http://blog.sina.com.cn/s/blog_8c7d49f20102v4qm.html

Patch is just an ordered list of vertices (在tessellation shader里面比较重要的概念就是这个patch,patch是一系列的顶点,OpenGL规定patch的vertex数量必须至少大于等于3)

The tessellation process doesn’t operate on OpenGL’s classic geometric primitives: points, lines, and triangles, but uses a new primitive called a patch (Tessellation shader就是针对patch来进行处理的而并非点,县,三角形)

 

Patch – are processed by all of active shading stage in the pipeline (patch可被所有的shader处理)

Two shading stages of Tessellation shading:

1.      Tessellation Control Shader (optional)

In tessellation control shader, gl_PatchVerticesIn provides the number of elements in gl_in (gl_in 用于在tessellation control shader里面访问传进的来patch里面的顶点)

gl_out 用于控制patch里面的vertex从tessellation control shader输出后的属性

gl_InvocationID is used to access the specific vertex of a patch (gl_InvocationID 用于访问传入patch里的特定顶点)

Layout (vertices = *) out; (用于指定输出的patch里面有多少个顶点)

gl_TessLevelInner
              
Specify how the interior of the domain is subdivided and stored in
a two element array named gl_TessLevelInner(指定多边形内部如何细分)
         gl_TessLevelOuter

Control how the perimeter of the domain is subdivided, and is
stored in an implicitly declared four-element array named
gl_TessLevelOuter(指定多边形边界上的边被如何细分)
         
gl_TessLevelInner & gl_TessLevelOuter
根据多边形内部区域的类型会有不同的分割法

Tessellation Evaluation Shader (optional)

The TES is executed on all
generated domain
locations.Positions each of the
vertices in the final mesh (TES是针对从tessellation control shader
里面通过细分生成的顶点来进行运算,通过gl_in和gl_VocationID来访问我们传入的patch的一系列顶点信息,结合gl_TessCoord访问我们生成的新顶点的纹理信息来计算新的坐标位置,从而实现细分多边形的效果)

layout (quads, equal_spacing, ccw) in;
(指定新生成的多边形类型等信息)

glPatchParameteri() -- 告诉程序我们定义多少个顶点为一个patch

barrier() – .....

If you have additional per-vertex attribute values, either for
input or output, these need to be declared as either in or out
arrays in your tessellation control shader (需要传入TCS额外的顶点信息,需要定义额外in
& out array)

glPatchParameterfv() -- can be used to set the inner and
outer-tessellation levels(可以用于指定inner和outer的数值,当然我们也可以在tessellation
control shader里面通过gl_TessLevelInner &
gl_TessLevelOuter直接指定)

三中不同类型的domain -- 会决定我们inner和outer的具体含义:

Quad Tessellation:

…..

Isoline  Tessellation:

Use only two of the outer-tessellation levels to determine the
amount of subdivision

…..

Triangle Tessellation:

Triangular domains use barycentric coordinates to specify their
Tessellation coordinates

…..

最终渲染效果:

- Tessellation Shader" title="OpenGL - Tessellation Shader" height="291" width="452">

- Tessellation Shader" title="OpenGL - Tessellation Shader" height="318" width="448">

- Tessellation Shader" title="OpenGL - Tessellation Shader" height="291" width="455">

- Tessellation Shader" title="OpenGL - Tessellation Shader" height="298" width="444">

Main.cpp Source Code below:
#include

#include

#include

// TODO: 在此处引用程序需要的其他头文件
#include
#include "vgl.h"
#include "mat.h"
#include "LoadShaders.h"
#include "Shapes/Teapot.h"

using namespace std;

GLuint  PM;  // Projection
matrix
GLuint 
MVM;   
 // Model view matrix

GLuint  InnerL;  // Inner
tessellation paramter
GLuint  OuterL;  // Outer
tessellation paramter

GLfloat  Inner = 1.0;
GLfloat  Outer = 1.0;

#define SQUARE_VERTEX_NUMBER 4

//----------------------------------------------------------------------------

void init( void )
{
    //create
shader and link program first
    ShaderInfo
shaders[] = {
   
    {
GL_VERTEX_SHADER,         
"square.vert" },
   
    {
GL_TESS_CONTROL_SHADER,   
"square.cont" },
   
    {
GL_TESS_EVALUATION_SHADER, "square.eval" },
   
    {
GL_FRAGMENT_SHADER,       
"square.frag" },
   
    { GL_NONE,
NULL }
    };

//Shader
program
    GLuint
program;

//load
shaders
    program =
LoadShaders(shaders);

//link
shader program and use it
   
glUseProgram(program);

//vertex
array
    GLuint
VA;
   
glGenVertexArrays(1, &VA);
   
glBindVertexArray(VA);

//Vertex
buffer
    GLuint
VB;
   
glGenBuffers(1, &VB);
   
glBindBuffer(GL_ARRAY_BUFFER, VB);

//define
square vertex data
    const
GLfloat square_vertex_data[SQUARE_VERTEX_NUMBER][2] = {
   
    { 1.0, 1.0
},
   
    { -1.0,
1.0},
   
    { -1.0,
-1.0},
   
    { 1.0,
-1.0}
    };

//vertex
buffer data
   
glBufferData(GL_ARRAY_BUFFER, sizeof(square_vertex_data),
square_vertex_data, GL_STATIC_DRAW);

//Obtain
vertex position in program
    GLint
vPostionIndex = glGetAttribLocation(program, "vPosition");

//Enable
vertex attribute
   
glEnableVertexAttribArray( vPostionIndex );

//define an
array of generic vertex attribute data for shader
(让shader知道怎么去读取我们传入的vertex 信息)
   
glVertexAttribPointer(vPostionIndex, 2, GL_FLOAT, GL_FALSE, 0,
BUFFER_OFFSET(0));

//Obtain
uniform variable of program
    PM =
glGetUniformLocation(program, "PJ");
    MVM =
glGetUniformLocation(program, "MV");

InnerL =
glGetUniformLocation(program, "inner");
    OuterL =
glGetUniformLocation(program, "outer");

//Set some
default value for uniform variable
   
glUniform1f(InnerL, Inner);
   
glUniform1f(OuterL, Outer);

mat4  modelview = Translate( 0.0, 0.0, -2.0 )
*    RotateX(
-50.0 );
   
   
glUniformMatrix4fv( MVM, 1, GL_TRUE, modelview );

//Define how
many vertices composed one patch (定义多少个vertices定义一个patch)
   
glPatchParameteri(GL_PATCH_VERTICES, 4);

//Enable
some relative setting
   
glEnable(GL_DEPTH_TEST);

glClearColor( 0.0, 0.0, 0.0, 1.0 );
}

//----------------------------------------------------------------------------

void display( void )
{
    glClear(
GL_COLOR_BUFFER_BIT | GL_DEPTH_BUFFER_BIT );
   
   
glUniform1f(InnerL, Inner);

glUniform1f(OuterL, Outer);

glDrawArrays(GL_PATCHES, 0, SQUARE_VERTEX_NUMBER);

glutSwapBuffers();
}

//----------------------------------------------------------------------------

void reshape( int width, int height )
{
    glViewport(
0, 0, width, height );

GLfloat  aspect = GLfloat(width)/height;

mat4  projection = Perspective( 60.0, aspect, 1, 3
);
   
   
glUniformMatrix4fv( PM, 1, GL_TRUE, projection );

glutPostRedisplay();
}

//----------------------------------------------------------------------------

void keyboard( unsigned char key, int x, int y )
{
    switch ( key
) {
    case 'q':
case 'Q': case 033 :
   
    exit(
EXIT_SUCCESS );
   
    break;

case
'i':
   
   
Inner--;
   
    if ( Inner
< 1.0 )  Inner = 1.0;
   
    glUniform1f(
InnerL, Inner );
   
    break;

case
'I':
   
   
Inner++;
   
    if ( Inner
> 64 )  Inner = 64.0;
   
    glUniform1f(
InnerL, Inner );
   
    break;

case
'o':
   
   
Outer--;
   
    if ( Outer
< 1.0 )  Outer = 1.0;
   
    glUniform1f(
OuterL, Outer );
   
    break;

case
'O':
   
   
Outer++;
   
    if ( Outer
> 64 )  Outer = 64.0;
   
    glUniform1f(
OuterL, Outer );
   
    break;

case
'r':
   
    Inner =
1.0;
   
    Outer =
1.0;
   
    glUniform1f(
InnerL, Inner );
   
    glUniform1f(
OuterL, Outer );
   
    break;
   
    case 'm':
{
   
    static
GLenum mode = GL_FILL;
   
    mode = (
mode == GL_FILL ? GL_LINE : GL_FILL );
   
   
glPolygonMode( GL_FRONT_AND_BACK, mode );
   
   
   
  } break;
    }

glutPostRedisplay();
}

//----------------------------------------------------------------------------

int main( int argc, char *argv[] )
{
    glutInit(
&argc, argv );
   
glutInitDisplayMode( GLUT_RGBA | GLUT_DEPTH | GLUT_DOUBLE );
   
glutInitWindowSize( 512, 512 );
   
glutInitContextVersion( 3, 2 );
   
glutInitContextProfile( GLUT_CORE_PROFILE );
   
glutCreateWindow( "teapot" );

glewExperimental =
GL_TRUE;   
//注意这里有个坑 --
这一句很关键,不加这一句gl_GenVertexArray()会报错,好像说是GLEW对openGL的core
context有一些问题

//  
http://stackoverflow.com/questions/22813625/0xc0000005-access-violation-executing-location-0x00000000-opengl

GLenum error
= glewInit();

if( error !=
GLEW_OK )
    {
   
   
cout<<"glewInit failed, aborting"<<endl;
    }

if (
GLEW_ARB_vertex_array_object == NULL )
    {
   
   
cout<<"GLEW_ARB_vertex_array_object =
NULL"<<endl;
    }

cout<<"Error info:"<<glGetError()<<endl;

init();

glutDisplayFunc( display );
   
glutReshapeFunc( reshape );
   
glutKeyboardFunc( keyboard );

glutMainLoop();
    return
0;
}

square.vert source code below:

#version 400 core

in  vec4  vPosition;

void main()
{
   
//gl_Position is used to access the vertex position that is input
from application
    gl_Position
= vPosition;
}

square.cont source code below:

#version 400 core

//The main purpose of tessellation control shader is:
//Generate the tessellation output patch vertices that are passed
to the tessellation
//evaluation shader, as well as update any per-vertex

//Define how many vertices will be used as one patch
layout (vertices = 4) out;

//uniform type is used to define the variable that can be used to
communicate between shader and application
uniform float inner;

uniform float outer;

void main()
{
   
//gl_TessLevelInner is used to define:
    //how the
interior of the domain is subdivided and stored in a two element
array named
   
   
gl_TessLevelInner[0] = inner;
   
gl_TessLevelInner[1] = inner;
   
   
//gl_TessLevelOuter is used to define:
    //how the
perimeter of the domain is subdivided, and is stored in an
    //implicitly
declared four-element array
   
gl_TessLevelOuter[0] = outer;
   
gl_TessLevelOuter[1] = outer;
   
gl_TessLevelOuter[2] = outer;
   
gl_TessLevelOuter[3] = outer;
   
    //gl_in is
used to access the number of elements that are define by
glPatchParameteri()
   
//glPatchParameteri() define how many vertices as a patch
    //gl_out is
used to access the output vertex position of tessellation control
shader
    //gl_in
vertex shader structure below:
    //in
gl_PerVertex {
   
//   
    vec4
gl_Position;
   
//   
    float
gl_PointSize;
   
//   
    float
gl_ClipDistance[]
    //}
gl_in[gl_PatchVerticesIn];

//gl_out
vertex shader structure is similar to gl_in structure
   
//gl_InvocationID is used to access the specific vertex of a
patch
   
gl_out[gl_InvocationID].gl_Position = 
gl_in[gl_InvocationID].gl_Position;
}

square.eval source code below:

#version 400 core

//The main purpose of tessellation evaluation shader is:
//configure the primitive generator, which is done using a layout
directive
//Specifying the face winding for generated primitives
//(the order the vertices are issued determines the facedness of
the primitive)
layout (quads, equal_spacing, ccw) in;

//The TES is executed on all generated domain locations.
//The bound tessellation evaluation shader is
//executed one for each tessellation coordinate that the primitive
generator
//Tessellation Shaders emits, and is responsible for determining
the position
//of the vertex derived from the tessellation coordinate.

uniform mat4 PJ;

uniform mat4 MV;

#define M_PI   
   
3.14159265358979323846

//----------------------------------------------------------------------------

float Hanning( vec2 p )
{
    p -= 0.5; //
map unit square to [-.5, .5]

float r =
length( p );

r = cos(
M_PI * r / 2.0 );
    r *=
r;

return
r;
}

void main()
{
   
//gl_TessCoord is used to access the Tessellation coordinates that
are
    //generated
by tessellation control shader
    float u =
gl_TessCoord.x;
    float v =
gl_TessCoord.y;
   
    //use
Tessellation coordinates to calculate position for new vertex
that
    //is
generated by tessellation control shader
    #define
p(i)  gl_in[i].gl_Position

vec4 pos =
v*(u*p(0) + (1-u)*p(1)) + (1-v)*(u*p(3) + (1-u)*p(2));
    pos.z =
Hanning( gl_TessCoord.xy );
   
    gl_Position
= PJ * MV * pos;
}

square.frag source code below:

#version 400 core

out  vec4 fColor;

void main()
{
    fColor = (1
- gl_FragCoord.z) * vec4( 1.0, 0.0, 0.0, 1.0 );
}

总结:

tessellation shader是可选的shader,不是必须的

tessellation shader与vertex shader不一样,tessellation
shader是针对patch(一系列顶点)来处理而不是一个顶点 (因为tessellation
shader需要通过传入的patch(一系列顶点)来计算新顶点的位置信息)

tessellation control shader负责对patch的细分设定

tessellation evaluation
shader负责对TCS细分出来的顶点进行位置等信息运算从而实现LOD(level of detail --
根据与camera的距离不同而细分程度不同)等效果

Bezier曲线在这里是一种细分后位置的计算方法来实现曲面的平滑效果

tessellation shader有个重要的应用就是LOD(Level of
Detail),通过在判断物体与视线的距离来设定tessellation control里面的factor
level从而实现近细分多,远细分少的效果

还有一个应用叫displacement mapping,在tessellation evaluation
shader里面通过tessellation coordinate的值来映射纹理(sample a texture)

最新文章

  1. .NET重构(类型码的设计、重构方法)
  2. jquery easyui 1.4.1 API( CHM版)
  3. ebs R12.2启动报错&quot;failed to start a managed process after the maximum retry limit&quot;
  4. UISegmentedControl 的使用
  5. javabean连数据库
  6. 【Hadoop学习】HDFS 短路本地读
  7. [C++程序设计]返回指针值的函数
  8. 在SQL中用正则表达式替换html标签(2)
  9. EL表达式得不到后台传过来的值
  10. haproxy配置文件详解和ACL功能
  11. (模拟) codeVs1160 蛇形矩阵
  12. ecplise多个版本tomcat的使用
  13. Linux记录- Linux下限制rm -rf /
  14. windows 环境使用 kafka
  15. python基础知识-GUI编程-TK-StringVar
  16. unix网络编程——TCP套接字编程
  17. PHP正则表达式函数学习
  18. jersey 过滤器名称绑定的问题 NameBinding Provider
  19. sliverlight资源文件的URI调用
  20. Spring 中好用的泛型操作API

热门文章

  1. 智能DNS解析之edns-client-subnet篇
  2. Asp.net HttpWebRequest和HttpWebResponse发送和接受任何类型数据
  3. 大数据学习——akka学习
  4. 思路清奇:通过 JavaScript 获取移动设备的型号
  5. 使用 D8 分析 javascript 如何被 V8 引擎优化的
  6. 【Luogu】P3761城市(dfs)
  7. [SCOI2016] 背单词 (Trie 树,贪心)
  8. java面试题之有哪几种方式可以让线程阻塞
  9. Linux 之 rsync实现服务器的文件同步
  10. ORA-12514: TNS:listener does not currently know of service requested in connect