Programmable Graphics Processing UnitGPU

1998 NVIDIA modern GPU GPU

modern GPU transistorsmicrochip

Modern GPU 1998 4 NVIDIA

1999 2000 modern GPU GPU

2001 modern GPU GPU vertex

modern GPU 2002 2003 NVIDIA GeForce

1. vertex programmability fragment programmability

GPU Feng Liu Platform Independent

1.2 GPU VS CPU

GPU CPU GPUGPU

GPU ALUArithmetic Logic Unit

Control ALU ALU

GPU Visualization in Scientific Computing

GPU GPU CPU

color buffer frame buffer

Because, your application supplies the geometric data as a collection of vertices,

2.1.1 object space world space

When an artist creates a 3D model of an object, the artist selects a convenient

object space object space coordinate

world space coordinate

world coordinate space

object space GPU

2.1.2 world space eye space

eye space camera

2.1.3 eye space project and clip space

(-1,-1,-1)(1,1,1)Canonical view volume,

2.2 Primitive Assembly && Triangle setup

Primitive Assembly primitive,

View Frustum Culling

Rasterization is the process of

2.3.2 Pixel Operation

Pixel operation Raster Operation2 Raster

cd = aica + (1 a )cs over

Z buffer depth buffer

steve Baker Learning to love your

2.4.3 Stencil Buffer

Stencil buffer, buffer

2.4.4 Frame Buffer

A framebuffer is a video output device that drives a video display from a

Frame buffer buffer

shader language GLSL High Level Shading

shader language C++\JAVA

3.1 Shader Language

shader language shader program

GPU Programmable Vertex Processor

Vertex program Fragment program

3.2 Vertex Shader Program

Vertex shader program Fragment shader program

3.3 Fragment Shader Program

3.4 CG VS GLSL VS HLSL

Shader language 3 OpenGL GLSLOpenGL

GLSL HLSL OpenGL Direct3D

ATI OpenGL OSG

CgC for Graphcis NVIDIA Microsoft

Cg OpenGL Direct3D API

Cg Toolkit NVIDIA Corporation\Cg\bin

cgc [options] file

GPU编程与 CG语言中文版 PDF

GPU编程与 CG语言中文版 PDF