问题描述
我正在研究一个适用于Quartz Composer的自定义几何库，并尝试在Plug In中绘制一些凹面多边形.
I'm working on a custom geometry library adapted to Quartz Composer and trying to draw some concave polygons in a Plug In.
我实现了poly2tri库，因此用户可以选择是否进行三角剖分，但这不适用于按帧进行多边形变换的渲染.
I implemented the poly2tri library, so the user can choose to triangulate or not, but it's not suitable for a per-frame polygon transformations rendering.
我是OpenGL的新手，我已经阅读和测试了很多有关模版缓冲区和奇/偶操作的内容，但是即使代码似乎对其他人也有用，但对我却不起作用.
I'm a noob in OpenGL and I've been reading and testing a lot, about stencil buffer and odd/even operations, but even code that seem to work for other people, doesn't work for me.
渲染上下文是CGLContextObj，我正在使用NVidia GEForce GT650开发MacBook Pro Retina显示器.我读到所有配置都没有模版缓冲区，但是看起来有时它可以工作，尽管不是我想要的那样.
The render context is a CGLContextObj, and I'm working on a MacBook Pro Retina Display, with NVidia GEForce GT650. I read that all configurations don't have stencil buffers, but it look like it works sometimes, though not as I would like it to.
我想知道是否有人使用相同的配置，然后使用的代码可以正常工作，并且可以查看我的代码.尤其是，我也对请求的通过次数感到好奇，根据我猜想的顶点或凸性缺陷"的数量...
I was wondering if someone with the same kind of config was using a code that works and could take a look at my code. In particular, I'm curious too about the number of passes requested, according to the number of vertices or "convexity defects" I guess...
我从以下位置获取了我的信息:
I took my infos from :
- http://fly.cc.fer.hr/~unreal /theredbook/chapter13.html
- http://commaexcess.com/articles/7/concave-polygon -triangulation-shortcut
- http://graphicsbb.itgo.com/solutions/extrude.html
- http: //analysesmusings.wordpress.com/2012/07/13/drawing-filled-concave-polygons-using-the-stencil-buffer/
- http://fly.cc.fer.hr/~unreal/theredbook/chapter13.html
- http://commaexcess.com/articles/7/concave-polygon-triangulation-shortcut
- http://graphicsbb.itgo.com/solutions/extrude.html
- http://analysesmusings.wordpress.com/2012/07/13/drawing-filled-concave-polygons-using-the-stencil-buffer/
...但仍然不清楚...
... but still not clear...
这是我的代码(实际上，当我测试了如此多的配置时，其中之一)和结果的图片.实际上，我通常将实际的渲染放入每个多边形调用的方法中，但我将其重写为更加清晰:
Here is my code (one of them in fact, as I tested so much configurations) and a picture of the result. Actually I use to put the actual rendering in a method called for each polygon, but I rewrote it to be much clear :
编辑
事实上，我理解我必须绘制每个三角形，以便反转模板缓冲区中的位值.因此，我将代码重写为:
In fact I understood that I have to draw each triangle, in order to invert the bit value in the stencil buffer. So I rewrote my code into this :
CGLContextObj cgl_ctx = [context CGLContextObj];
CGLLockContext(cgl_ctx);
GLenum error;
if(cgl_ctx == NULL)
return NO;
glPushAttrib(GL_ALL_ATTRIB_BITS);
glPushClientAttrib(GL_CLIENT_ALL_ATTRIB_BITS);
glEnable(GL_BLEND);
glBlendFunc(GL_SRC_ALPHA, GL_ONE_MINUS_SRC_ALPHA);
glDisable(GL_CULL_FACE);
glClear(GL_STENCIL_BUFFER_BIT);
glClearStencil(0);
glEnable(GL_STENCIL_TEST);
glStencilOp(GL_INVERT, GL_INVERT, GL_INVERT);
glStencilFunc(GL_ALWAYS, 1, 1);
glColorMask(GL_FALSE, GL_FALSE, GL_FALSE, GL_FALSE);
// glColor4d(1., 1., 1., 1.); ----> does it make sense ?
glBegin(GL_TRIANGLE_FAN); {
for (int i = 1; i < [vertices count] - 1; i++) {
// Allways drawing the first vertex
glVertex2d([[[vertices objectAtIndex:0] objectAtIndex:0] doubleValue], [[[vertices objectAtIndex:0] objectAtIndex:1] doubleValue]);
// Then two others to make a triangle
glVertex2d([[[vertices objectAtIndex:i] objectAtIndex:0] doubleValue], [[[vertices objectAtIndex:i] objectAtIndex:1] doubleValue]);
glVertex2d([[[vertices objectAtIndex:i+1] objectAtIndex:0] doubleValue], [[[vertices objectAtIndex:i+1] objectAtIndex:1] doubleValue]);
}
}
glEnd();
glStencilOp(GL_ZERO, GL_ZERO, GL_ZERO);
glStencilFunc(GL_EQUAL, 1, 1);
glColorMask(GL_TRUE, GL_TRUE, GL_TRUE, GL_TRUE);
glColor4d(1., 0., 0., 0.5);
glBegin(GL_TRIANGLE_FAN); {
for (id vertex in vertices) {
glVertex2d([[vertex objectAtIndex:0] doubleValue], [[vertex objectAtIndex:1] doubleValue]);
}
glVertex2d([[[vertices objectAtIndex:0] objectAtIndex:0] doubleValue], [[[vertices objectAtIndex:0] objectAtIndex:1] doubleValue]);
}
glEnd();
glDisable (GL_STENCIL_TEST);
glDisable(GL_BLEND);
glPopClientAttrib();
glPopAttrib();
if((error = glGetError()))
NSLog(@"OpenGL error %04X", error);
CGLUnlockContext(cgl_ctx);
return (error ? NO : YES);
但是它仍然不起作用.这是我的结果，以及原始图像和说明的链接.
But it still doesn't work. Here is my result and the link to the original image and the explanation.
实际上，由Quartz Composer启用的上下文没有实现模板缓冲区.用模板缓冲区直接在OpenGL中渲染似乎是不可能的.
In fact, the context enabled by Quartz Composer doesn't implement a stencil buffer. It seems impossible to render directly in OpenGL with the stencil buffer.
推荐答案
...
glClearStencil(0);
...
请注意， glClearStencil() 只是设置了一点状态，实际上并不会清除模板缓冲区.
Be aware that glClearStencil() just sets a bit of state and doesn't actually clear the stencil buffer.
尝试在每个多边形之前的某个位置添加glClear( GL_STENCIL_BUFFER_BIT ).
Try adding a glClear( GL_STENCIL_BUFFER_BIT ) somewhere before each polygon.
像这样:
#include <GL/glut.h>
#include <glm/glm.hpp>
#include <glm/gtc/type_ptr.hpp>
#include <vector>
std::vector< glm::vec2 > pts;
bool leftHeld = true;
glm::vec2* dragPt = NULL;
void mouse( int button, int state, int x, int y )
{
glm::vec2 pt( x, glutGet( GLUT_WINDOW_HEIGHT ) - y );
// left mouse button starts dragging a point
dragPt = NULL;
leftHeld = false;
if( button == GLUT_LEFT_BUTTON && state == GLUT_DOWN )
{
leftHeld = true;
size_t minIdx = 0;
for( size_t i = 0; i < pts.size(); ++i )
{
float newDist = glm::distance( pt, pts[ i ] );
float oldDist = glm::distance( pt, pts[ minIdx ] );
if( newDist <= oldDist && newDist < 15.0f )
{
minIdx = i;
dragPt = &pts[ minIdx ];
}
}
}
// middle mouse button clears all points
if( button == GLUT_MIDDLE_BUTTON && state == GLUT_UP )
{
pts.clear();
}
// right mouse button adds a point
if( button == GLUT_RIGHT_BUTTON && state == GLUT_UP )
{
pts.push_back( pt );
}
glutPostRedisplay();
}
void motion( int x, int y )
{
glm::vec2 pt( x, glutGet( GLUT_WINDOW_HEIGHT ) - y );
if( dragPt && leftHeld )
{
*dragPt = pt;
glutPostRedisplay();
}
}
void glLine( const std::vector< glm::vec2 >& line, GLenum mode )
{
glBegin( mode );
for( size_t i = 0; i < line.size(); ++i )
{
glVertex2f( line[i].x, line[i].y );
}
glEnd();
}
void display()
{
glClear( GL_COLOR_BUFFER_BIT | GL_DEPTH_BUFFER_BIT );
glMatrixMode( GL_PROJECTION );
glLoadIdentity();
double w = glutGet( GLUT_WINDOW_WIDTH );
double h = glutGet( GLUT_WINDOW_HEIGHT );
glOrtho( 0, w, 0, h, -1, 1 );
glMatrixMode( GL_MODELVIEW );
glLoadIdentity();
// draw polygon
glClear( GL_STENCIL_BUFFER_BIT );
{
// fill stencil buffer
glEnable( GL_STENCIL_TEST );
glColorMask( GL_FALSE,GL_FALSE,GL_FALSE,GL_FALSE );
glStencilOp( GL_KEEP, GL_KEEP, GL_INVERT );
glStencilFunc( GL_ALWAYS, 0x1, 0x1 );
glBegin( GL_TRIANGLES );
for( size_t i = 1; i+1 < pts.size(); ++i )
{
glVertex2fv( glm::value_ptr( pts[ 0 ] ) );
glVertex2fv( glm::value_ptr( pts[ i ] ) );
glVertex2fv( glm::value_ptr( pts[ i+1 ] ) );
}
glEnd();
// fill color buffer
glColor3ub( 0, 128, 0 );
glColorMask( GL_TRUE, GL_TRUE, GL_TRUE, GL_TRUE );
glStencilFunc( GL_EQUAL, 0x1, 0x1 );
glBegin( GL_TRIANGLES );
for( size_t i = 1; i+1 < pts.size(); ++i )
{
glVertex2fv( glm::value_ptr( pts[ 0 ] ) );
glVertex2fv( glm::value_ptr( pts[ i ] ) );
glVertex2fv( glm::value_ptr( pts[ i+1 ] ) );
}
glEnd();
glDisable( GL_STENCIL_TEST );
}
// draw polygon boundary
glLineWidth( 1 );
glColor3ub( 255, 255, 255 );
glLine( pts, GL_LINE_LOOP );
// draw vertexes
glPointSize( 9 );
glColor3ub( 255, 0, 0 );
glLine( pts, GL_POINTS );
glutSwapBuffers();
}
int main( int argc, char **argv )
{
glutInit( &argc, argv );
glutInitDisplayMode( GLUT_RGBA | GLUT_DEPTH | GLUT_DOUBLE | GLUT_STENCIL );
glutInitWindowSize( 640, 480 );
glutCreateWindow( "GLUT" );
glutMouseFunc( mouse );
glutMotionFunc( motion );
glutDisplayFunc( display );
glutMainLoop();
return 0;
}
这篇关于使用GL_STENCIL_TEST渲染凹面的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持！