glsl チュートリアル
glslを使い始める
サーチ…
備考
このセクションでは、glslの概要と、なぜ開発者がそれを使いたいのかを概説します。
また、glsl内の大きなテーマについても言及し、関連するトピックにリンクする必要があります。 glslのドキュメンテーションは新しいので、それらの関連トピックの初期バージョンを作成する必要があるかもしれません。
バージョン
OpenGLシェーディング言語
| GLSLバージョン | OpenGLバージョン | シェーダプリプロセッサ | 発売日 |
|---|---|---|---|
| 1.10 | 2.0 | #version 110 | 2004-09-07 |
| 1.20 | 2.1 | #version 120 | 2006年7月2日 |
| 1.30 | 3.0 | #version 130 | 2008年8月11日 |
| 1.40 | 3.1 | #version 140 | 2009年3月24日 |
| 1.50 | 3.2 | #version 150 | 2009-08-03 |
| 3.30 | 3.3 | #version 330 | 2010-02-12 |
| 4.00 | 4.0 | #version 400 | 2010-03-11 |
| 4.10 | 4.1 | #version 410 | 2010-07-26 |
| 4.20 | 4.2 | #version 420 | 2011-08-08 |
| 4.30 | 4.3 | #version 430 | 2012-08-06 |
| 4.40 | 4.4 | #version 440 | 2013年7月22日 |
| 4.50 | 4.5 | #version 450 | 2014-08-11 |
OpenGL ESシェーディング言語
| GLSL ESバージョン | OpenGL ESバージョン | シェーダプリプロセッサ | 発売日 |
|---|---|---|---|
| 1.00 ES | 2.0 | #version 100 | 2007-03-05 |
| 3.00 ES | 3.0 | #version 300 es | 2012-08-06 |
| 3.10 ES | 3.1 | #version 310 es | 2014-03-17 |
| 3.20 ES | 3.2 | #version 320 es | 2015-08-10 |
インストールまたはセットアップ
glslのセットアップまたはインストールに関する詳細な手順。
最初のOGL 4.0 GLSLシェーダプログラム
頂点位置と色属性を持つ単純なOGL 4.0 GLSLシェーダプログラムです。プログラムはphytonスクリプトで実行されます。スクリプトを実行するには、PyOpenGLをインストールする必要があります。
シェーダプログラムは、少なくとも頂点シェーダとフレグマントシェーダから構成されます（コンピュータシェーダを除く）。 1番目のシェーダステージは頂点シェーダで、最後のシェーダステージはフラグメントシェーダです（その間に任意のステージが可能ですが、これ以上は説明しません）。
頂点シェーダ
first.vet
頂点シェーダは、描画コマンドで指定された頂点と関連する属性を処理します。頂点シェーダは、入力ストリームから頂点を処理し、任意の方法で頂点を操作できます。頂点シェーダは、入力ストリームから単一の頂点を受け取り、出力頂点ストリームに対して1つの単一の頂点を生成します。
この例では、単一の三角形を描画するので、頂点シェーダは三角形の各頂点に3回ずつ実行されます。この場合、頂点シェーダへの入力in vec3 inPos頂点位置in vec3 inPosのcolor属性in vec3 inCol 。カラー属性は、次のシェーダステージ（ out vec3 vertCol ）に渡されます。
#version 400
layout (location = 0) in vec3 inPos;
layout (location = 1) in vec3 inCol;
out vec3 vertCol;
void main()
{
vertCol = inCol;
gl_Position = vec4( inPos, 1.0 );
}
フラグメントシェーダ
first.frag
この例では、頂点シェーダの直後にフラグメントシェーダが続きます。頂点の位置と属性は、各フラグメントの各面内で補間されます。フラグメントシェーダは、三角形全体のフラグメントごとに1回実行され、フレッジシェーダからカラー属性を受け取ります。三角形が描かれるので、描かれた三角形に基づくフラグメントの重心座標に従って色属性が補間される。
#version 400
in vec3 vertCol;
out vec4 fragColor;
void main()
{
fragColor = vec4( vertCol, 1.0 );
}
Phytonスクリプト
Pythonスクリプトは、シェーダプログラムをコンパイル、リンク、実行し、ジオメトリを描画するだけです。 Cや他の何かで自明に書き直すことができます。この文書では、最も注意を払うべきではありません。
from OpenGL.GL import *
from OpenGL.GLUT import *
from OpenGL.GLU import *
from sys import *
from array import array
# draw event
def OnDraw():
glClear( GL_COLOR_BUFFER_BIT | GL_DEPTH_BUFFER_BIT )
glUseProgram( shaderProgram )
glBindVertexArray( vaObj )
glDrawArrays( GL_TRIANGLES, 0, 3 )
glutSwapBuffers()
# read vertex shader program
with open( 'first.vert', 'r' ) as vertFile:
vertCode = vertFile.read()
print( '\nvertex shader code:' )
print( vertCode )
# read fragment shader program
with open( 'first.frag', 'r' ) as fragFile:
fragCode = fragFile.read()
print( '\nfragment shader code:' )
print( fragCode )
# initialize glut
glutInit()
# create window
wndW = 800
wndH = 600
glutInitDisplayMode(GLUT_RGBA | GLUT_DOUBLE | GLUT_ALPHA | GLUT_DEPTH)
glutInitWindowPosition(0, 0)
glutInitWindowSize(wndW, wndH)
wndID = glutCreateWindow(b'OGL window')
glutDisplayFunc(OnDraw)
glutIdleFunc(OnDraw)
# define triangle data
posData = [ -0.636, -0.45, 0.0, 0.636, -0.45, 0.0, 0.0, 0.9, 0.0 ]
colData = [ 1.0, 0.0, 0.0, 1.0, 1.0, 0.0, 0.0, 0.0, 1.0 ]
posAr = array( "f", posData )
colAr = array( "f", colData )
# create buffers
posBuffer = glGenBuffers(1)
glBindBuffer( GL_ARRAY_BUFFER, posBuffer )
glBufferData( GL_ARRAY_BUFFER, posAr.tostring(), GL_STATIC_DRAW )
colBuffer = glGenBuffers(1)
glBindBuffer( GL_ARRAY_BUFFER, colBuffer )
glBufferData( GL_ARRAY_BUFFER, colAr.tostring(), GL_STATIC_DRAW )
# create vertex array opject
vaObj = glGenVertexArrays( 1 )
glBindVertexArray( vaObj )
glEnableVertexAttribArray( 0 )
glEnableVertexAttribArray( 1 )
glBindBuffer( GL_ARRAY_BUFFER, posBuffer )
glVertexAttribPointer( 0, 3, GL_FLOAT, GL_FALSE, 0, None )
glBindBuffer( GL_ARRAY_BUFFER, colBuffer )
glVertexAttribPointer( 1, 3, GL_FLOAT, GL_FALSE, 0, None )
# compile vertex shader
vertShader = glCreateShader( GL_VERTEX_SHADER )
glShaderSource( vertShader, vertCode )
glCompileShader( vertShader )
result = glGetShaderiv( vertShader, GL_COMPILE_STATUS )
if not (result):
print( glGetShaderInfoLog( vertShader ) )
sys.exit()
# compile fragment shader
fragShader = glCreateShader( GL_FRAGMENT_SHADER )
glShaderSource( fragShader, fragCode )
glCompileShader( fragShader )
result = glGetShaderiv( fragShader, GL_COMPILE_STATUS )
if not (result):
print( glGetShaderInfoLog( fragShader ) )
sys.exit()
# link shader program
shaderProgram = glCreateProgram()
glAttachShader( shaderProgram, vertShader )
glAttachShader( shaderProgram, fragShader )
glLinkProgram( shaderProgram )
result = glGetProgramiv( shaderProgram, GL_LINK_STATUS )
if not (result):
print( 'link error:' )
print( glGetProgramInfoLog( shaderProgram ) )
sys.exit()
# start main loop
glutMainLoop()
OGL 4.0 GLSLでのモデル、ビュー、プロジェクションマトリックスの使用
モデル、ビュー、および投影行列の使用を示すシンプルなOGL 4.0 GLSLシェーダプログラムです。プログラムはフィトンスクリプトで実行されます。スクリプトを実行するには、PyOpenGLとNumPyをインストールする必要があります。
投影行列：投影行列は、世界の3D点からビューポートの2D点へのピンホールカメラのマッピングを記述する。この例では、視野が90度の投影行列を使用します。
ビュー行列：ビュー行列は、 目の位置とシーンに視線方向を規定します。この例では、シーンの中心に向かって視聴方向を維持しながらシーンの周りを循環しています。
モデル行列：モデル行列は、シーン内のオブジェクトの位置と相対的なサイズを定義します。この例では、モデル行列はオブジェクトを上下に移動します。
頂点シェーダ
mvp.vet
#version 400
layout (location = 0) in vec3 inPos;
layout (location = 1) in vec3 inCol;
out vec3 vertCol;
uniform mat4 projectionMat44;
uniform mat4 viewMat44;
uniform mat4 modelMat44;
void main()
{
vertCol = inCol;
vec4 modolPos = modelMat44 * vec4( inPos, 1.0 );
vec4 viewPos = viewMat44 * modolPos;
gl_Position = projectionMat44 * viewPos;
}
フラグメントシェーダ
mvp.frag
#version 400
in vec3 vertCol;
out vec4 fragColor;
void main()
{
fragColor = vec4( vertCol, 1.0 );
}
Phytonスクリプト
from OpenGL.GL import *
from OpenGL.GLUT import *
from OpenGL.GLU import *
import numpy as np
from time import time
import math
import sys
# draw event
def OnDraw():
currentTime = time()
# set up projection matrix
prjMat = perspective( 90.0, wndW/wndH, 0.5, 100.0)
# set up view matrix
viewMat = Translate( np.matrix(np.identity(4), copy=False, dtype='float32'), np.array( [0.0, 0.0, -8.0] ) )
viewMat = RotateView( viewMat, [10.0, CalcAng( currentTime, 10.0 ), 0.0] )
# set up tetrahedron model matrix
tetModelMat = np.matrix(np.identity(4), copy=False, dtype='float32')
tetModelMat = RotateX( tetModelMat, -90.0 )
tetModelMat = Scale( tetModelMat, np.repeat( 2.0, 3 ) )
tetModelMat = Translate( tetModelMat, np.array( [-2.0, 0.0, CalcMove(currentTime, 6.0, [-1.0, 1.0])] ) )
# set up icosahedron model matrix
icoModelMat = np.matrix(np.identity(4), copy=False, dtype='float32')
icoModelMat = RotateX( icoModelMat, -90.0 )
icoModelMat = Scale( icoModelMat, np.repeat( 2.0, 3 ) )
icoModelMat = Translate( icoModelMat, np.array( [2.0, 0.0, CalcMove(currentTime, 6.0, [1.0, -1.0])] ) )
# set up attributes and shader program
glEnable( GL_DEPTH_TEST )
glClear( GL_COLOR_BUFFER_BIT | GL_DEPTH_BUFFER_BIT )
glUseProgram( shaderProgram )
glUniformMatrix4fv( projectionMatLocation, 1, GL_FALSE, prjMat )
glUniformMatrix4fv( viewMatLocation, 1, GL_FALSE, viewMat )
# draw tetrahedron
glUniformMatrix4fv( modelMatLocation, 1, GL_FALSE, tetModelMat )
glBindVertexArray( tetVAObj )
glDrawElements(GL_TRIANGLES, len(tetIndices), GL_UNSIGNED_INT, tetIndices)
# draw tetrahedron
glUniformMatrix4fv( modelMatLocation, 1, GL_FALSE, icoModelMat )
glBindVertexArray( icoVAObj )
glDrawArrays( GL_TRIANGLES, 0, len(icoPosData) )
glutSwapBuffers()
def Fract(val): return val - math.trunc(val)
def CalcAng(currentTime, intervall): return Fract( (currentTime - startTime) / intervall ) * 360.0
def CalcMove(currentTime, intervall, range):
pos = Fract( (currentTime - startTime) / intervall ) * 2.0
pos = pos if pos < 1.0 else (2.0-pos)
return range[0] + (range[1] - range[0]) * pos
# read shader program and compile shader
def CompileShader( sourceFileName, shaderStage ):
with open( sourceFileName, 'r' ) as sourceFile:
sourceCode = sourceFile.read()
nameMap = { GL_VERTEX_SHADER: 'vertex', GL_FRAGMENT_SHADER: 'fragment' }
print( '\n%s shader code:' % nameMap.get(shaderStage, '') )
print( sourceCode )
shaderObj = glCreateShader( shaderStage )
glShaderSource( shaderObj, sourceCode )
glCompileShader( shaderObj )
result = glGetShaderiv( shaderObj, GL_COMPILE_STATUS )
if not (result):
print( glGetShaderInfoLog( shaderObj ) )
sys.exit()
return shaderObj
# linke shader objects to shader program
def LinkProgram( shaderObjs ):
shaderProgram = glCreateProgram()
for shObj in shaderObjs:
glAttachShader( shaderProgram, shObj )
glLinkProgram( shaderProgram )
result = glGetProgramiv( shaderProgram, GL_LINK_STATUS )
if not (result):
print( 'link error:' )
print( glGetProgramInfoLog( shaderProgram ) )
sys.exit()
return shaderProgram
# create vertex array opject
def CreateVAO( dataArrays ):
noOfBuffers = len(dataArrays)
buffers = glGenBuffers(noOfBuffers)
newVAObj = glGenVertexArrays( 1 )
glBindVertexArray( newVAObj )
for inx in range(0, noOfBuffers):
vertexSize, dataArr = dataArrays[inx]
arr = np.array( dataArr, dtype='float32' )
glBindBuffer( GL_ARRAY_BUFFER, buffers[inx] )
glBufferData( GL_ARRAY_BUFFER, arr, GL_STATIC_DRAW )
glEnableVertexAttribArray( inx )
glVertexAttribPointer( inx, vertexSize, GL_FLOAT, GL_FALSE, 0, None )
return newVAObj
def Translate(matA, trans):
matB = np.copy(matA)
for i in range(0, 4): matB[3,i] = matA[0,i] * trans[0] + matA[1,i] * trans[1] + matA[2,i] * trans[2] + matA[3,i]
return matB
def Scale(matA, s):
matB = np.copy(matA)
for i0 in range(0, 3):
for i1 in range(0, 4): matB[i0,i1] = matA[i0,i1] * s[i0]
return matB
def RotateHlp(matA, angDeg, a0, a1):
matB = np.copy(matA)
ang = math.radians(angDeg)
sinAng, cosAng = math.sin(ang), math.cos(ang)
for i in range(0, 4):
matB[a0,i] = matA[a0,i] * cosAng + matA[a1,i] * sinAng
matB[a1,i] = matA[a0,i] * -sinAng + matA[a1,i] * cosAng
return matB
def RotateX(matA, angDeg): return RotateHlp(matA, angDeg, 1, 2)
def RotateY(matA, angDeg): return RotateHlp(matA, angDeg, 2, 0)
def RotateZ(matA, angDeg): return RotateHlp(matA, angDeg, 0, 1)
def RotateView(matA, angDeg): return RotateZ(RotateY(RotateX(matA, angDeg[0]), angDeg[1]), angDeg[2])
def perspective(fov, aspectRatio, near, far):
fn, f_n = far + near, far - near
r, t = aspectRatio, 1.0 / math.tan( math.radians(fov) / 2.0 )
return np.matrix( [ [t/r,0,0,0], [0,t,0,0], [0,0,-fn/f_n,-2.0*far*near/f_n], [0,0,-1,0] ] )
# initialize glut
glutInit()
# create window
wndW, wndH = 800, 600
glutInitDisplayMode(GLUT_RGBA | GLUT_DOUBLE | GLUT_ALPHA | GLUT_DEPTH)
glutInitWindowPosition(0, 0)
glutInitWindowSize(wndW, wndH)
wndID = glutCreateWindow(b'OGL window')
glutDisplayFunc(OnDraw)
glutIdleFunc(OnDraw)
# define tetrahedron vertex array opject
sin120 = 0.8660254
tetPposData = [ 0.0, 0.0, 1.0, 0.0, -sin120, -0.5, sin120 * sin120, 0.5 * sin120, -0.5, -sin120 * sin120, 0.5 * sin120, -0.5 ]
tetColData = [ 1.0, 0.0, 0.0, 1.0, 1.0, 0.0, 0.0, 0.0, 1.0, 0.0, 1.0, 0.0, ]
tetIndices = [ 0, 1, 2, 0, 2, 3, 0, 3, 1, 1, 3, 2 ]
tetVAObj = CreateVAO( [ (3, tetPposData), (3, tetColData) ] )
tetInxArr = np.array( tetIndices, dtype='uint' )
# define icosahedron vertex array opject
icoPts = [
[ 0.000, 0.000, 1.000], [ 0.894, 0.000, 0.447], [ 0.276, 0.851, 0.447], [-0.724, 0.526, 0.447],
[-0.724, -0.526, 0.447], [ 0.276, -0.851, 0.447], [ 0.724, 0.526, -0.447], [-0.276, 0.851, -0.447],
[-0.894, 0.000, -0.447], [-0.276, -0.851, -0.447], [ 0.724, -0.526, -0.447], [ 0.000, 0.000, -1.000] ]
icoCol = [ [1.0, 0.0, 0.0], [0.0, 0.0, 1.0], [1.0, 1.0, 0.0], [0.0, 1.0, 0.0], [1.0, 0.5, 0.0], [1.0, 0.0, 1.0] ]
icoIndices = [
2, 0, 1, 3, 0, 2, 4, 0, 3, 5, 0, 4, 1, 0, 5, 11, 7, 6, 11, 8, 7, 11, 9, 8, 11, 10, 9, 11, 6, 10,
1, 6, 2, 2, 7, 3, 3, 8, 4, 4, 9, 5, 5, 10, 1, 2, 6, 7, 3, 7, 8, 4, 8, 9, 5, 9, 10, 1, 10, 6 ]
icoPosData = []
for inx in icoIndices:
for inx_s in range(0, 3):
icoPosData.append( icoPts[inx][inx_s] )
icoColData = []
for inx in range(0, len(icoPosData) // 9):
inx_col = inx % len(icoCol)
for inx_p in range(0, 3):
for inx_s in range(0, 3):
icoColData.append( icoCol[inx_col][inx_s] )
icoVAObj = CreateVAO( [ (3, icoPosData), (3, icoColData) ] )
# load, compile and link shader
shaderProgram = LinkProgram( [
CompileShader( 'mvp.vert', GL_VERTEX_SHADER ),
CompileShader( 'mvp.frag', GL_FRAGMENT_SHADER )
] )
projectionMatLocation = glGetUniformLocation(shaderProgram, "projectionMat44")
viewMatLocation = glGetUniformLocation(shaderProgram, "viewMat44")
modelMatLocation = glGetUniformLocation(shaderProgram, "modelMat44")
# start main loop
startTime = time()
glutMainLoop()
モデルにテクスチャを置き、OGL 4.0 GLSLでテクスチャマトリックスを使用する
メッシュ上に2Dテクスチャをマッピングする方法を示すシンプルなOGL 4.0 GLSLシェーダプログラムです。プログラムはphytonスクリプトで実行されます。スクリプトを実行するには、PyOpenGLとNumPyをインストールする必要があります。
テクスチャマトリクスは、テクスチャがどのようにメッシュにマッピングされるかを定義します。テクスチャマトリクスを操作することによって、テクスチャを移動、拡大縮小、回転させることができます。
頂点シェーダ
tex.vert
#version 400
layout (location = 0) in vec3 inPos;
layout (location = 1) in vec2 inTex;
out vec2 vertTex;
uniform mat4 u_projectionMat44;
uniform mat4 u_viewMat44;
uniform mat4 u_modelMat44;
uniform mat4 u_textureMat44;
void main()
{
vertTex = ( u_textureMat44 * vec4( inTex, 0.0, 1.0 ) ).st;
vec4 modolPos = u_modelMat44 * vec4( inPos, 1.0 );
vec4 viewPos = u_viewMat44 * modolPos;
gl_Position = u_projectionMat44 * viewPos;
}
フラグメントシェーダ
tex.frag
#version 400
in vec2 vertTex;
out vec4 fragColor;
uniform sampler2D u_texture;
void main()
{
vec4 texCol = texture( u_texture, vertTex.st );
fragColor = vec4( texCol.rgb, 1.0 );
}
Phytonスクリプト
from OpenGL.GL import *
from OpenGL.GLUT import *
from OpenGL.GLU import *
import numpy as np
from time import time
import math
import sys
# draw event
def OnDraw():
currentTime = time()
# set up projection matrix
prjMat = perspective( 90.0, wndW/wndH, 0.5, 100.0)
# set up view matrix
viewMat = Translate( np.matrix(np.identity(4), copy=False, dtype='float32'), np.array( [0.0, 0.0, -15.0] ) )
viewMat = RotateView( viewMat, [30.0, CalcAng( currentTime, 60.0 ), 0.0] )
# set up tetrahedron model matrix
cubeModelMat = np.matrix(np.identity(4), copy=False, dtype='float32')
cubeModelMat = RotateX( cubeModelMat, -90.0 )
cubeModelMat = Scale( cubeModelMat, np.repeat( 5.0, 3 ) )
# set up texture matrix
texMat = np.matrix(np.identity(4), copy=False, dtype='float32')
deltaT = Fract( (currentTime - startTime) / 28.0 ) * 28.0
if deltaT < 7.0 or deltaT >= 21.0:
texMat = Scale( texMat, np.repeat( CalcMove(currentTime, 7.0, [1.0, 2.0]), 3 ) )
if deltaT >= 7.0 and deltaT < 14.0 or deltaT >= 21.0:
transAng = math.radians( CalcAng(currentTime, 7.0) )
texMat = Translate( texMat, np.array( [math.sin(transAng)*0.5, math.cos(transAng)*0.5-0.5, 0.0] ) )
if deltaT >= 14.0:
texMat = RotateZ( texMat, CalcAng(currentTime, 7.0) )
# set up attributes and shader program
glEnable( GL_DEPTH_TEST )
glClear( GL_COLOR_BUFFER_BIT | GL_DEPTH_BUFFER_BIT )
glUseProgram( shaderProgram )
glUniformMatrix4fv( projectionMatLocation, 1, GL_FALSE, prjMat )
glUniformMatrix4fv( viewMatLocation, 1, GL_FALSE, viewMat )
glUniformMatrix4fv( textureMatLocation, 1, GL_FALSE, texMat )
glUniform1i( textureLocation, 0 )
# draw cube
glUniformMatrix4fv( modelMatLocation, 1, GL_FALSE, cubeModelMat )
glBindVertexArray( cubeVAObj )
glDrawElements(GL_TRIANGLES, len(cubeIndices), GL_UNSIGNED_INT, cubeIndices)
glutSwapBuffers()
def Fract(val): return val - math.trunc(val)
def CalcAng(currentTime, intervall): return Fract( (currentTime - startTime) / intervall ) * 360.0
def CalcMove(currentTime, intervall, range):
pos = Fract( (currentTime - startTime) / intervall ) * 2.0
pos = pos if pos < 1.0 else (2.0-pos)
return range[0] + (range[1] - range[0]) * pos
# read shader program and compile shader
def CompileShader( sourceFileName, shaderStage ):
with open( sourceFileName, 'r' ) as sourceFile:
sourceCode = sourceFile.read()
nameMap = { GL_VERTEX_SHADER: 'vertex', GL_FRAGMENT_SHADER: 'fragment' }
print( '\n%s shader code:' % nameMap.get(shaderStage, '') )
print( sourceCode )
shaderObj = glCreateShader( shaderStage )
glShaderSource( shaderObj, sourceCode )
glCompileShader( shaderObj )
result = glGetShaderiv( shaderObj, GL_COMPILE_STATUS )
if not (result):
print( glGetShaderInfoLog( shaderObj ) )
sys.exit()
return shaderObj
# linke shader objects to shader program
def LinkProgram( shaderObjs ):
shaderProgram = glCreateProgram()
for shObj in shaderObjs:
glAttachShader( shaderProgram, shObj )
glLinkProgram( shaderProgram )
result = glGetProgramiv( shaderProgram, GL_LINK_STATUS )
if not (result):
print( 'link error:' )
print( glGetProgramInfoLog( shaderProgram ) )
sys.exit()
return shaderProgram
# create vertex array object
def CreateVAO( dataArrays ):
noOfBuffers = len(dataArrays)
buffers = glGenBuffers(noOfBuffers)
newVAObj = glGenVertexArrays( 1 )
glBindVertexArray( newVAObj )
for inx in range(0, noOfBuffers):
vertexSize, dataArr = dataArrays[inx]
arr = np.array( dataArr, dtype='float32' )
glBindBuffer( GL_ARRAY_BUFFER, buffers[inx] )
glBufferData( GL_ARRAY_BUFFER, arr, GL_STATIC_DRAW )
glEnableVertexAttribArray( inx )
glVertexAttribPointer( inx, vertexSize, GL_FLOAT, GL_FALSE, 0, None )
return newVAObj
def Translate(matA, trans):
matB = np.copy(matA)
for i in range(0, 4): matB[3,i] = matA[0,i] * trans[0] + matA[1,i] * trans[1] + matA[2,i] * trans[2] + matA[3,i]
return matB
def Scale(matA, s):
matB = np.copy(matA)
for i0 in range(0, 3):
for i1 in range(0, 4): matB[i0,i1] = matA[i0,i1] * s[i0]
return matB
def RotateHlp(matA, angDeg, a0, a1):
matB = np.copy(matA)
ang = math.radians(angDeg)
sinAng, cosAng = math.sin(ang), math.cos(ang)
for i in range(0, 4):
matB[a0,i] = matA[a0,i] * cosAng + matA[a1,i] * sinAng
matB[a1,i] = matA[a0,i] * -sinAng + matA[a1,i] * cosAng
return matB
def RotateX(matA, angDeg): return RotateHlp(matA, angDeg, 1, 2)
def RotateY(matA, angDeg): return RotateHlp(matA, angDeg, 2, 0)
def RotateZ(matA, angDeg): return RotateHlp(matA, angDeg, 0, 1)
def RotateView(matA, angDeg): return RotateZ(RotateY(RotateX(matA, angDeg[0]), angDeg[1]), angDeg[2])
def perspective(fov, aspectRatio, near, far):
fn, f_n = far + near, far - near
r, t = aspectRatio, 1.0 / math.tan( math.radians(fov) / 2.0 )
return np.matrix( [ [t/r,0,0,0], [0,t,0,0], [0,0,-fn/f_n,-2.0*far*near/f_n], [0,0,-1,0] ] )
# initialize glut
glutInit()
# create window
wndW, wndH = 800, 600
glutInitDisplayMode(GLUT_RGBA | GLUT_DOUBLE | GLUT_ALPHA | GLUT_DEPTH)
glutInitWindowPosition(0, 0)
glutInitWindowSize(wndW, wndH)
wndID = glutCreateWindow(b'OGL window')
glutDisplayFunc(OnDraw)
glutIdleFunc(OnDraw)
# define cube vertex array opject
icoPts = [
[-1.0, -1.0, 1.0], [ 1.0, -1.0, 1.0], [ 1.0, 1.0, 1.0], [-1.0, 1.0, 1.0],
[-1.0, -1.0, -1.0], [ 1.0, -1.0, -1.0], [ 1.0, 1.0, -1.0], [-1.0, 1.0, -1.0] ]
cubePosData = []
for inx in [ 0, 1, 2, 3, 1, 5, 6, 2, 5, 4, 7, 6, 4, 0, 3, 7, 3, 2, 6, 7, 1, 0, 4, 5 ]:
for inx_s in range(0, 3): cubePosData.append( icoPts[inx][inx_s] )
cubeTexData = []
for inx in range(0, 6):
for texCoord in [-0.5, -0.5, 0.5, -0.5, 0.5, 0.5, -0.5, 0.5]: cubeTexData.append( texCoord )
icoCol = [ [1.0, 0.0, 0.0], [1.0, 0.5, 0.0], [1.0, 0.0, 1.0], [1.0, 1.0, 0.0], [0.0, 1.0, 0.0], [0.0, 0.0, 1.0] ]
cubeIndices = []
for inx in range(0, 6):
for inx_s in [0, 1, 2, 0, 2, 3]: cubeIndices.append( inx * 4 + inx_s )
cubeVAObj = CreateVAO( [ (3, cubePosData), (2, cubeTexData) ] )
cubeInxArr = np.array( cubeIndices, dtype='uint' )
# load, compile and link shader
shaderProgram = LinkProgram( [
CompileShader( 'python/ogl4tex/tex.vert', GL_VERTEX_SHADER ),
CompileShader( 'python/ogl4tex/tex.frag', GL_FRAGMENT_SHADER )
] )
projectionMatLocation = glGetUniformLocation(shaderProgram, "u_projectionMat44")
viewMatLocation = glGetUniformLocation(shaderProgram, "u_viewMat44")
modelMatLocation = glGetUniformLocation(shaderProgram, "u_modelMat44")
textureMatLocation = glGetUniformLocation(shaderProgram, "u_textureMat44")
textureLocation = glGetUniformLocation(shaderProgram, "u_texture")
# create texture
texCX, texCY = 128, 128
texPlan = np.zeros( texCX * texCY * 4, dtype=np.uint8 )
for inx_x in range(0, texCX):
for inx_y in range(0, texCY):
val_x = math.sin( math.pi * 6.0 * inx_x / texCX )
val_y = math.sin( math.pi * 6.0 * inx_y / texCY )
inx_tex = inx_y * texCX * 4 + inx_x * 4
texPlan[inx_tex + 0] = int( 128 + 127 * val_x )
texPlan[inx_tex + 1] = 63
texPlan[inx_tex + 2] = int( 128 + 127 * val_y )
texPlan[inx_tex + 3] = 255
glActiveTexture( GL_TEXTURE0 )
texObj = glGenTextures( 1 )
glBindTexture( GL_TEXTURE_2D, texObj )
glTexImage2D(GL_TEXTURE_2D, 0, GL_RGBA, texCX, texCY, 0, GL_RGBA, GL_UNSIGNED_BYTE, texPlan)
glTexParameterf(GL_TEXTURE_2D, GL_TEXTURE_MAG_FILTER, GL_LINEAR)
glTexParameterf(GL_TEXTURE_2D, GL_TEXTURE_MIN_FILTER, GL_LINEAR)
glTexParameterf(GL_TEXTURE_2D, GL_TEXTURE_WRAP_S, GL_REPEAT)
glTexParameterf(GL_TEXTURE_2D, GL_TEXTURE_WRAP_T, GL_REPEAT)
# start main loop
startTime = time()
glutMainLoop()
インターフェイスブロックとユニフォームブロックの使用：OGL 4.0 GLSLのCook-Torranceライトモデル
簡単なOGL 4.0 GLSLシェーダプログラムで、Cook-Torranceマイクロファセットライトモデル実装でインターフェイスブロックと均一ブロックの使用を示します。プログラムはphytonスクリプトで実行されます。スクリプトを実行するには、PyOpenGLとNumPyをインストールする必要があります。
インターフェースブロックは、GLSLの入力、出力、統一、またはストレージバッファ変数のグループです。制服Blockisは、ストレージ修飾子を持つインタフェースブロックであるuniform 。
頂点シェーダ
ibub.vert
#version 400
layout (location = 0) in vec3 inPos;
layout (location = 1) in vec3 inNV;
layout (location = 2) in vec3 inCol;
out TVertexData
{
vec3 pos;
vec3 nv;
vec3 col;
} outData;
uniform mat4 u_projectionMat44;
uniform mat4 u_modelViewMat44;
uniform mat3 u_normalMat33;
void main()
{
vec4 viewPos = u_modelViewMat44 * vec4( inPos, 1.0 );
outData.pos = viewPos.xyz / viewPos.w;
outData.nv = u_normalMat33 * normalize( inNV );
outData.col = inCol;
gl_Position = u_projectionMat44 * viewPos;
}
フラグメントシェーダ
ibub.frag
#version 400
in TVertexData
{
vec3 pos;
vec3 nv;
vec3 col;
} inData;
out vec4 fragColor;
uniform UB_material
{
float u_roughness;
float u_fresnel0;
vec4 u_specularTint;
};
struct TLightSource
{
vec4 ambient;
vec4 diffuse;
vec4 specular;
vec4 dir;
};
uniform UB_lightSource
{
TLightSource u_lightSource;
};
vec3 CookTorrance( vec3 esPt, vec3 esPtNV, vec3 col, vec4 specularTint, float roughness, float fresnel0 )
{
vec3 esVLight = normalize( -u_lightSource.dir.xyz );
vec3 esVEye = normalize( -esPt );
vec3 halfVector = normalize( esVEye + esVLight );
vec3 reflVector = normalize( reflect( -esVLight, esPtNV ) );
float VdotR = dot( esVEye, reflVector );
float HdotL = dot( halfVector, esVLight );
float NdotL = dot( esPtNV, esVLight );
float NdotV = dot( esPtNV, esVEye );
float NdotH = dot( esPtNV, halfVector );
float NdotH2 = NdotH * NdotH;
float NdotL_clamped = max( NdotL, 0.0 );
float NdotV_clamped = max( NdotV, 0.0 );
float m2 = roughness * roughness;
// Lambertian diffuse
float k_diffuse = NdotL_clamped;
// Cook-Torrance fresnel
float theta = HdotL;
float n = (1.0 + sqrt(fresnel0)) / (1.0 - sqrt(fresnel0));
float g = sqrt( n*n + theta * theta + 1.0 );
float gc = g + theta;
float g_c = g - theta;
float q = (gc * theta - 1.0) / (g_c * theta + 1.0);
float fresnel = 0.5 * (g_c * g_c) / (gc * gc) * (1.0 + q * q);
// Gaussian distribution
float psi = acos( VdotR );
float distribution = max( 0.0, HdotL * exp( - psi * psi / m2 ) );
// Torrance-Sparrow geometric term
float geometric_att = min( 1.0, min( 2.0 * NdotH * NdotV_clamped / HdotL, 2.0 * NdotH * NdotL_clamped / HdotL ) );
// Microfacet bidirectional reflectance distribution function
float brdf_spec = fresnel * distribution * geometric_att / ( 4.0 * NdotL_clamped * NdotV_clamped );
float k_specular = brdf_spec;
vec3 lightColor = col.rgb * u_lightSource.ambient.rgb
+ max( 0.0, k_diffuse ) * col.rgb * u_lightSource.diffuse.rgb +
+ max( 0.0, k_specular ) * mix( col.rgb, specularTint.rgb, specularTint.a ) * u_lightSource.specular.rgb;
return lightColor;
}
void main()
{
vec3 lightCol = CookTorrance( inData.pos, inData.nv, inData.col, u_specularTint, u_roughness, u_fresnel0 );
fragColor = vec4( lightCol, 1.0 );
}
Phytonスクリプト
from OpenGL.GL import *
from OpenGL.GLUT import *
from OpenGL.GLU import *
import numpy as np
from time import time
import math
import sys
sin120 = 0.8660254
rotateCamera = False
# draw event
def OnDraw():
dist = 3.0
currentTime = time()
comeraRotAng = CalcAng( currentTime, 10.0 )
# set up projection matrix
prjMat = Perspective(90.0, wndW/wndH, 0.5, 100.0)
# set up view matrix
viewMat = Translate( np.matrix(np.identity(4), copy=False, dtype='float32'), np.array( [0.0, 0.0, -12.0] ) )
viewMat = RotateView( viewMat, [30.0, comeraRotAng if rotateCamera else 0.0, 0.0] )
# set up light source
lightSourceBuffer.BindDataFloat(b'u_lightSource.dir', TransformVec4([-3.0, -2.0, -1.0, 0.0], viewMat) )
# set up tetrahedron model matrix
tetModelMat = np.matrix(np.identity(4), copy=False, dtype='float32')
if not rotateCamera: tetModelMat = RotateY( tetModelMat, comeraRotAng )
tetModelMat = RotateX( tetModelMat, -90.0 )
tetModelMat = Scale( tetModelMat, np.repeat( 2.4, 3 ) )
tetModelMat = Translate( tetModelMat, np.array( [0.0, dist, 0.0] ) )
tetModelMat = RotateY( tetModelMat, CalcAng( currentTime, 20.0 ) )
tetModelMat = RotateX( tetModelMat, CalcAng( currentTime, 9.0 ) )
# set up icosahedron model matrix
icoModelMat = np.matrix(np.identity(4), copy=False, dtype='float32')
if not rotateCamera: icoModelMat = RotateY( icoModelMat, comeraRotAng )
icoModelMat = RotateX( icoModelMat, -90.0 )
icoModelMat = Scale( icoModelMat, np.repeat( 2.0, 3 ) )
icoModelMat = Translate( icoModelMat, np.array( [dist * -sin120, dist * -0.5, 0.0] ) )
icoModelMat = RotateY( icoModelMat, CalcAng( currentTime, 20.0 ) )
icoModelMat = RotateX( icoModelMat, CalcAng( currentTime, 11.0 ) )
# set up cube model matrix
cubeModelMat = np.matrix(np.identity(4), copy=False, dtype='float32')
if not rotateCamera: cubeModelMat = RotateY( cubeModelMat, comeraRotAng )
cubeModelMat = RotateX( cubeModelMat, -90.0 )
cubeModelMat = Scale( cubeModelMat, np.repeat( 1.6, 3 ) )
cubeModelMat = Translate( cubeModelMat, np.array( [dist * sin120, dist * -0.5, 0.0] ) )
cubeModelMat = RotateY( cubeModelMat, CalcAng( currentTime, 20.0 ) )
cubeModelMat = RotateX( cubeModelMat, CalcAng( currentTime, 13.0 ) )
# set up attributes and shader program
glEnable( GL_DEPTH_TEST )
glClear( GL_COLOR_BUFFER_BIT | GL_DEPTH_BUFFER_BIT )
glUseProgram( shaderProgram )
glUniformMatrix4fv( projectionMatLocation, 1, GL_FALSE, prjMat )
lightSourceBuffer.BindToTarget()
# draw tetrahedron
tetMaterialBuffer.BindToTarget()
modelViewMat = Multiply(viewMat, tetModelMat)
glUniformMatrix4fv( modelViewMatLocation, 1, GL_FALSE, modelViewMat )
glUniformMatrix3fv( normalMatLocation, 1, GL_FALSE, ToMat33(modelViewMat) )
glBindVertexArray( tetVAObj )
glDrawArrays( GL_TRIANGLES, 0, len(tetPosData) )
# draw icosahedron
icoMaterialBuffer.BindToTarget()
modelViewMat = Multiply(viewMat, icoModelMat)
glUniformMatrix4fv( modelViewMatLocation, 1, GL_FALSE, modelViewMat )
glUniformMatrix3fv( normalMatLocation, 1, GL_FALSE, ToMat33(modelViewMat) )
glBindVertexArray( icoVAObj )
glDrawArrays( GL_TRIANGLES, 0, len(icoPosData) )
# draw cube
cubeMaterialBuffer.BindToTarget()
modelViewMat = Multiply(viewMat, cubeModelMat)
glUniformMatrix4fv( modelViewMatLocation, 1, GL_FALSE, modelViewMat )
glUniformMatrix3fv( normalMatLocation, 1, GL_FALSE, ToMat33(modelViewMat) )
glBindVertexArray( cubeVAObj )
glDrawElements(GL_TRIANGLES, len(cubeIndices), GL_UNSIGNED_INT, cubeIndices)
glutSwapBuffers()
def Fract(val): return val - math.trunc(val)
def CalcAng(currentTime, intervall): return Fract( (currentTime - startTime) / intervall ) * 360.0
def CalcMove(currentTime, intervall, range):
pos = Fract( (currentTime - startTime) / intervall ) * 2.0
pos = pos if pos < 1.0 else (2.0-pos)
return range[0] + (range[1] - range[0]) * pos
# read shader program and compile shader
def CompileShader( sourceFileName, shaderStage ):
with open( sourceFileName, 'r' ) as sourceFile:
sourceCode = sourceFile.read()
nameMap = { GL_VERTEX_SHADER: 'vertex', GL_FRAGMENT_SHADER: 'fragment' }
print( '\n%s shader code:' % nameMap.get(shaderStage, '') )
print( sourceCode )
shaderObj = glCreateShader( shaderStage )
glShaderSource( shaderObj, sourceCode )
glCompileShader( shaderObj )
result = glGetShaderiv( shaderObj, GL_COMPILE_STATUS )
if not (result):
print( glGetShaderInfoLog( shaderObj ) )
sys.exit()
return shaderObj
# linke shader objects to shader program
def LinkProgram( shaderObjs ):
shaderProgram = glCreateProgram()
for shObj in shaderObjs:
glAttachShader( shaderProgram, shObj )
glLinkProgram( shaderProgram )
result = glGetProgramiv( shaderProgram, GL_LINK_STATUS )
if not (result):
print( 'link error:' )
print( glGetProgramInfoLog( shaderProgram ) )
sys.exit()
return shaderProgram
# create vertex array object
def CreateVAO( dataArrays ):
noOfBuffers = len(dataArrays)
buffers = glGenBuffers(noOfBuffers)
newVAObj = glGenVertexArrays( 1 )
glBindVertexArray( newVAObj )
for inx in range(0, noOfBuffers):
vertexSize, dataArr = dataArrays[inx]
arr = np.array( dataArr, dtype='float32' )
glBindBuffer( GL_ARRAY_BUFFER, buffers[inx] )
glBufferData( GL_ARRAY_BUFFER, arr, GL_STATIC_DRAW )
glEnableVertexAttribArray( inx )
glVertexAttribPointer( inx, vertexSize, GL_FLOAT, GL_FALSE, 0, None )
return newVAObj
# representation of a uniform block
class UniformBlock:
def __init__(self, shaderProg, name):
self.shaderProg = shaderProg
self.name = name
def Link(self, bindingPoint):
self.bindingPoint = bindingPoint
self.noOfUniforms = glGetProgramiv(self.shaderProg, GL_ACTIVE_UNIFORMS)
self.maxUniformNameLen = glGetProgramiv(self.shaderProg, GL_ACTIVE_UNIFORM_MAX_LENGTH)
self.index = glGetUniformBlockIndex(self.shaderProg, self.name)
intData = np.zeros(1, dtype=int)
glGetActiveUniformBlockiv(self.shaderProg, self.index, GL_UNIFORM_BLOCK_ACTIVE_UNIFORMS, intData)
self.count = intData[0]
self.indices = np.zeros(self.count, dtype=int)
glGetActiveUniformBlockiv(self.shaderProg, self.index, GL_UNIFORM_BLOCK_ACTIVE_UNIFORM_INDICES, self.indices)
self.offsets = np.zeros(self.count, dtype=int)
glGetActiveUniformsiv(self.shaderProg, self.count, self.indices, GL_UNIFORM_OFFSET, self.offsets)
self.size = 0
strLengthData = np.zeros(1, dtype=int)
arraysizeData = np.zeros(1, dtype=int)
typeData = np.zeros(1, dtype='uint32')
nameData = np.chararray(self.maxUniformNameLen+1)
self.namemap = {}
self.dataSize = 0
for inx in range(0, len(self.indices)):
glGetActiveUniform( self.shaderProg, self.indices[inx], self.maxUniformNameLen, strLengthData, arraysizeData, typeData, nameData.data )
name = nameData.tostring()[:strLengthData[0]]
self.namemap[name] = inx
self.dataSize = max(self.dataSize, self.offsets[inx] + arraysizeData * 16)
glUniformBlockBinding(self.shaderProg, self.index, self.bindingPoint)
print('\nuniform block %s size:%4d' % (self.name, self.dataSize))
for uName in self.namemap:
print( ' %-40s index:%2d offset:%4d' % (uName, self.indices[self.namemap[uName]], self.offsets [self.namemap[uName]]) )
# representation of a uniform block buffer
class UniformBlockBuffer:
def __init__(self, ub):
self.namemap = ub.namemap
self.offsets = ub.offsets
self.bindingPoint = ub.bindingPoint
self.object = glGenBuffers(1)
self.dataSize = ub.dataSize
glBindBuffer(GL_UNIFORM_BUFFER, self.object)
dataArray = np.zeros(self.dataSize//4, dtype='float32')
glBufferData(GL_UNIFORM_BUFFER, self.dataSize, dataArray, GL_DYNAMIC_DRAW)
def BindToTarget(self):
glBindBuffer(GL_UNIFORM_BUFFER, self.object)
glBindBufferBase(GL_UNIFORM_BUFFER, self.bindingPoint, self.object)
def BindDataFloat(self, name, dataArr):
glBindBuffer(GL_UNIFORM_BUFFER, self.object)
dataArray = np.array(dataArr, dtype='float32')
glBufferSubData(GL_UNIFORM_BUFFER, self.offsets[self.namemap[name]], len(dataArr)*4, dataArray)
def Translate(matA, trans):
matB = np.copy(matA)
for i in range(0, 4): matB[3,i] = matA[0,i] * trans[0] + matA[1,i] * trans[1] + matA[2,i] * trans[2] + matA[3,i]
return matB
def Scale(matA, s):
matB = np.copy(matA)
for i0 in range(0, 3):
for i1 in range(0, 4): matB[i0,i1] = matA[i0,i1] * s[i0]
return matB
def RotateHlp(matA, angDeg, a0, a1):
matB = np.copy(matA)
ang = math.radians(angDeg)
sinAng, cosAng = math.sin(ang), math.cos(ang)
for i in range(0, 4):
matB[a0,i] = matA[a0,i] * cosAng + matA[a1,i] * sinAng
matB[a1,i] = matA[a0,i] * -sinAng + matA[a1,i] * cosAng
return matB
def RotateX(matA, angDeg): return RotateHlp(matA, angDeg, 1, 2)
def RotateY(matA, angDeg): return RotateHlp(matA, angDeg, 2, 0)
def RotateZ(matA, angDeg): return RotateHlp(matA, angDeg, 0, 1)
def RotateView(matA, angDeg): return RotateZ(RotateY(RotateX(matA, angDeg[0]), angDeg[1]), angDeg[2])
def Multiply(matA, matB):
matC = np.copy(matA)
for i0 in range(0, 4):
for i1 in range(0, 4):
matC[i0,i1] = matB[i0,0] * matA[0,i1] + matB[i0,1] * matA[1,i1] + matB[i0,2] * matA[2,i1] + matB[i0,3] * matA [3,i1]
return matC
def ToMat33(mat44):
mat33 = np.matrix(np.identity(3), copy=False, dtype='float32')
for i0 in range(0, 3):
for i1 in range(0, 3): mat33[i0, i1] = mat44[i0, i1]
return mat33
def TransformVec4(vecA,mat44):
vecB = np.zeros(4, dtype='float32')
for i0 in range(0, 4):
vecB[i0] = vecA[0] * mat44[0,i0] + vecA[1] * mat44[1,i0] + vecA[2] * mat44[2,i0] + vecA[3] * mat44[3,i0]
return vecB
def Perspective(fov, aspectRatio, near, far):
fn, f_n = far + near, far - near
r, t = aspectRatio, 1.0 / math.tan( math.radians(fov) / 2.0 )
return np.matrix( [ [t/r,0,0,0], [0,t,0,0], [0,0,-fn/f_n,-2.0*far*near/f_n], [0,0,-1,0] ] )
def AddToBuffer( buffer, data, count=1 ):
for inx_c in range(0, count):
for inx_s in range(0, len(data)): buffer.append( data[inx_s] )
# initialize glut
glutInit()
# create window
wndW, wndH = 800, 600
glutInitDisplayMode(GLUT_RGBA | GLUT_DOUBLE | GLUT_ALPHA | GLUT_DEPTH)
glutInitWindowPosition(0, 0)
glutInitWindowSize(wndW, wndH)
wndID = glutCreateWindow(b'OGL window')
glutDisplayFunc(OnDraw)
glutIdleFunc(OnDraw)
# define tetrahedron vertex array opject
tetPts = [ (0.0, 0.0, 1.0), (0.0, -sin120, -0.5), (sin120 * sin120, 0.5 * sin120, -0.5), (-sin120 * sin120, 0.5 * sin120, -0.5) ]
tetCol = [ [1.0, 0.0, 0.0], [1.0, 1.0, 0.0], [0.0, 0.0, 1.0], [0.0, 1.0, 0.0], ]
tetInxdices = [ 0, 1, 2, 0, 2, 3, 0, 3, 1, 1, 3, 2 ]
tetPosData = []
for inx in tetInxdices: AddToBuffer( tetPosData, tetPts[inx] )
tetNVData = []
for inx_nv in range(0, len(tetInxdices) // 3):
nv = [0.0, 0.0, 0.0]
for inx_p in range(0, 3):
for inx_s in range(0, 3): nv[inx_s] += tetPts[ tetInxdices[inx_nv*3 + inx_p] ][inx_s]
AddToBuffer( tetNVData, nv, 3 )
tetColData = []
for inx_col in range(0, len(tetInxdices) // 3): AddToBuffer( tetColData, tetCol[inx_col % len(tetCol)], 3 )
tetVAObj = CreateVAO( [ (3, tetPosData), (3, tetNVData), (3, tetColData) ] )
# define icosahedron vertex array opject
icoPts = [
( 0.000, 0.000, 1.000), ( 0.894, 0.000, 0.447), ( 0.276, 0.851, 0.447), (-0.724, 0.526, 0.447),
(-0.724, -0.526, 0.447), ( 0.276, -0.851, 0.447), ( 0.724, 0.526, -0.447), (-0.276, 0.851, -0.447),
(-0.894, 0.000, -0.447), (-0.276, -0.851, -0.447), ( 0.724, -0.526, -0.447), ( 0.000, 0.000, -1.000) ]
icoCol = [ [1.0, 0.0, 0.0], [0.0, 0.0, 1.0], [1.0, 1.0, 0.0], [0.0, 1.0, 0.0], [1.0, 0.5, 0.0], [1.0, 0.0, 1.0] ]
icoIndices = [
2, 0, 1, 3, 0, 2, 4, 0, 3, 5, 0, 4, 1, 0, 5, 11, 7, 6, 11, 8, 7, 11, 9, 8, 11, 10, 9, 11, 6, 10,
1, 6, 2, 2, 7, 3, 3, 8, 4, 4, 9, 5, 5, 10, 1, 2, 6, 7, 3, 7, 8, 4, 8, 9, 5, 9, 10, 1, 10, 6 ]
icoPosData = []
for inx in icoIndices: AddToBuffer( icoPosData, icoPts[inx] )
icoNVData = []
for inx in icoIndices: AddToBuffer( icoNVData, icoPts[inx] )
#for inx_nv in range(0, len(icoIndices) // 3):
# nv = [0.0, 0.0, 0.0]
# for inx_p in range(0, 3):
# for inx_s in range(0, 3): nv[inx_s] += icoPts[ icoIndices[inx_nv*3 + inx_p] ][inx_s]
# AddToBuffer( icoNVData, nv, 3 )
icoColData = []
for inx_col in range(0, len(icoIndices) // 3): AddToBuffer( icoColData, icoCol[inx_col % len(icoCol)], 3 )
icoVAObj = CreateVAO( [ (3, icoPosData), (3, icoNVData), (3, icoColData) ] )
# define cube vertex array opject
cubePts = [
(-1.0, -1.0, 1.0), ( 1.0, -1.0, 1.0), ( 1.0, 1.0, 1.0), (-1.0, 1.0, 1.0),
(-1.0, -1.0, -1.0), ( 1.0, -1.0, -1.0), ( 1.0, 1.0, -1.0), (-1.0, 1.0, -1.0) ]
cubeCol = [ [1.0, 0.0, 0.0], [1.0, 0.5, 0.0], [1.0, 0.0, 1.0], [1.0, 1.0, 0.0], [0.0, 1.0, 0.0], [0.0, 0.0, 1.0] ]
cubeHlpInx = [ 0, 1, 2, 3, 1, 5, 6, 2, 5, 4, 7, 6, 4, 0, 3, 7, 3, 2, 6, 7, 1, 0, 4, 5 ]
cubePosData = []
for inx in cubeHlpInx: AddToBuffer( cubePosData, cubePts[inx] )
cubeNVData = []
for inx_nv in range(0, len(cubeHlpInx) // 4):
nv = [0.0, 0.0, 0.0]
for inx_p in range(0, 4):
for inx_s in range(0, 3): nv[inx_s] += cubePts[ cubeHlpInx[inx_nv*4 + inx_p] ][inx_s]
AddToBuffer( cubeNVData, nv, 4 )
cubeColData = []
for inx_col in range(0, 6):
AddToBuffer( cubeColData, cubeCol[inx_col % len(cubeCol)], 4 )
cubeIndices = []
for inx in range(0, 6):
for inx_s in [0, 1, 2, 0, 2, 3]: cubeIndices.append( inx * 4 + inx_s )
cubeVAObj = CreateVAO( [ (3, cubePosData), (3, cubeNVData), (3, cubeColData) ] )
# load, compile and link shader
shaderProgram = LinkProgram( [
CompileShader( 'ibub.vert', GL_VERTEX_SHADER ),
CompileShader( 'ibub.frag', GL_FRAGMENT_SHADER )
] )
# get unifor locations
projectionMatLocation = glGetUniformLocation(shaderProgram, "u_projectionMat44")
modelViewMatLocation = glGetUniformLocation(shaderProgram, "u_modelViewMat44")
normalMatLocation = glGetUniformLocation(shaderProgram, "u_normalMat33")
# linke uniform blocks
ubMaterial = UniformBlock(shaderProgram, "UB_material")
ubLightSource = UniformBlock(shaderProgram, "UB_lightSource")
ubMaterial.Link(1)
ubLightSource.Link(2)
# create uniform block buffers
lightSourceBuffer = UniformBlockBuffer(ubLightSource)
lightSourceBuffer.BindDataFloat(b'u_lightSource.ambient', [0.1, 0.1, 0.1, 1.0])
lightSourceBuffer.BindDataFloat(b'u_lightSource.diffuse', [0.4, 0.4, 0.4, 1.0])
lightSourceBuffer.BindDataFloat(b'u_lightSource.specular', [1.0, 1.0, 1.0, 1.0])
tetMaterialBuffer = UniformBlockBuffer(ubMaterial)
tetMaterialBuffer.BindDataFloat(b'u_roughness', [0.3])
tetMaterialBuffer.BindDataFloat(b'u_fresnel0', [0.5])
tetMaterialBuffer.BindDataFloat(b'u_specularTint',[1.0, 1.0, 1.0, 0.7])
icoMaterialBuffer = UniformBlockBuffer(ubMaterial)
icoMaterialBuffer.BindDataFloat(b'u_roughness', [0.1])
icoMaterialBuffer.BindDataFloat(b'u_fresnel0', [0.2])
icoMaterialBuffer.BindDataFloat(b'u_specularTint',[1.0, 1.0, 1.0, 0.7])
cubeMaterialBuffer = UniformBlockBuffer(ubMaterial)
cubeMaterialBuffer.BindDataFloat(b'u_roughness', [0.5])
cubeMaterialBuffer.BindDataFloat(b'u_fresnel0', [0.3])
cubeMaterialBuffer.BindDataFloat(b'u_specularTint',[1.0, 1.0, 1.0, 0.7])
# start main loop
startTime = time()
glutMainLoop()
OGL 4.0 GLSLでジオメトリシェーダを使用してジオメトリを作成する
ジオメトリシェーダの使用を示すシンプルなOGL 4.0 GLSLシェーダプログラムです。プログラムはphytonスクリプトで実行されます。スクリプトを実行するには、PyOpenGLとNumPyをインストールする必要があります。
この例では、ジオメトリシェーダでジオメトリ全体（シリンダ）が生成されます。
頂点シェーダ
geo.vert
#version 400
layout (location = 0) in vec3 inPos;
layout (location = 1) in vec3 inNormal;
layout (location = 2) in vec3 inTangent;
out TVertexData
{
mat3 orientationMat;
} outData;
void main()
{
vec3 normal = normalize( inNormal );
vec3 tangent = normalize( inTangent );
vec3 binormal = cross( tangent, normal );
outData.orientationMat = mat3( normal, cross( binormal, normal ), binormal );
gl_Position = vec4( inPos, 1.0 );
}
ジオメトリシェーダ
geo.geo
#version 400
layout( invocations = 3 ) in;
layout( points ) in;
layout( triangle_strip, max_vertices = 160 ) out;
in TVertexData
{
mat3 orientationMat;
} inData[];
out TGeometryData
{
vec3 pos;
vec3 nv;
vec3 col;
} outData;
uniform mat4 u_projectionMat44;
uniform mat4 u_viewMat44;
uniform mat4 u_modelMat44;
void NewVertex( in vec3 pt, in mat4 transMat )
{
vec4 viewPos = transMat * vec4( pt, 1.0 );
outData.pos = viewPos.xyz / viewPos.w;
gl_Position = u_projectionMat44 * viewPos;
EmitVertex();
}
const int circumferenceTile = 36;
void main()
{
vec4 origin = gl_in[0].gl_Position;
origin /= origin.w;
mat4 orintationMat = mat4( vec4( inData[0].orientationMat[0], 0.0 ),
vec4( inData[0].orientationMat[1], 0.0 ),
vec4( inData[0].orientationMat[2], 0.0 ),
origin );
mat4 modelViewMat = u_viewMat44 * u_modelMat44 * orintationMat;
mat3 normalMat = mat3( modelViewMat );
outData.col = vec3( 0.5, 0.7, 0.6 );
if ( gl_InvocationID == 0 ) // top of the cylinder
{
outData.nv = normalMat * vec3(0.0, 0.0, 1.0);
vec2 prevPt = vec2( 0.0, 1.0 );
for ( int inx = 1; inx <= circumferenceTile; inx += 2 )
{
float ang1 = 2.0 * 3.14159 * float(inx) / float(circumferenceTile);
float ang2 = 2.0 * 3.14159 * float(inx+1) / float(circumferenceTile);
vec2 actPt1 = vec2( sin(ang1), cos(ang1) );
vec2 actPt2 = vec2( sin(ang2), cos(ang2) );
NewVertex( vec3(prevPt.xy, 1.0), modelViewMat );
NewVertex( vec3(actPt1.xy, 1.0), modelViewMat );
NewVertex( vec3(0.0, 0.0, 1.0), modelViewMat );
NewVertex( vec3(actPt2.xy, 1.0), modelViewMat );
EndPrimitive();
prevPt = actPt2;
}
}
if ( gl_InvocationID == 1 ) // bottom of the cylinder
{
outData.nv = normalMat * vec3(0.0, 0.0, -1.0);
vec2 prevPt = vec2( 0.0, 1.0 );
for ( int inx = circumferenceTile-1; inx >= 0; inx -= 2 )
{
float ang1 = 2.0 * 3.14159 * float(inx) / float(circumferenceTile);
float ang2 = 2.0 * 3.14159 * float(inx-1) / float(circumferenceTile);
vec2 actPt1 = vec2( sin(ang1), cos(ang1) );
vec2 actPt2 = vec2( sin(ang2), cos(ang2) );
NewVertex( vec3(prevPt.xy, -1.0), modelViewMat );
NewVertex( vec3(actPt1.xy, -1.0), modelViewMat );
NewVertex( vec3(0.0, 0.0, -1.0), modelViewMat );
NewVertex( vec3(actPt2.xy, -1.0), modelViewMat );
EndPrimitive();
prevPt = actPt2;
}
}
if ( gl_InvocationID == 2 ) // hull of the cylinder
{
vec2 prevPt = vec2( 0.0, 1.0 );
for ( int inx = 1; inx <= circumferenceTile; ++ inx )
{
float ang = 2.0 * 3.14159 * float(inx) / float(circumferenceTile);
vec2 actPt = vec2( sin(ang), cos(ang) );
outData.nv = normalMat * vec3(prevPt, 0.0);
NewVertex( vec3(prevPt.xy, -1.0), modelViewMat );
outData.nv = normalMat * vec3(actPt, 0.0);
NewVertex( vec3(actPt.xy, -1.0), modelViewMat );
outData.nv = normalMat * vec3(prevPt, 0.0);
NewVertex( vec3(prevPt.xy, 1.0), modelViewMat );
outData.nv = normalMat * vec3(actPt, 0.0);
NewVertex( vec3(actPt.xy, 1.0), modelViewMat );
prevPt = actPt;
}
EndPrimitive();
}
}
フラグメントシェーダ
geo.frag
#version 400
in TGeometryData
{
vec3 pos;
vec3 nv;
vec3 col;
} inData;
out vec4 fragColor;
uniform UB_material
{
float u_roughness;
float u_fresnel0;
vec4 u_specularTint;
};
struct TLightSource
{
vec4 ambient;
vec4 diffuse;
vec4 specular;
vec4 dir;
};
uniform UB_lightSource
{
TLightSource u_lightSource;
};
float Fresnel_Schlick( float theta )
{
float m = clamp( 1.0 - theta, 0.0, 1.0 );
float m2 = m * m;
return m2 * m2 * m; // pow( m, 5.0 )
}
vec3 LightModel( vec3 esPt, vec3 esPtNV, vec3 col, vec4 specularTint, float roughness, float fresnel0 )
{
vec3 esVLight = normalize( -u_lightSource.dir.xyz );
vec3 esVEye = normalize( -esPt );
vec3 halfVector = normalize( esVEye + esVLight );
float HdotL = dot( halfVector, esVLight );
float NdotL = dot( esPtNV, esVLight );
float NdotV = dot( esPtNV, esVEye );
float NdotH = dot( esPtNV, halfVector );
float NdotH2 = NdotH * NdotH;
float NdotL_clamped = max( NdotL, 0.0 );
float NdotV_clamped = max( NdotV, 0.0 );
float m2 = roughness * roughness;
// Lambertian diffuse
float k_diffuse = NdotL_clamped;
// Schlick approximation
float fresnel = fresnel0 + ( 1.0 - fresnel0 ) * Fresnel_Schlick( HdotL );
// Beckmann distribution
float distribution = max( 0.0, exp( ( NdotH2 - 1.0 ) / ( m2 * NdotH2 ) ) / ( 3.14159265 * m2 * NdotH2 * NdotH2 ) );
// Torrance-Sparrow geometric term
float geometric_att = min( 1.0, min( 2.0 * NdotH * NdotV_clamped / HdotL, 2.0 * NdotH * NdotL_clamped / HdotL ) );
// Microfacet bidirectional reflectance distribution function
float k_specular = fresnel * distribution * geometric_att / ( 4.0 * NdotL_clamped * NdotV_clamped );
vec3 lightColor = col.rgb * u_lightSource.ambient.rgb +
max( 0.0, k_diffuse ) * col.rgb * u_lightSource.diffuse.rgb +
max( 0.0, k_specular ) * mix( col.rgb, specularTint.rgb, specularTint.a ) * u_lightSource.specular.rgb;
return lightColor;
}
void main()
{
vec3 lightCol = LightModel( inData.pos, inData.nv, inData.col, u_specularTint, u_roughness, u_fresnel0 );
fragColor = vec4( clamp( lightCol, 0.0, 1.0 ), 1.0 );
}
Phytonスクリプト
from OpenGL.GL import *
from OpenGL.GLUT import *
from OpenGL.GLU import *
import numpy as np
from time import time
import math
import sys
sin120 = 0.8660254
rotateCamera = False
# draw event
def OnDraw():
dist = 3.0
currentTime = time()
comeraRotAng = CalcAng( currentTime, 10.0 )
# set up projection matrix
prjMat = Perspective(90.0, wndW/wndH, 0.5, 100.0)
# set up view matrix
viewMat = np.matrix(np.identity(4), copy=False, dtype='float32')
viewMat = Translate( viewMat, np.array( [0.0, 0.0, -12.0] ) )
viewMat = RotateView( viewMat, [30.0, comeraRotAng if rotateCamera else 0.0, 0.0] )
# set up light source
lightSourceBuffer.BindDataFloat(b'u_lightSource.dir', TransformVec4([-0.1, 1.0, -5.0, 0.0], viewMat) )
# set up the model matrix
modelMat = np.matrix(np.identity(4), copy=False, dtype='float32')
if not rotateCamera: modelMat = RotateY( modelMat, comeraRotAng )
modelMat = Scale( modelMat, np.repeat( 4, 3 ) )
#modelMat = Translate( modelMat, np.array( [0.0, 0.0, 1.0] ) )
#modelMat = RotateY( modelMat, CalcAng( currentTime, 20.0 ) )
modelMat = RotateX( modelMat, CalcAng( currentTime, 9.0 ) )
# set up attributes and shader program
glEnable( GL_DEPTH_TEST )
glClear( GL_COLOR_BUFFER_BIT | GL_DEPTH_BUFFER_BIT )
glUseProgram( shaderProgram )
glUniformMatrix4fv( projectionMatLocation, 1, GL_FALSE, prjMat )
glUniformMatrix4fv( viewMatLocation, 1, GL_FALSE, viewMat )
lightSourceBuffer.BindToTarget()
# draw point
materialBuffer.BindToTarget()
glUniformMatrix4fv( modelMatLocation, 1, GL_FALSE, modelMat )
glBindVertexArray( pointVAObj )
glDrawArrays( GL_POINTS, 0, 1 )
glutSwapBuffers()
def Fract(val): return val - math.trunc(val)
def CalcAng(currentTime, intervall): return Fract( (currentTime - startTime) / intervall ) * 360.0
def CalcMove(currentTime, intervall, range):
pos = Fract( (currentTime - startTime) / intervall ) * 2.0
pos = pos if pos < 1.0 else (2.0-pos)
return range[0] + (range[1] - range[0]) * pos
# read shader program and compile shader
def CompileShader( sourceFileName, shaderStage ):
with open( sourceFileName, 'r' ) as sourceFile:
sourceCode = sourceFile.read()
nameMap = { GL_VERTEX_SHADER: 'vertex', GL_GEOMETRY_SHADER: 'geometry', GL_FRAGMENT_SHADER: 'fragment' }
print( '\n%s shader code:' % nameMap.get(shaderStage, '') )
print( sourceCode )
shaderObj = glCreateShader( shaderStage )
glShaderSource( shaderObj, sourceCode )
glCompileShader( shaderObj )
result = glGetShaderiv( shaderObj, GL_COMPILE_STATUS )
if not (result):
print( glGetShaderInfoLog( shaderObj ) )
sys.exit()
return shaderObj
# linke shader objects to shader program
def LinkProgram( shaderObjs ):
shaderProgram = glCreateProgram()
for shObj in shaderObjs:
glAttachShader( shaderProgram, shObj )
glLinkProgram( shaderProgram )
result = glGetProgramiv( shaderProgram, GL_LINK_STATUS )
if not (result):
print( 'link error:' )
print( glGetProgramInfoLog( shaderProgram ) )
sys.exit()
return shaderProgram
# create vertex array object
def CreateVAO( dataArrays ):
noOfBuffers = len(dataArrays)
buffers = glGenBuffers(noOfBuffers)
newVAObj = glGenVertexArrays( 1 )
glBindVertexArray( newVAObj )
for inx in range(0, noOfBuffers):
vertexSize, dataArr = dataArrays[inx]
arr = np.array( dataArr, dtype='float32' )
glBindBuffer( GL_ARRAY_BUFFER, buffers[inx] )
glBufferData( GL_ARRAY_BUFFER, arr, GL_STATIC_DRAW )
glEnableVertexAttribArray( inx )
glVertexAttribPointer( inx, vertexSize, GL_FLOAT, GL_FALSE, 0, None )
return newVAObj
# representation of a uniform block
class UniformBlock:
def __init__(self, shaderProg, name):
self.shaderProg = shaderProg
self.name = name
def Link(self, bindingPoint):
self.bindingPoint = bindingPoint
self.noOfUniforms = glGetProgramiv(self.shaderProg, GL_ACTIVE_UNIFORMS)
self.maxUniformNameLen = glGetProgramiv(self.shaderProg, GL_ACTIVE_UNIFORM_MAX_LENGTH)
self.index = glGetUniformBlockIndex(self.shaderProg, self.name)
intData = np.zeros(1, dtype=int)
glGetActiveUniformBlockiv(self.shaderProg, self.index, GL_UNIFORM_BLOCK_ACTIVE_UNIFORMS, intData)
self.count = intData[0]
self.indices = np.zeros(self.count, dtype=int)
glGetActiveUniformBlockiv(self.shaderProg, self.index, GL_UNIFORM_BLOCK_ACTIVE_UNIFORM_INDICES, self.indices)
self.offsets = np.zeros(self.count, dtype=int)
glGetActiveUniformsiv(self.shaderProg, self.count, self.indices, GL_UNIFORM_OFFSET, self.offsets)
strLengthData = np.zeros(1, dtype=int)
arraysizeData = np.zeros(1, dtype=int)
typeData = np.zeros(1, dtype='uint32')
nameData = np.chararray(self.maxUniformNameLen+1)
self.namemap = {}
self.dataSize = 0
for inx in range(0, len(self.indices)):
glGetActiveUniform( self.shaderProg, self.indices[inx], self.maxUniformNameLen, strLengthData, arraysizeData, typeData, nameData.data )
name = nameData.tostring()[:strLengthData[0]]
self.namemap[name] = inx
self.dataSize = max(self.dataSize, self.offsets[inx] + arraysizeData * 16)
glUniformBlockBinding(self.shaderProg, self.index, self.bindingPoint)
print('\nuniform block %s size:%4d' % (self.name, self.dataSize))
for uName in self.namemap:
print( ' %-40s index:%2d offset:%4d' % (uName, self.indices[self.namemap[uName]], self.offsets [self.namemap[uName]]) )
# representation of a uniform block buffer
class UniformBlockBuffer:
def __init__(self, ub):
self.namemap = ub.namemap
self.offsets = ub.offsets
self.bindingPoint = ub.bindingPoint
self.object = glGenBuffers(1)
self.dataSize = ub.dataSize
glBindBuffer(GL_UNIFORM_BUFFER, self.object)
dataArray = np.zeros(self.dataSize//4, dtype='float32')
glBufferData(GL_UNIFORM_BUFFER, self.dataSize, dataArray, GL_DYNAMIC_DRAW)
def BindToTarget(self):
glBindBuffer(GL_UNIFORM_BUFFER, self.object)
glBindBufferBase(GL_UNIFORM_BUFFER, self.bindingPoint, self.object)
def BindDataFloat(self, name, dataArr):
glBindBuffer(GL_UNIFORM_BUFFER, self.object)
dataArray = np.array(dataArr, dtype='float32')
glBufferSubData(GL_UNIFORM_BUFFER, self.offsets[self.namemap[name]], len(dataArr)*4, dataArray)
def Translate(matA, trans):
matB = np.copy(matA)
for i in range(0, 4): matB[3,i] = matA[0,i] * trans[0] + matA[1,i] * trans[1] + matA[2,i] * trans[2] + matA[3,i]
return matB
def Scale(matA, s):
matB = np.copy(matA)
for i0 in range(0, 3):
for i1 in range(0, 4): matB[i0,i1] = matA[i0,i1] * s[i0]
return matB
def RotateHlp(matA, angDeg, a0, a1):
matB = np.copy(matA)
ang = math.radians(angDeg)
sinAng, cosAng = math.sin(ang), math.cos(ang)
for i in range(0, 4):
matB[a0,i] = matA[a0,i] * cosAng + matA[a1,i] * sinAng
matB[a1,i] = matA[a0,i] * -sinAng + matA[a1,i] * cosAng
return matB
def RotateX(matA, angDeg): return RotateHlp(matA, angDeg, 1, 2)
def RotateY(matA, angDeg): return RotateHlp(matA, angDeg, 2, 0)
def RotateZ(matA, angDeg): return RotateHlp(matA, angDeg, 0, 1)
def RotateView(matA, angDeg): return RotateZ(RotateY(RotateX(matA, angDeg[0]), angDeg[1]), angDeg[2])
def Multiply(matA, matB):
matC = np.copy(matA)
for i0 in range(0, 4):
for i1 in range(0, 4):
matC[i0,i1] = matB[i0,0] * matA[0,i1] + matB[i0,1] * matA[1,i1] + matB[i0,2] * matA[2,i1] + matB[i0,3] * matA [3,i1]
return matC
def ToMat33(mat44):
mat33 = np.matrix(np.identity(3), copy=False, dtype='float32')
for i0 in range(0, 3):
for i1 in range(0, 3): mat33[i0, i1] = mat44[i0, i1]
return mat33
def TransformVec4(vecA,mat44):
vecB = np.zeros(4, dtype='float32')
for i0 in range(0, 4):
vecB[i0] = vecA[0] * mat44[0,i0] + vecA[1] * mat44[1,i0] + vecA[2] * mat44[2,i0] + vecA[3] * mat44[3,i0]
return vecB
def Perspective(fov, aspectRatio, near, far):
fn, f_n = far + near, far - near
r, t = aspectRatio, 1.0 / math.tan( math.radians(fov) / 2.0 )
return np.matrix( [ [t/r,0,0,0], [0,t,0,0], [0,0,-fn/f_n,-2.0*far*near/f_n], [0,0,-1,0] ] )
def AddToBuffer( buffer, data, count=1 ):
for inx_c in range(0, count):
for inx_s in range(0, len(data)): buffer.append( data[inx_s] )
# initialize glut
glutInit()
# create window
wndW, wndH = 800, 600
glutInitDisplayMode(GLUT_RGBA | GLUT_DOUBLE | GLUT_ALPHA | GLUT_DEPTH)
glutInitWindowPosition(0, 0)
glutInitWindowSize(wndW, wndH)
wndID = glutCreateWindow(b'OGL window')
glutDisplayFunc(OnDraw)
glutIdleFunc(OnDraw)
# define location vertex array opject
pointVAObj = CreateVAO( [ (3, [0.0, 0.0, 0.0] ), (3, [0.0, 0.0, -1.0]), (3, [1.0, 0.0, 0.0]) ] )
# load, compile and link shader
shaderProgram = LinkProgram( [
CompileShader( 'geo.vert', GL_VERTEX_SHADER ),
CompileShader( 'geo.geo', GL_GEOMETRY_SHADER ),
CompileShader( 'geo.frag', GL_FRAGMENT_SHADER )
] )
# get unifor locations
projectionMatLocation = glGetUniformLocation(shaderProgram, "u_projectionMat44")
viewMatLocation = glGetUniformLocation(shaderProgram, "u_viewMat44")
modelMatLocation = glGetUniformLocation(shaderProgram, "u_modelMat44")
# linke uniform blocks
ubMaterial = UniformBlock(shaderProgram, "UB_material")
ubLightSource = UniformBlock(shaderProgram, "UB_lightSource")
ubMaterial.Link(1)
ubLightSource.Link(2)
# create uniform block buffers
lightSourceBuffer = UniformBlockBuffer(ubLightSource)
lightSourceBuffer.BindDataFloat(b'u_lightSource.ambient', [0.2, 0.2, 0.2, 1.0])
lightSourceBuffer.BindDataFloat(b'u_lightSource.diffuse', [0.2, 0.2, 0.2, 1.0])
lightSourceBuffer.BindDataFloat(b'u_lightSource.specular', [1.0, 1.0, 1.0, 1.0])
materialBuffer = UniformBlockBuffer(ubMaterial)
materialBuffer.BindDataFloat(b'u_roughness', [0.5])
materialBuffer.BindDataFloat(b'u_fresnel0', [0.2])
materialBuffer.BindDataFloat(b'u_specularTint',[1.0, 0.5, 0.5, 0.8])
# start main loop
startTime = time()
glutMainLoop()
OGL 4.0 GLSLのサブルーチンを使用してジオメトリとサーフェス表現を切り替える
使用シェーダサブルーチンを示すシンプルなOGL 4.0 GLSLシェーダプログラム。プログラムはphytonスクリプトで実行されます。スクリプトを実行するには、PyOpenGLとNumPyをインストールする必要があります。
サブルーチンは、ジオメトリシェーダで生成された異なるジオメトリを切り替え、サーフェスの表現を変更します。
頂点シェーダ
subr.vert
#version 400
layout (location = 0) in vec3 inPos;
layout (location = 1) in vec3 inNormal;
layout (location = 2) in vec3 inTangent;
out TVertexData
{
mat3 orientationMat;
} outData;
void main()
{
vec3 normal = normalize( inNormal );
vec3 tangent = normalize( inTangent );
vec3 binormal = cross( tangent, normal );
outData.orientationMat = mat3( normal, cross( binormal, normal ), binormal );
gl_Position = vec4( inPos, 1.0 );
}
ジオメトリシェーダ
サブレッジョ
#version 400
layout( points ) in;
layout( triangle_strip, max_vertices = 512 ) out;
in TVertexData
{
mat3 orientationMat;
} inData[];
out TGeometryData
{
vec3 pos;
vec3 nv;
vec2 tex;
} outData;
uniform mat4 u_projectionMat44;
uniform mat4 u_viewMat44;
uniform mat4 u_modelMat44;
uniform mat4 u_textureMat44;
void SetTextureCoord( in vec2 tecCoord )
{
vec4 tex = u_textureMat44 * vec4( tecCoord, 0.0, 1.0 );
outData.tex = tex.xy;
}
void NewVertex( in vec3 pt, in mat4 transMat )
{
vec4 viewPos = transMat * vec4( pt, 1.0 );
outData.pos = viewPos.xyz / viewPos.w;
gl_Position = u_projectionMat44 * viewPos;
EmitVertex();
}
void NewVertexAndTex( in vec3 pt, in mat4 transMat )
{
SetTextureCoord( pt.xy * 0.5 + 0.5 );
NewVertex( pt, transMat );
}
void NewVertexNvTex( in vec3 pt, in mat4 transMat, in vec3 nv, in vec2 tex )
{
outData.nv = nv;
SetTextureCoord( tex );
vec4 viewPos = transMat * vec4( pt, 1.0 );
outData.pos = viewPos.xyz / viewPos.w;
gl_Position = u_projectionMat44 * viewPos;
EmitVertex();
}
subroutine void TShape( in mat4 );
subroutine uniform TShape su_shape;
void main()
{
vec4 origin = gl_in[0].gl_Position;
origin /= origin.w;
mat4 orintationMat = mat4( vec4( inData[0].orientationMat[0], 0.0 ),
vec4( inData[0].orientationMat[1], 0.0 ),
vec4( inData[0].orientationMat[2], 0.0 ),
origin );
mat4 modelMat = u_modelMat44 * orintationMat;
su_shape( modelMat );
}
subroutine(TShape) void DrawSphere( in mat4 modelMat )
{
const int circumferenceTile = 18;
const int layersTile = 11;
mat4 modelViewMat = u_viewMat44 * modelMat;
mat3 normalMat = mat3( modelViewMat );
float preStepLay = 0.0;
vec2 prePtLay = vec2( 0.0, -1.0 );
for ( int inxLay = 1; inxLay <= layersTile; ++ inxLay )
{
float stepLay = float(inxLay) / float(layersTile);
float angLay = 3.14159 * stepLay;
vec2 ptLay = vec2( sin(angLay), -cos(angLay) );
float preStepCir = 0.0;
vec2 prePtCir = vec2( 0.0, 1.0 );
for ( int inxCir = 0; inxCir <= circumferenceTile; ++ inxCir )
{
float stepCir = float(inxCir) / float(circumferenceTile);
float angCir = 2.0 * 3.14159 * stepCir;
vec2 ptCir = vec2( sin(angCir), cos(angCir) );
if ( inxLay == 1 )
{
if ( inxCir >= 0 )
{
vec3 pt1 = vec3( ptLay.x * prePtCir.x, ptLay.x * prePtCir.y, ptLay.y );
vec3 pt2 = vec3( 0.0, 0.0, -1.0 );
vec3 pt3 = vec3( ptLay.x * ptCir.x, ptLay.x * ptCir.y, ptLay.y );
NewVertexNvTex( pt1, modelViewMat, normalMat * pt1, vec2( preStepCir * 2.0, stepLay ) );
NewVertexNvTex( pt2, modelViewMat, normalMat * pt2, vec2( preStepCir + stepCir, preStepLay ) );
NewVertexNvTex( pt3, modelViewMat, normalMat * pt3, vec2( stepCir * 2.0, stepLay ) );
EndPrimitive();
}
}
else if ( inxLay == layersTile )
{
if ( inxCir > 0 )
{
vec3 pt1 = vec3( prePtLay.x * prePtCir.x, prePtLay.x * prePtCir.y, prePtLay.y );
vec3 pt2 = vec3( prePtLay.x * ptCir.x, prePtLay.x * ptCir.y, prePtLay.y );
vec3 pt3 = vec3( 0.0, 0.0, 1.0 );
NewVertexNvTex( pt1, modelViewMat, normalMat * pt1, vec2( preStepCir * 2.0, preStepLay ) );
NewVertexNvTex( pt2, modelViewMat, normalMat * pt2, vec2( stepCir * 2.0, preStepLay ) );
NewVertexNvTex( pt3, modelViewMat, normalMat * pt3, vec2( preStepCir + stepCir, stepLay ) );
EndPrimitive();
}
}
else
{
vec3 pt1 = vec3( prePtLay.x * ptCir.x, prePtLay.x * ptCir.y, prePtLay.y );
vec3 pt2 = vec3( ptLay.x * ptCir.x, ptLay.x * ptCir.y, ptLay.y );
NewVertexNvTex( pt1, modelViewMat, normalMat * pt1, vec2( stepCir * 2.0, preStepLay ) );
NewVertexNvTex( pt2, modelViewMat, normalMat * pt2, vec2( stepCir * 2.0, stepLay ) );
}
preStepCir = stepCir;
prePtCir = ptCir;
}
if ( inxLay > 1 && inxLay < layersTile )
EndPrimitive();
preStepLay = stepLay;
prePtLay = ptLay;
}
}
subroutine(TShape) void DrawTorus( in mat4 modelMat )
{
const int circumferenceTile = 12;
const int layersTile = 18;
const float torusRad = 0.8;
const float ringRad = 0.4;
mat4 modelViewMat = u_viewMat44 * modelMat;
mat3 normalMat = mat3( modelViewMat );
float preStepLay = 0.0;
mat4 prePosMat;
for ( int inxLay = 0; inxLay <= layersTile; ++ inxLay )
{
float stepLay = float(inxLay) / float(layersTile);
float angLay = 2.0 * 3.14159 * stepLay;
mat4 posMat = mat4(
vec4( cos(angLay), sin(angLay), 0.0, 0.0 ),
vec4( sin(angLay), cos(angLay), 0.0, 0.0 ),
vec4( 0.0, 0.0, 1.0, 0.0 ),
vec4( cos(angLay) * torusRad, sin(angLay) * torusRad, 0.0, 1.0 ) );
for ( int inxCir = 0; inxLay > 0 && inxCir <= circumferenceTile; ++ inxCir )
{
float stepCir = float(inxCir) / float(circumferenceTile);
float angCir = 2.0 * 3.14159 * stepCir;
vec2 ptCir = vec2( sin(angCir), cos(angCir) );
vec4 tempPt = vec4( ptCir.x * ringRad, 0.0, ptCir.y * ringRad, 1.0 );
vec4 pt1 = prePosMat * tempPt;
vec4 pt2 = posMat * tempPt;
NewVertexNvTex( pt1.xyz, modelViewMat, normalMat * normalize(pt1.xyz - prePosMat[3].xyz), vec2(stepCir, preStepLay*2.0) );
NewVertexNvTex( pt2.xyz, modelViewMat, normalMat * normalize(pt2.xyz - posMat[3].xyz), vec2(stepCir, stepLay*2.0) );
}
EndPrimitive();
preStepLay = stepLay;
prePosMat = posMat;
}
}
フラグメントシェーダ
subr.frag
#version 400
in TGeometryData
{
vec3 pos;
vec3 nv;
vec2 tex;
} inData;
out vec4 fragColor;
uniform sampler2D u_texture;
uniform UB_material
{
float u_roughness;
float u_fresnel0;
vec4 u_color;
vec4 u_specularTint;
};
struct TLightSource
{
vec4 ambient;
vec4 diffuse;
vec4 specular;
vec4 dir;
};
uniform UB_lightSource
{
TLightSource u_lightSource;
};
subroutine vec4 TSurface( void );
subroutine uniform TSurface su_surface;
float Fresnel_Schlick( in float theta );
vec3 LightModel( in vec3 esPt, in vec3 esPtNV, in vec3 col, in vec4 specularTint, in float roughness, in float fresnel0 );
void main()
{
vec4 fragCol = su_surface();
vec3 lightCol = LightModel( inData.pos, inData.nv, fragCol.rgb, u_specularTint, u_roughness, u_fresnel0 );
fragColor = vec4( clamp( lightCol, 0.0, 1.0 ), fragCol.a );
}
subroutine(TSurface) vec4 SurfaceColor( void )
{
return u_color;
}
subroutine(TSurface) vec4 SurfaceTexture( void )
{
return texture( u_texture, inData.tex.st );
}
float Fresnel_Schlick( in float theta )
{
float m = clamp( 1.0 - theta, 0.0, 1.0 );
float m2 = m * m;
return m2 * m2 * m; // pow( m, 5.0 )
}
vec3 LightModel( in vec3 esPt, in vec3 esPtNV, in vec3 col, in vec4 specularTint, in float roughness, in float fresnel0 )
{
vec3 esVLight = normalize( -u_lightSource.dir.xyz );
vec3 esVEye = normalize( -esPt );
vec3 halfVector = normalize( esVEye + esVLight );
float HdotL = dot( halfVector, esVLight );
float NdotL = dot( esPtNV, esVLight );
float NdotV = dot( esPtNV, esVEye );
float NdotH = dot( esPtNV, halfVector );
float NdotH2 = NdotH * NdotH;
float NdotL_clamped = max( NdotL, 0.0 );
float NdotV_clamped = max( NdotV, 0.0 );
float m2 = roughness * roughness;
// Lambertian diffuse
float k_diffuse = NdotL_clamped;
// Schlick approximation
float fresnel = fresnel0 + ( 1.0 - fresnel0 ) * Fresnel_Schlick( HdotL );
// Beckmann distribution
float distribution = max( 0.0, exp( ( NdotH2 - 1.0 ) / ( m2 * NdotH2 ) ) / ( 3.14159265 * m2 * NdotH2 * NdotH2 ) );
// Torrance-Sparrow geometric term
float geometric_att = min( 1.0, min( 2.0 * NdotH * NdotV_clamped / HdotL, 2.0 * NdotH * NdotL_clamped / HdotL ) );
// Microfacet bidirectional reflectance distribution function
float k_specular = fresnel * distribution * geometric_att / ( 4.0 * NdotL_clamped * NdotV_clamped );
vec3 lightColor = col.rgb * u_lightSource.ambient.rgb +
max( 0.0, k_diffuse ) * col.rgb * u_lightSource.diffuse.rgb +
max( 0.0, k_specular ) * mix( col.rgb, specularTint.rgb, specularTint.a ) * u_lightSource.specular.rgb;
return lightColor;
}
Phytonスクリプト
from OpenGL.GL import *
from OpenGL.GLUT import *
from OpenGL.GLU import *
import numpy as np
from time import time
import math
import sys
sin120 = 0.8660254
rotateCamera = False
# draw event
def OnDraw():
dist = 3.0
currentTime = time()
comeraRotAng = CalcAng( currentTime, 10.0 )
# set up projection matrix
prjMat = Perspective(90.0, wndW/wndH, 0.5, 100.0)
# set up view matrix
viewMat = np.matrix(np.identity(4), copy=False, dtype='float32')
viewMat = Translate( viewMat, np.array( [0.0, 0.0, -14.0] ) )
viewMat = RotateView( viewMat, [30.0, comeraRotAng if rotateCamera else 0.0, 0.0] )
# set up light source
lightSourceBuffer.BindDataFloat(b'u_lightSource.dir', TransformVec4([-1.0, -1.0, -5.0, 0.0], viewMat) )
# set up model matrices
modelMat = []
for inx in range(0, 2):
modelMat.append( np.matrix(np.identity(4), copy=False, dtype='float32') )
if not rotateCamera: modelMat[inx] = RotateY( modelMat[inx], comeraRotAng )
modelMat[0] = Scale( modelMat[0], np.repeat( 3, 3 ) )
modelMat[0] = Translate( modelMat[0], np.array( [0.0, 0.0, -2.0] ) )
modelMat[0] = RotateY( modelMat[0], CalcAng( currentTime, 23.0 ) )
modelMat[0] = RotateX( modelMat[0], CalcAng( currentTime, 13.0 ) )
modelMat[1] = Scale( modelMat[1], np.repeat( 3, 3 ) )
modelMat[1] = Translate( modelMat[1], np.array( [0.0, 0.0, 2.0] ) )
modelMat[1] = RotateY( modelMat[1], CalcAng( currentTime, 17.0 ) )
modelMat[1] = RotateX( modelMat[1], CalcAng( currentTime, 9.0 ) )
# set up texture matrix
texMat = np.matrix(np.identity(4), copy=False, dtype='float32')
# set up attributes and shader program
glEnable( GL_DEPTH_TEST )
glClear( GL_COLOR_BUFFER_BIT | GL_DEPTH_BUFFER_BIT )
glUseProgram( shaderProgram )
glUniformMatrix4fv( projectionMatLocation, 1, GL_FALSE, prjMat )
glUniformMatrix4fv( viewMatLocation, 1, GL_FALSE, viewMat )
glUniformMatrix4fv( textureMatLocation, 1, GL_FALSE, texMat )
glUniform1i( textureLocation, 0 )
lightSourceBuffer.BindToTarget()
# draw points
glBindVertexArray( pointVAObj )
for inx in range(0, 2):
# set up geometry shader subroutine
shape = 1 if inx==0 else 0 # 0: sphere, 1: torus
glUniformSubroutinesuiv(GL_GEOMETRY_SHADER, 1, np.array( [shape], dtype='uint' ))
# set up fragment shader subroutine
surfaceKind = inx # 0: color, 1: texture
glUniformSubroutinesuiv(GL_FRAGMENT_SHADER, 1, np.array( [surfaceKind], dtype='uint' ))
materialBuffer[inx].BindToTarget()
glUniformMatrix4fv( modelMatLocation, 1, GL_FALSE, modelMat[inx] )
glDrawArrays( GL_POINTS, 0, 1 )
glutSwapBuffers()
def Fract(val): return val - math.trunc(val)
def CalcAng(currentTime, intervall): return Fract( (currentTime - startTime) / intervall ) * 360.0
def CalcMove(currentTime, intervall, range):
pos = Fract( (currentTime - startTime) / intervall ) * 2.0
pos = pos if pos < 1.0 else (2.0-pos)
return range[0] + (range[1] - range[0]) * pos
# read shader program and compile shader
def CompileShader( sourceFileName, shaderStage ):
with open( sourceFileName, 'r' ) as sourceFile:
sourceCode = sourceFile.read()
nameMap = { GL_VERTEX_SHADER: 'vertex', GL_GEOMETRY_SHADER: 'geometry', GL_FRAGMENT_SHADER: 'fragment' }
print( '\n%s shader code:' % nameMap.get(shaderStage, '') )
print( sourceCode )
shaderObj = glCreateShader( shaderStage )
glShaderSource( shaderObj, sourceCode )
glCompileShader( shaderObj )
result = glGetShaderiv( shaderObj, GL_COMPILE_STATUS )
if not (result):
print( glGetShaderInfoLog( shaderObj ) )
sys.exit()
return shaderObj
# linke shader objects to shader program
def LinkProgram( shaderObjs ):
shaderProgram = glCreateProgram()
for shObj in shaderObjs:
glAttachShader( shaderProgram, shObj )
glLinkProgram( shaderProgram )
result = glGetProgramiv( shaderProgram, GL_LINK_STATUS )
if not (result):
print( 'link error:' )
print( glGetProgramInfoLog( shaderProgram ) )
sys.exit()
return shaderProgram
# create vertex array object
def CreateVAO( dataArrays ):
noOfBuffers = len(dataArrays)
buffers = glGenBuffers(noOfBuffers)
newVAObj = glGenVertexArrays( 1 )
glBindVertexArray( newVAObj )
for inx in range(0, noOfBuffers):
vertexSize, dataArr = dataArrays[inx]
arr = np.array( dataArr, dtype='float32' )
glBindBuffer( GL_ARRAY_BUFFER, buffers[inx] )
glBufferData( GL_ARRAY_BUFFER, arr, GL_STATIC_DRAW )
glEnableVertexAttribArray( inx )
glVertexAttribPointer( inx, vertexSize, GL_FLOAT, GL_FALSE, 0, None )
return newVAObj
# representation of a uniform block
class UniformBlock:
def __init__(self, shaderProg, name):
self.shaderProg = shaderProg
self.name = name
def Link(self, bindingPoint):
self.bindingPoint = bindingPoint
self.noOfUniforms = glGetProgramiv(self.shaderProg, GL_ACTIVE_UNIFORMS)
self.maxUniformNameLen = glGetProgramiv(self.shaderProg, GL_ACTIVE_UNIFORM_MAX_LENGTH)
self.index = glGetUniformBlockIndex(self.shaderProg, self.name)
intData = np.zeros(1, dtype=int)
glGetActiveUniformBlockiv(self.shaderProg, self.index, GL_UNIFORM_BLOCK_ACTIVE_UNIFORMS, intData)
self.count = intData[0]
self.indices = np.zeros(self.count, dtype=int)
glGetActiveUniformBlockiv(self.shaderProg, self.index, GL_UNIFORM_BLOCK_ACTIVE_UNIFORM_INDICES, self.indices)
self.offsets = np.zeros(self.count, dtype=int)
glGetActiveUniformsiv(self.shaderProg, self.count, self.indices, GL_UNIFORM_OFFSET, self.offsets)
strLengthData = np.zeros(1, dtype=int)
arraysizeData = np.zeros(1, dtype=int)
typeData = np.zeros(1, dtype='uint32')
nameData = np.chararray(self.maxUniformNameLen+1)
self.namemap = {}
self.dataSize = 0
for inx in range(0, len(self.indices)):
glGetActiveUniform( self.shaderProg, self.indices[inx], self.maxUniformNameLen, strLengthData, arraysizeData, typeData, nameData.data )
name = nameData.tostring()[:strLengthData[0]]
self.namemap[name] = inx
self.dataSize = max(self.dataSize, self.offsets[inx] + arraysizeData * 16)
glUniformBlockBinding(self.shaderProg, self.index, self.bindingPoint)
print('\nuniform block %s size:%4d' % (self.name, self.dataSize))
for uName in self.namemap:
print( ' %-40s index:%2d offset:%4d' % (uName, self.indices[self.namemap[uName]], self.offsets[self.namemap [uName]]) )
# representation of a uniform block buffer
class UniformBlockBuffer:
def __init__(self, ub):
self.namemap = ub.namemap
self.offsets = ub.offsets
self.bindingPoint = ub.bindingPoint
self.object = glGenBuffers(1)
self.dataSize = ub.dataSize
glBindBuffer(GL_UNIFORM_BUFFER, self.object)
dataArray = np.zeros(self.dataSize//4, dtype='float32')
glBufferData(GL_UNIFORM_BUFFER, self.dataSize, dataArray, GL_DYNAMIC_DRAW)
def BindToTarget(self):
glBindBuffer(GL_UNIFORM_BUFFER, self.object)
glBindBufferBase(GL_UNIFORM_BUFFER, self.bindingPoint, self.object)
def BindDataFloat(self, name, dataArr):
glBindBuffer(GL_UNIFORM_BUFFER, self.object)
dataArray = np.array(dataArr, dtype='float32')
glBufferSubData(GL_UNIFORM_BUFFER, self.offsets[self.namemap[name]], len(dataArr)*4, dataArray)
def Translate(matA, trans):
matB = np.copy(matA)
for i in range(0, 4): matB[3,i] = matA[0,i] * trans[0] + matA[1,i] * trans[1] + matA[2,i] * trans[2] + matA[3,i]
return matB
def Scale(matA, s):
matB = np.copy(matA)
for i0 in range(0, 3):
for i1 in range(0, 4): matB[i0,i1] = matA[i0,i1] * s[i0]
return matB
def RotateHlp(matA, angDeg, a0, a1):
matB = np.copy(matA)
ang = math.radians(angDeg)
sinAng, cosAng = math.sin(ang), math.cos(ang)
for i in range(0, 4):
matB[a0,i] = matA[a0,i] * cosAng + matA[a1,i] * sinAng
matB[a1,i] = matA[a0,i] * -sinAng + matA[a1,i] * cosAng
return matB
def RotateX(matA, angDeg): return RotateHlp(matA, angDeg, 1, 2)
def RotateY(matA, angDeg): return RotateHlp(matA, angDeg, 2, 0)
def RotateZ(matA, angDeg): return RotateHlp(matA, angDeg, 0, 1)
def RotateView(matA, angDeg): return RotateZ(RotateY(RotateX(matA, angDeg[0]), angDeg[1]), angDeg[2])
def Multiply(matA, matB):
matC = np.copy(matA)
for i0 in range(0, 4):
for i1 in range(0, 4):
matC[i0,i1] = matB[i0,0] * matA[0,i1] + matB[i0,1] * matA[1,i1] + matB[i0,2] * matA[2,i1] + matB[i0,3] * matA[3,i1]
return matC
def ToMat33(mat44):
mat33 = np.matrix(np.identity(3), copy=False, dtype='float32')
for i0 in range(0, 3):
for i1 in range(0, 3): mat33[i0, i1] = mat44[i0, i1]
return mat33
def TransformVec4(vecA,mat44):
vecB = np.zeros(4, dtype='float32')
for i0 in range(0, 4):
vecB[i0] = vecA[0] * mat44[0,i0] + vecA[1] * mat44[1,i0] + vecA[2] * mat44[2,i0] + vecA[3] * mat44[3,i0]
return vecB
def Perspective(fov, aspectRatio, near, far):
fn, f_n = far + near, far - near
r, t = aspectRatio, 1.0 / math.tan( math.radians(fov) / 2.0 )
return np.matrix( [ [t/r,0,0,0], [0,t,0,0], [0,0,-fn/f_n,-2.0*far*near/f_n], [0,0,-1,0] ] )
def AddToBuffer( buffer, data, count=1 ):
for inx_c in range(0, count):
for inx_s in range(0, len(data)): buffer.append( data[inx_s] )
# initialize glut
glutInit()
# create window
wndW, wndH = 800, 600
glutInitDisplayMode(GLUT_RGBA | GLUT_DOUBLE | GLUT_ALPHA | GLUT_DEPTH)
glutInitWindowPosition(0, 0)
glutInitWindowSize(wndW, wndH)
wndID = glutCreateWindow(b'OGL window')
glutDisplayFunc(OnDraw)
glutIdleFunc(OnDraw)
# define location vertex array opject
pointVAObj = CreateVAO( [ (3, [0.0, 0.0, 0.0] ), (3, [0.0, 0.0, 1.0]), (3, [1.0, 0.0, 0.0]) ] )
# create texture
texCX, texCY = 128, 128
texPlan = np.zeros( texCX * texCY * 4, dtype=np.uint8 )
for inx_x in range(0, texCX):
for inx_y in range(0, texCY):
val_x = math.sin( math.pi * 6.0 * inx_x / texCX )
val_y = math.sin( math.pi * 6.0 * inx_y / texCY )
inx_tex = inx_y * texCX * 4 + inx_x * 4
texPlan[inx_tex + 0] = int( 128 + 127 * val_x )
texPlan[inx_tex + 1] = 63
texPlan[inx_tex + 2] = int( 128 + 127 * val_y )
texPlan[inx_tex + 3] = 255
glActiveTexture( GL_TEXTURE0 )
texObj = glGenTextures( 1 )
glBindTexture( GL_TEXTURE_2D, texObj )
glTexImage2D(GL_TEXTURE_2D, 0, GL_RGBA, texCX, texCY, 0, GL_RGBA, GL_UNSIGNED_BYTE, texPlan)
glTexParameterf(GL_TEXTURE_2D, GL_TEXTURE_MAG_FILTER, GL_LINEAR)
glTexParameterf(GL_TEXTURE_2D, GL_TEXTURE_MIN_FILTER, GL_LINEAR)
glTexParameterf(GL_TEXTURE_2D, GL_TEXTURE_WRAP_S, GL_REPEAT)
glTexParameterf(GL_TEXTURE_2D, GL_TEXTURE_WRAP_T, GL_REPEAT)
# load, compile and link shader
shaderProgram = LinkProgram( [
CompileShader( 'python/ogl4subr/subr.vert', GL_VERTEX_SHADER ),
CompileShader( 'python/ogl4subr/subr.geo', GL_GEOMETRY_SHADER ),
CompileShader( 'python/ogl4subr/subr.frag', GL_FRAGMENT_SHADER )
] )
# get unifor locations
projectionMatLocation = glGetUniformLocation(shaderProgram, "u_projectionMat44")
viewMatLocation = glGetUniformLocation(shaderProgram, "u_viewMat44")
modelMatLocation = glGetUniformLocation(shaderProgram, "u_modelMat44")
textureMatLocation = glGetUniformLocation(shaderProgram, "u_textureMat44")
textureLocation = glGetUniformLocation(shaderProgram, "u_texture")
# linke uniform blocks
ubMaterial = UniformBlock(shaderProgram, "UB_material")
ubLightSource = UniformBlock(shaderProgram, "UB_lightSource")
ubMaterial.Link(1)
ubLightSource.Link(2)
# create uniform block buffers
lightSourceBuffer = UniformBlockBuffer(ubLightSource)
lightSourceBuffer.BindDataFloat(b'u_lightSource.ambient', [0.2, 0.2, 0.2, 1.0])
lightSourceBuffer.BindDataFloat(b'u_lightSource.diffuse', [0.2, 0.2, 0.2, 1.0])
lightSourceBuffer.BindDataFloat(b'u_lightSource.specular', [1.0, 1.0, 1.0, 1.0])
materialBuffer = [ UniformBlockBuffer(ubMaterial), UniformBlockBuffer(ubMaterial) ]
materialBuffer[0].BindDataFloat(b'u_roughness', [0.45])
materialBuffer[0].BindDataFloat(b'u_fresnel0', [0.45])
materialBuffer[0].BindDataFloat(b'u_color', [0.5, 0.7, 0.6, 1.0])
materialBuffer[0].BindDataFloat(b'u_specularTint',[1.0, 0.5, 0.5, 0.8])
materialBuffer[1].BindDataFloat(b'u_roughness', [0.4])
materialBuffer[1].BindDataFloat(b'u_fresnel0', [0.4])
materialBuffer[1].BindDataFloat(b'u_color', [0.7, 0.5, 0.6, 1.0])
materialBuffer[1].BindDataFloat(b'u_specularTint',[0.5, 1.0, 0.5, 0.8])
# start main loop
startTime = time()
glutMainLoop()
OGL 4.0 GLSLのテッセレーションシェーダでジオメトリを変更する
ジオメトリにテッセレーションシェーダの詳細を追加する方法を示すシンプルなOGL 4.0 GLSLシェーダプログラム。プログラムはpythonスクリプトで実行されます。スクリプトを実行するには、PyOpenGLとNumPyをインストールする必要があります。
この例の基本的なメッシュは20の三角形からなる20面体です。テッセレーション制御シェーダは、各三角形が多くの小さな部分のセットにどのように分割されるかを定義します。三角形をテッセレーションするとき、生成されたデータは元の三角形に基づく重心座標です。テッセレーション評価シェーダは、この方法で取得したデータから新しいジオメトリを生成します。この例では、各三角形は中央にピークを持ち、icosaderの中心から外側に向かっています。このようにして、元の20面体よりはるかに複雑な形状が生成される。
頂点シェーダ
tess.vert
layout (location = 0) in vec3 inPos;
layout (location = 1) in vec3 inNV;
out TVertexData
{
vec3 pos;
vec3 nv;
} outData;
uniform mat4 u_projectionMat44;
uniform mat4 u_modelViewMat44;
uniform mat3 u_normalMat33;
void main()
{
vec4 viewPos = u_modelViewMat44 * vec4( inPos, 1.0 );
outData.pos = viewPos.xyz / viewPos.w;
outData.nv = u_normalMat33 * normalize( inNV );
gl_Position = u_projectionMat44 * viewPos;
}
テッセレーションコントロールシェーダ
tess.tctrl
#version 400
layout( vertices=3 ) out;
in TVertexData
{
vec3 pos;
vec3 nv;
} inData[];
out TVertexData
{
vec3 pos;
vec3 nv;
} outData[];
void main()
{
outData[gl_InvocationID].pos = inData[gl_InvocationID].pos;
outData[gl_InvocationID].nv = inData[gl_InvocationID].nv;
if ( gl_InvocationID == 0 )
{
gl_TessLevelOuter[0] = 10.0;
gl_TessLevelOuter[1] = 10.0;
gl_TessLevelOuter[2] = 10.0;
gl_TessLevelInner[0] = 10.0;
}
}
テッセレーション評価シェーダ
tess.teval
#version 400
layout(triangles, equal_spacing, ccw) in;
in TVertexData
{
vec3 pos;
vec3 nv;
} inData[];
out TTessData
{
vec3 pos;
vec3 nv;
float height;
} outData;
uniform mat4 u_projectionMat44;
void main()
{
float sideLen[3] = float[3]
(
length( inData[1].pos - inData[0].pos ),
length( inData[2].pos - inData[1].pos ),
length( inData[0].pos - inData[2].pos )
);
float s = ( sideLen[0] + sideLen[1] + sideLen[2] ) / 2.0;
float rad = sqrt( (s - sideLen[0]) * (s - sideLen[1]) * (s - sideLen[2]) / s );
vec3 cpt = ( inData[0].pos + inData[1].pos + inData[2].pos ) / 3.0;
vec3 pos = inData[0].pos * gl_TessCoord.x + inData[1].pos * gl_TessCoord.y + inData[2].pos * gl_TessCoord.z;
vec3 nv = normalize( inData[0].nv * gl_TessCoord.x + inData[1].nv * gl_TessCoord.y + inData[2].nv * gl_TessCoord.z );
float cptDist = length( cpt - pos );
float sizeRelation = 1.0 - min( rad, cptDist ) / rad;
float height = pow( sizeRelation, 2.0 );
outData.pos = pos + nv * height * rad;
outData.nv = mix( nv, normalize( pos - cpt ), height );
outData.height = height;
gl_Position = u_projectionMat44 * vec4( outData.pos, 1.0 );
}
フラグメントシェーダ
tess.frag
#version 400
in TTessData
{
vec3 pos;
vec3 nv;
float height;
} inData;
out vec4 fragColor;
uniform sampler2D u_texture;
uniform UB_material
{
float u_roughness;
float u_fresnel0;
vec4 u_color;
vec4 u_specularTint;
};
struct TLightSource
{
vec4 ambient;
vec4 diffuse;
vec4 specular;
vec4 dir;
};
uniform UB_lightSource
{
TLightSource u_lightSource;
};
float Fresnel_Schlick( in float theta );
vec3 LightModel( in vec3 esPt, in vec3 esPtNV, in vec3 col, in vec4 specularTint, in float roughness, in float fresnel0 );
void main()
{
vec3 col = mix( u_color.rgb, vec3( 1.0, 1.0, 1.0 ), inData.height );
vec3 lightCol = LightModel( inData.pos, inData.nv, col, u_specularTint, u_roughness, u_fresnel0 );
fragColor = vec4( clamp( lightCol, 0.0, 1.0 ), 1.0 );
}
float Fresnel_Schlick( in float theta )
{
float m = clamp( 1.0 - theta, 0.0, 1.0 );
float m2 = m * m;
return m2 * m2 * m; // pow( m, 5.0 )
}
vec3 LightModel( in vec3 esPt, in vec3 esPtNV, in vec3 col, in vec4 specularTint, in float roughness, in float fresnel0 )
{
vec3 esVLight = normalize( -u_lightSource.dir.xyz );
vec3 esVEye = normalize( -esPt );
vec3 halfVector = normalize( esVEye + esVLight );
float HdotL = dot( halfVector, esVLight );
float NdotL = dot( esPtNV, esVLight );
float NdotV = dot( esPtNV, esVEye );
float NdotH = dot( esPtNV, halfVector );
float NdotH2 = NdotH * NdotH;
float NdotL_clamped = max( NdotL, 0.0 );
float NdotV_clamped = max( NdotV, 0.0 );
float m2 = roughness * roughness;
// Lambertian diffuse
float k_diffuse = NdotL_clamped;
// Schlick approximation
float fresnel = fresnel0 + ( 1.0 - fresnel0 ) * Fresnel_Schlick( HdotL );
// Beckmann distribution
float distribution = max( 0.0, exp( ( NdotH2 - 1.0 ) / ( m2 * NdotH2 ) ) / ( 3.14159265 * m2 * NdotH2 * NdotH2 ) );
// Torrance-Sparrow geometric term
float geometric_att = min( 1.0, min( 2.0 * NdotH * NdotV_clamped / HdotL, 2.0 * NdotH * NdotL_clamped / HdotL ) );
// Microfacet bidirectional reflectance distribution function
float k_specular = fresnel * distribution * geometric_att / ( 4.0 * NdotL_clamped * NdotV_clamped );
vec3 lightColor = col.rgb * u_lightSource.ambient.rgb +
max( 0.0, k_diffuse ) * col.rgb * u_lightSource.diffuse.rgb +
max( 0.0, k_specular ) * mix( col.rgb, specularTint.rgb, specularTint.a ) * u_lightSource.specular.rgb;
return lightColor;
}
Pythonスクリプト
from OpenGL.GL import *
from OpenGL.GLUT import *
from OpenGL.GLU import *
import numpy as np
from time import time
import math
import sys
sin120 = 0.8660254
rotateCamera = False
# draw event
def OnDraw():
dist = 3.0
currentTime = time()
comeraRotAng = CalcAng( currentTime, 10.0 )
# set up projection matrix
prjMat = Perspective(90.0, wndW/wndH, 0.5, 100.0)
# set up view matrix
viewMat = np.matrix(np.identity(4), copy=False, dtype='float32')
viewMat = Translate( viewMat, np.array( [0.0, 0.0, -12.0] ) )
viewMat = RotateView( viewMat, [30.0, comeraRotAng if rotateCamera else 0.0, 0.0] )
# set up light source
lightSourceBuffer.BindDataFloat(b'u_lightSource.dir', TransformVec4([-1.0, -1.0, -5.0, 0.0], viewMat) )
# set up icosahedron model matrix
icoModelMat = np.matrix(np.identity(4), copy=False, dtype='float32')
if not rotateCamera: icoModelMat = RotateY( icoModelMat, comeraRotAng )
icoModelMat = Scale( icoModelMat, np.repeat( 5, 3 ) )
icoModelMat = RotateY( icoModelMat, CalcAng( currentTime, 17.0 ) )
icoModelMat = RotateX( icoModelMat, CalcAng( currentTime, 13.0 ) )
# set up attributes and shader program
glEnable( GL_DEPTH_TEST )
glClear( GL_COLOR_BUFFER_BIT | GL_DEPTH_BUFFER_BIT )
glUseProgram( shaderProgram )
glUniformMatrix4fv( projectionMatLocation, 1, GL_FALSE, prjMat )
lightSourceBuffer.BindToTarget()
# draw icosahedron
icoMaterialBuffer.BindToTarget()
modelViewMat = Multiply(viewMat, icoModelMat)
glUniformMatrix4fv( modelViewMatLocation, 1, GL_FALSE, modelViewMat )
glUniformMatrix3fv( normalMatLocation, 1, GL_FALSE, ToMat33(modelViewMat) )
glBindVertexArray( icoVAObj )
glPatchParameteri( GL_PATCH_VERTICES, 3 )
glDrawArrays( GL_PATCHES, 0, len(icoPosData) )
glutSwapBuffers()
def Fract(val): return val - math.trunc(val)
def CalcAng(currentTime, intervall): return Fract( (currentTime - startTime) / intervall ) * 360.0
def CalcMove(currentTime, intervall, range):
pos = Fract( (currentTime - startTime) / intervall ) * 2.0
pos = pos if pos < 1.0 else (2.0-pos)
return range[0] + (range[1] - range[0]) * pos
# read shader program and compile shader
def CompileShader( sourceFileName, shaderStage ):
with open( sourceFileName, 'r' ) as sourceFile:
sourceCode = sourceFile.read()
nameMap = { GL_VERTEX_SHADER: 'vertex', GL_GEOMETRY_SHADER: 'geometry', GL_FRAGMENT_SHADER: 'fragment' }
print( '\n%s shader code:' % nameMap.get(shaderStage, '') )
print( sourceCode )
shaderObj = glCreateShader( shaderStage )
glShaderSource( shaderObj, sourceCode )
glCompileShader( shaderObj )
result = glGetShaderiv( shaderObj, GL_COMPILE_STATUS )
if not (result):
print( glGetShaderInfoLog( shaderObj ) )
sys.exit()
return shaderObj
# link shader objects to shader program
def LinkProgram( shaderObjs ):
shaderProgram = glCreateProgram()
for shObj in shaderObjs:
glAttachShader( shaderProgram, shObj )
glLinkProgram( shaderProgram )
result = glGetProgramiv( shaderProgram, GL_LINK_STATUS )
if not (result):
print( 'link error:' )
print( glGetProgramInfoLog( shaderProgram ) )
sys.exit()
return shaderProgram
# create vertex array object
def CreateVAO( dataArrays ):
noOfBuffers = len(dataArrays)
buffers = glGenBuffers(noOfBuffers)
newVAObj = glGenVertexArrays( 1 )
glBindVertexArray( newVAObj )
for inx in range(0, noOfBuffers):
vertexSize, dataArr = dataArrays[inx]
arr = np.array( dataArr, dtype='float32' )
glBindBuffer( GL_ARRAY_BUFFER, buffers[inx] )
glBufferData( GL_ARRAY_BUFFER, arr, GL_STATIC_DRAW )
glEnableVertexAttribArray( inx )
glVertexAttribPointer( inx, vertexSize, GL_FLOAT, GL_FALSE, 0, None )
return newVAObj
# representation of a uniform block
class UniformBlock:
def __init__(self, shaderProg, name):
self.shaderProg = shaderProg
self.name = name
def Link(self, bindingPoint):
self.bindingPoint = bindingPoint
self.noOfUniforms = glGetProgramiv(self.shaderProg, GL_ACTIVE_UNIFORMS)
self.maxUniformNameLen = glGetProgramiv(self.shaderProg, GL_ACTIVE_UNIFORM_MAX_LENGTH)
self.index = glGetUniformBlockIndex(self.shaderProg, self.name)
intData = np.zeros(1, dtype=int)
glGetActiveUniformBlockiv(self.shaderProg, self.index, GL_UNIFORM_BLOCK_ACTIVE_UNIFORMS, intData)
self.count = intData[0]
self.indices = np.zeros(self.count, dtype=int)
glGetActiveUniformBlockiv(self.shaderProg, self.index, GL_UNIFORM_BLOCK_ACTIVE_UNIFORM_INDICES, self.indices)
self.offsets = np.zeros(self.count, dtype=int)
glGetActiveUniformsiv(self.shaderProg, self.count, self.indices, GL_UNIFORM_OFFSET, self.offsets)
strLengthData = np.zeros(1, dtype=int)
arraysizeData = np.zeros(1, dtype=int)
typeData = np.zeros(1, dtype='uint32')
nameData = np.chararray(self.maxUniformNameLen+1)
self.namemap = {}
self.dataSize = 0
for inx in range(0, len(self.indices)):
glGetActiveUniform( self.shaderProg, self.indices[inx], self.maxUniformNameLen, strLengthData, arraysizeData, typeData, nameData.data )
name = nameData.tostring()[:strLengthData[0]]
self.namemap[name] = inx
self.dataSize = max(self.dataSize, self.offsets[inx] + arraysizeData * 16)
glUniformBlockBinding(self.shaderProg, self.index, self.bindingPoint)
print('\nuniform block %s size:%4d' % (self.name, self.dataSize))
for uName in self.namemap:
print( ' %-40s index:%2d offset:%4d' % (uName, self.indices[self.namemap[uName]], self.offsets[self.namemap [uName]]) )
# representation of a uniform block buffer
class UniformBlockBuffer:
def __init__(self, ub):
self.namemap = ub.namemap
self.offsets = ub.offsets
self.bindingPoint = ub.bindingPoint
self.object = glGenBuffers(1)
self.dataSize = ub.dataSize
glBindBuffer(GL_UNIFORM_BUFFER, self.object)
dataArray = np.zeros(self.dataSize//4, dtype='float32')
glBufferData(GL_UNIFORM_BUFFER, self.dataSize, dataArray, GL_DYNAMIC_DRAW)
def BindToTarget(self):
glBindBuffer(GL_UNIFORM_BUFFER, self.object)
glBindBufferBase(GL_UNIFORM_BUFFER, self.bindingPoint, self.object)
def BindDataFloat(self, name, dataArr):
glBindBuffer(GL_UNIFORM_BUFFER, self.object)
dataArray = np.array(dataArr, dtype='float32')
glBufferSubData(GL_UNIFORM_BUFFER, self.offsets[self.namemap[name]], len(dataArr)*4, dataArray)
def Translate(matA, trans):
matB = np.copy(matA)
for i in range(0, 4): matB[3,i] = matA[0,i] * trans[0] + matA[1,i] * trans[1] + matA[2,i] * trans[2] + matA[3,i]
return matB
def Scale(matA, s):
matB = np.copy(matA)
for i0 in range(0, 3):
for i1 in range(0, 4): matB[i0,i1] = matA[i0,i1] * s[i0]
return matB
def RotateHlp(matA, angDeg, a0, a1):
matB = np.copy(matA)
ang = math.radians(angDeg)
sinAng, cosAng = math.sin(ang), math.cos(ang)
for i in range(0, 4):
matB[a0,i] = matA[a0,i] * cosAng + matA[a1,i] * sinAng
matB[a1,i] = matA[a0,i] * -sinAng + matA[a1,i] * cosAng
return matB
def RotateX(matA, angDeg): return RotateHlp(matA, angDeg, 1, 2)
def RotateY(matA, angDeg): return RotateHlp(matA, angDeg, 2, 0)
def RotateZ(matA, angDeg): return RotateHlp(matA, angDeg, 0, 1)
def RotateView(matA, angDeg): return RotateZ(RotateY(RotateX(matA, angDeg[0]), angDeg[1]), angDeg[2])
def Multiply(matA, matB):
matC = np.copy(matA)
for i0 in range(0, 4):
for i1 in range(0, 4):
matC[i0,i1] = matB[i0,0] * matA[0,i1] + matB[i0,1] * matA[1,i1] + matB[i0,2] * matA[2,i1] + matB[i0,3] * matA[3,i1]
return matC
def ToMat33(mat44):
mat33 = np.matrix(np.identity(3), copy=False, dtype='float32')
for i0 in range(0, 3):
for i1 in range(0, 3): mat33[i0, i1] = mat44[i0, i1]
return mat33
def TransformVec4(vecA,mat44):
vecB = np.zeros(4, dtype='float32')
for i0 in range(0, 4):
vecB[i0] = vecA[0] * mat44[0,i0] + vecA[1] * mat44[1,i0] + vecA[2] * mat44[2,i0] + vecA[3] * mat44[3,i0]
return vecB
def Perspective(fov, aspectRatio, near, far):
fn, f_n = far + near, far - near
r, t = aspectRatio, 1.0 / math.tan( math.radians(fov) / 2.0 )
return np.matrix( [ [t/r,0,0,0], [0,t,0,0], [0,0,-fn/f_n,-2.0*far*near/f_n], [0,0,-1,0] ] )
def AddToBuffer( buffer, data, count=1 ):
for inx_c in range(0, count):
for inx_s in range(0, len(data)): buffer.append( data[inx_s] )
# initialize glut
glutInit()
# create window
wndW, wndH = 800, 600
glutInitDisplayMode(GLUT_RGBA | GLUT_DOUBLE | GLUT_ALPHA | GLUT_DEPTH)
glutInitWindowPosition(0, 0)
glutInitWindowSize(wndW, wndH)
wndID = glutCreateWindow(b'OGL window')
glutDisplayFunc(OnDraw)
glutIdleFunc(OnDraw)
# define icosahedron vertex array opject
icoPts = [
( 0.000, 0.000, 1.000), ( 0.894, 0.000, 0.447), ( 0.276, 0.851, 0.447), (-0.724, 0.526, 0.447),
(-0.724, -0.526, 0.447), ( 0.276, -0.851, 0.447), ( 0.724, 0.526, -0.447), (-0.276, 0.851, -0.447),
(-0.894, 0.000, -0.447), (-0.276, -0.851, -0.447), ( 0.724, -0.526, -0.447), ( 0.000, 0.000, -1.000) ]
icoCol = [ [1.0, 0.0, 0.0], [0.0, 0.0, 1.0], [1.0, 1.0, 0.0], [0.0, 1.0, 0.0], [1.0, 0.5, 0.0], [1.0, 0.0, 1.0] ]
icoIndices = [
2, 0, 1, 3, 0, 2, 4, 0, 3, 5, 0, 4, 1, 0, 5, 11, 7, 6, 11, 8, 7, 11, 9, 8, 11, 10, 9, 11, 6, 10,
1, 6, 2, 2, 7, 3, 3, 8, 4, 4, 9, 5, 5, 10, 1, 2, 6, 7, 3, 7, 8, 4, 8, 9, 5, 9, 10, 1, 10, 6 ]
icoPosData = []
for inx in icoIndices: AddToBuffer( icoPosData, icoPts[inx] )
icoNVData = []
for inx_nv in range(0, len(icoIndices) // 3):
nv = [0.0, 0.0, 0.0]
for inx_p in range(0, 3):
for inx_s in range(0, 3): nv[inx_s] += icoPts[ icoIndices[inx_nv*3 + inx_p] ][inx_s]
AddToBuffer( icoNVData, nv, 3 )
icoVAObj = CreateVAO( [ (3, icoPosData), (3, icoNVData) ] )
# load, compile and link shader
shaderProgram = LinkProgram( [
CompileShader( 'tess.vert', GL_VERTEX_SHADER ),
CompileShader( 'tess.tctrl', GL_TESS_CONTROL_SHADER ),
CompileShader( 'tess.teval', GL_TESS_EVALUATION_SHADER ),
CompileShader( 'tess.frag', GL_FRAGMENT_SHADER )
] )
# get unifor locations
projectionMatLocation = glGetUniformLocation(shaderProgram, "u_projectionMat44")
modelViewMatLocation = glGetUniformLocation(shaderProgram, "u_modelViewMat44")
normalMatLocation = glGetUniformLocation(shaderProgram, "u_normalMat33")
# linke uniform blocks
ubMaterial = UniformBlock(shaderProgram, "UB_material")
ubLightSource = UniformBlock(shaderProgram, "UB_lightSource")
ubMaterial.Link(1)
ubLightSource.Link(2)
# create uniform block buffers
lightSourceBuffer = UniformBlockBuffer(ubLightSource)
lightSourceBuffer.BindDataFloat(b'u_lightSource.ambient', [0.2, 0.2, 0.2, 1.0])
lightSourceBuffer.BindDataFloat(b'u_lightSource.diffuse', [0.2, 0.2, 0.2, 1.0])
lightSourceBuffer.BindDataFloat(b'u_lightSource.specular', [1.0, 1.0, 1.0, 1.0])
icoMaterialBuffer = UniformBlockBuffer(ubMaterial)
icoMaterialBuffer.BindDataFloat(b'u_roughness', [0.45])
icoMaterialBuffer.BindDataFloat(b'u_fresnel0', [0.4])
icoMaterialBuffer.BindDataFloat(b'u_color', [0.6, 0.5, 0.8, 1.0])
icoMaterialBuffer.BindDataFloat(b'u_specularTint',[1.0, 0.5, 0.5, 0.8])
# start main loop
startTime = time()
glutMainLoop()
Modified text is an extract of the original Stack Overflow Documentation
ライセンスを受けた CC BY-SA 3.0
所属していない Stack Overflow
