What is a standard reason for this behavior? I have 2 the same buffers with 2 textures of mesh with premultiplied alpha and alpha background: in the first I have object on background, but the second leaves object trail behind itself when I move it on this alpha background. I can't realise why! FBO's are similar.
Thanks!
My mistake was I forgot to bind FBO before clearing:
glClearColor(0.0f, 0.0f, 0.0f, 1.0f);
glClear(GL_COLOR_BUFFER_BIT | GL_DEPTH_BUFFER_BIT);
glBindFramebuffer(GL_FRAMEBUFFER, FBO);
glBindFramebuffer(GL_FRAMEBUFFER, 0); // WRONG!
glBindFramebuffer(GL_FRAMEBUFFER, FBO);
glClearColor(0.0f, 0.0f, 0.0f, 1.0f);
glClear(GL_COLOR_BUFFER_BIT | GL_DEPTH_BUFFER_BIT)
glBindFramebuffer(GL_FRAMEBUFFER, 0); // RIGHT!
Related
I am studying OpenGL at Learn-OpenGL.
Here I learned about skybox and framebuffer.
So I can see that I can blur the skybox by capturing it in the framebuffer.
I want to see a screen like this image:
But, when rendering with this framebuffer, the object is overwritten and only the skybox is visible.
this is my code:
glEnable(GL_DEPTH_TEST);
glEnable(GL_BLEND);
glBindFramebuffer(GL_FRAMEBUFFER, skybox_framebuffer->get_fbo());
{
// skybox capture
glViewport(0, 0, skybox_framebuffer->get_width(), skybox_framebuffer->get_height());
glClearColor(0.0f, 0.0f, 0.0f, 1.0f);
glClear(GL_COLOR_BUFFER_BIT | GL_DEPTH_BUFFER_BIT);
skybox.draw(camera, projection);
}
glBindFramebuffer(GL_FRAMEBUFFER, 0);
{
glViewport(0, 0, g_window.get_width(), g_window.get_height());
glClearColor(0.0f, 0.0f, 0.0f, 1.0f);
glClear(GL_COLOR_BUFFER_BIT | GL_DEPTH_BUFFER_BIT);
skybox_framebuffer->draw();
draw_model(ourModel, ourShader, view, projection);
}
Even changing the drawing order gave the same result:
.
And, if I draw the SKYBOX directly instead of the framebuffer, I get this result:
.
Solution1
I think the solution is to use the stencil buffer.
my solution code:
glBindFramebuffer(GL_FRAMEBUFFER, 0);
{
glViewport(0, 0, g_window.get_width(), g_window.get_height());
glEnable(GL_DEPTH_TEST);
glEnable(GL_STENCIL_TEST);
glStencilOp(GL_KEEP, GL_KEEP, GL_REPLACE);
glClearColor(0.0f, 0.0f, 0.0f, 1.0f);
glClear(GL_COLOR_BUFFER_BIT | GL_DEPTH_BUFFER_BIT | GL_STENCIL_BUFFER_BIT);
// draw model
glStencilFunc(GL_ALWAYS, 1, 0xFF);
glStencilMask(0xFF);
draw_model(ourModel, ourShader, view, projection);
// draw skybox framebuffer
glStencilFunc(GL_NOTEQUAL, 1, 0xFF);
glStencilMask(0x00); // disable writing to the stencil buffer
glDisable(GL_DEPTH_TEST);
skybox_framebuffer->draw();
glStencilMask(0xFF);
glStencilFunc(GL_ALWAYS, 1, 0xFF);
glEnable(GL_DEPTH_TEST);
}
and my solution result:
Solution2
using glDepthMask
This shows the same screen as solution1 without the stencil buffer.
glBindFramebuffer(GL_FRAMEBUFFER, 0);
{
glViewport(0, 0, g_window.get_width(), g_window.get_height());
glEnable(GL_DEPTH_TEST);
glClearColor(0.0f, 0.0f, 0.0f, 1.0f);
glClear(GL_COLOR_BUFFER_BIT | GL_DEPTH_BUFFER_BIT);
glDepthMask(GL_FALSE);
skybox_framebuffer->draw();
glDepthMask(GL_TRUE);
draw_model(ourModel, ourShader, view, projection);
}
question
I could implement it like this,
Is there any way to blur the skybox without using the framebuffer, other than using a blurred image?
Question:
I am trying to map a single 2D image to a 3D texture. When I visualize the output, OpenGL renders a white texture floating in black background. Why is this happening? Is it possible to map a 2D image to a 3D texture?
My goal is to place multiple 2D images along the z-dimension to resemble a volume. An example of such an approach is in this video: http://cvlab.epfl.ch/research/medical/em/synapses
Approaches attempted:
I took 2D images and mapped them to 2D textures. I have also used 2D texture arrays. Both of these approaches provide a good result. However, when I tried to paint the surface of a 3D texture with the same image, I get a white texture floating around in a black space.
Code
I have followed the texture mapping tutorial that is available on NeHe's website.
http://nehe.gamedev.net/tutorial/lesson_06_texturing_update/47002/
Except changes to these three functions, everything else in the program (present in the solution file in the website) is the same.
... // header file declarations
#include <glext.h>
PFNGLTEXIMAGE3DPROC glTexImage3D;
GLuint texture;
int LoadGLTextures() // Load Bitmaps And Convert To Textures
{
/* load an image file directly as a new OpenGL texture */
texture = SOIL_load_OGL_texture("Data/NeHe.bmp",SOIL_LOAD_AUTO,
SOIL_CREATE_NEW_ID,S OIL_FLAG_INVERT_Y);
// "Data/NeHe.bmp"
if(texture == 0) {return false;}
glTexImage3D = (PFNGLTEXIMAGE3DPROC) wglGetProcAddress("glTexImage3D");
if (glTexImage3D == NULL)
{
printf("Error in line %d: Couldn't load glTexImage3D function. Aborting.\n", __LINE__);
return -1;
}
glBindTexture(GL_TEXTURE_3D, texture);
glTexParameteri(GL_TEXTURE_3D,GL_TEXTURE_MIN_FILTER,GL_LINEAR);
glTexParameteri(GL_TEXTURE_3D,GL_TEXTURE_MAG_FILTER,GL_LINEAR);
//glTexImage3D(GL_TEXTURE_3D, 0, GL_RGBA, 4096, 4096, 1, 0, GL_RGBA, GL_UNSIGNED_BYTE, NULL);
return true;
}
int InitGL(GLvoid)
{
if (!LoadGLTextures())
{ return FALSE; }
glEnable(GL_TEXTURE_3D);
// glEnable(GL_TEXTURE_2D);
glShadeModel(GL_SMOOTH); //
glClearColor(0.0f, 0.0f, 0.0f, 0.5f); // Black Background
glClearDepth(1.0f); // Depth Buffer Setup
glEnable(GL_DEPTH_TEST);
glDepthFunc(GL_LEQUAL);
glHint(GL_PERSPECTIVE_CORRECTION_HINT, GL_NICEST); // Really Nice Perspective calculations
return TRUE; // Initialization Went OK
}
int DrawGLScene(GLvoid)
{
glClear(GL_COLOR_BUFFER_BIT | GL_DEPTH_BUFFER_BIT);
glLoadIdentity();
glTranslatef(0.0f,0.0f,-5.0f);
glRotatef(xrot,1.0f,0.0f,0.0f);
glRotatef(yrot,0.0f,1.0f,0.0f);
glRotatef(zrot,0.0f,0.0f,1.0f);
// glBindTexture(GL_TEXTURE_2D, texture[0]);
glBindTexture(GL_TEXTURE_3D, texture);
glBegin(GL_QUADS);
glTexCoord3f(0.0f, 0.0f, 0.0f); glVertex3f(-1.0f, -1.0f, 0.0f);
glTexCoord3f(1.0f, 0.0f, 0.0f); glVertex3f( 1.0f, -1.0f, 0.0f);
glTexCoord3f(1.0f, 1.0f, 0.0f); glVertex3f( 1.0f, 1.0f, 0.0f);
glTexCoord3f(0.0f, 1.0f, 0.0f); glVertex3f(-1.0f, 1.0f, 0.0f);
glEnd();
xrot+=0.3f;
yrot+=0.2f;
zrot+=0.4f;
return TRUE;
}
...
It is better to use 3D textures instead of 2D ones as you will be able to apply rotations correctly using glRotate(). You will also need to do blending and/or alpha-testing to see different layers of your 2D image.
To create a 3D texture, you can create an array of dimension [width x height x number_of_slices] and then store the raw data slice by slice
Look at this tutorial: http://www.codeproject.com/Articles/352270/Getting-started-with-Volume-Rendering?msg=4729498#xx4729498xx
It's very good and it worked for me to do the same thing
I'm rendering textures on the two surfaces which actually have the same location and position.
In these kind of cases I want to display only one of the textures, while I get following
I.e. I want only the texture of first material to be displayed.
So I would like to know where I've to search for solution, should I play with the blending of the materials ?
This is a pretty old issue (z-fighting) as the card isn't sure which object to draw in front. The linked Wikipedia article has more examples of this.
To fix this, increase the depth buffer's accuracy (bit depth), reduce the depth of your view (distance far/near clip plane) or add a tiny offset so the coordinates are no longer exactly (or almost) the same. You could as well simply disable the depth buffer (or clear it) for rendering this (in which case whatever is rendered last will overlap everything else).
You mean you want to display few textures at one time? Use multi texturing then. Attach both textures to one Quad using:
glMultiTexCoord2fARB(GL_TEXTURE0_ARB, 0.0f, 0.0f);
glMultiTexCoord2fARB(GL_TEXTURE1_ARB, 0.0f, 0.0f);
And play with glTexEnvi to combine them as you want. If you want to show only one of your textures then draw two quads with separate textures on each of them. And quad witch will be drawn later will be visible. Then if you want to turn from one to another, make separate -(void)'s for each of quad, with glClearColor before each quad drawing and then make button to call for example -(void)quad1 and button to call -(void)quad2.
Or just use depth buffer.
At your situation i would use multi texturing, because all buffers looked too hard for me all time.
I will show how would my code look like.
Code using multi texture:
glClear(GL_COLOR_BUFFER_BIT);
//Turn on blending and set glBlendFunc for your needs
glEnable(GL_BLEND);
glBlendFunc(GL_SRC_ALPHA, GL_ONE_MINUS_SRC_ALPHA);
//Describing multitexturing
glClientActiveTexture(GL_TEXTURE0_ARB); //Activating texture on unit 0
glActiveTexture(GL_TEXTURE0_ARB); //Activating texture on unit 0
glBindTexture(GL_TEXTURE_2D, Texture1); //Bind texture on unit 0
glEnable(GL_TEXTURE_2D); //Enable GL_TEXTURE_2D to set it's glTexEnvi for your needs
glTexEnvi( GL_TEXTURE_ENV, GL_TEXTURE_ENV_MODE, GL_COMBINE ); //Edit glTexEnvi for your needs (link bellow code)
glTexEnvi( GL_TEXTURE_ENV, GL_COMBINE_ALPHA, GL_INTERPOLATE );
glClientActiveTexture(GL_TEXTURE1_ARB); //Activating texture on unit 1
glActiveTexture(GL_TEXTURE1_ARB); //Activating texture on unit 1
glBindTexture(GL_TEXTURE_2D, Texture2); //Bind texture on unit 0
glEnable(GL_TEXTURE_2D); //Enable GL_TEXTURE_2D to set it's glTexEnvi for your needs
glTexEnvi( GL_TEXTURE_ENV, GL_TEXTURE_ENV_MODE, GL_COMBINE ); //Edit glTexEnvi for your needs (link bellow code)
glTexEnvi( GL_TEXTURE_ENV, GL_COMBINE_ALPHA, GL_REPLACE );
//Drawing quad with multi texture
glBegin(GL_QUADS);
glMultiTexCoord2fARB(GL_TEXTURE0_ARB, 0.0f, 0.0f); //Texture on unit 0 coords
glMultiTexCoord2fARB(GL_TEXTURE1_ARB, 0.0f, 0.0f); //Texture on unit 1 coords
glVertex2f(0.0, 500);
glMultiTexCoord2fARB(GL_TEXTURE0_ARB, 0.0f, 1.0f); //Texture on unit 0 coords
glMultiTexCoord2fARB(GL_TEXTURE1_ARB, 0.0f, 1.0f); //Texture on unit 1 coords
glVertex2f(0.0, 0.0);
glMultiTexCoord2fARB(GL_TEXTURE0_ARB, 1.0f, 1.0f); //Texture on unit 0 coords
glMultiTexCoord2fARB(GL_TEXTURE1_ARB, 1.0f, 1.0f); //Texture on unit 1 coords
glVertex2f(800, 0.0);
glMultiTexCoord2fARB(GL_TEXTURE0_ARB, 1.0f, 0.0f); //Texture on unit 0 coords
glMultiTexCoord2fARB(GL_TEXTURE1_ARB, 1.0f, 0.0f); //Texture on unit 1 coords
glVertex2f(800, 500);
glEnd();
//Disable GL_TEXTURE_2D. We don't need it anymore
glDisable(GL_TEXTURE_2D);
//Disable blending
glDisable(GL_BLEND);
//Flush everything
glFlush();
Code for using one quad for each texture
-(void) drawQuad1
{
glClear(GL_COLOR_BUFFER_BIT);
glEnable(GL_TEXTURE_2D);
glBindTexture(GL_TEXTURE_2D, Texture1);
glBegin(GL_QUADS);
glTexCoord2f(0.0f, 0.0f); glVertex2f(0.0, 500);
glTexCoord2f(0.0f, 1.0f); glVertex2f(0.0, 0.0);
glTexCoord2f(1.0f, 1.0f); glVertex2f(800, 0.0);
glTexCoord2f(1.0f, 0.0f); glVertex2f(800, 500);
glEnd();
glDisable(GL_TEXTURE_2D)
glFlush();
}
-(void) drawQuad2
{
glClear(GL_COLOR_BUFFER_BIT);
glEnable(GL_TEXTURE_2D);
glBindTexture(GL_TEXTURE_2D, Texture2);
glBegin(GL_QUADS);
glTexCoord2f(0.0f, 0.0f); glVertex2f(0.0, 500);
glTexCoord2f(0.0f, 1.0f); glVertex2f(0.0, 0.0);
glTexCoord2f(1.0f, 1.0f); glVertex2f(800, 0.0);
glTexCoord2f(1.0f, 0.0f); glVertex2f(800, 500);
glEnd();
glDisable(GL_TEXTURE_2D)
glFlush();
}
//and now when you want to draw `drawQuad1` you need to do `[self drawQuad1];`
//and where you want to draw `drawQuad2` you need to do `[self drawQuad2];`
Info about glTexEnvi here. Its just sample code. Fast written so could have mistakes. Sorry if i forgot something. Just i haven't been doing it for long time. Written it on Mac OS X, so if you are using other platform maby you will have to change few things (not opengl). And about depth buffer? I don't know how to use it. So I can't explain it for you. Sorry. Never used them.
I've begun switching my rendering code to support shaders, and that all works fine when rendering to the back buffer. So now I'm working towards rendering to FBOs, but all I get are white textures for both the color and normals.
Here is my FBO creation code:
void RenderTarget_GL::CreateFBO (void)
{
// if the machine supports the GL FBO extension
if (s_supportfbo)
{
// Create FBO
glGenFramebuffersEXT(1, &m_fbo);
glBindFramebufferEXT(GL_FRAMEBUFFER_EXT, m_fbo);
// Create default texture buffer
char *buffer = new char [static_cast<int>(g_window->GetWidth() * m_screenWidth) * static_cast<int>(g_window->GetHeight() * m_screenHeight) * 4];
std::memset(buffer, 0, static_cast<int>(g_window->GetWidth() * m_screenWidth) * static_cast<int>(g_window->GetHeight() * m_screenHeight) * 4);
// Create Render Texture
glGenTextures(1, &m_rendertexture);
glBindTexture(GL_TEXTURE_2D, m_rendertexture);
glTexImage2D(GL_TEXTURE_2D, 0, 4, static_cast<int>(g_window->GetWidth() * m_screenWidth), static_cast<int>(g_window->GetHeight() * m_screenHeight), 0, GL_RGBA, GL_UNSIGNED_BYTE, buffer);
// Bind Render Texture to FBO
glBindTexture(GL_TEXTURE_2D, 0);
glFramebufferTexture2DEXT(GL_FRAMEBUFFER_EXT, GL_COLOR_ATTACHMENT0_EXT, GL_TEXTURE_2D, m_rendertexture, 0);
// Create Normal Texture if this FBO will be rendering normals
if (m_hasnormal)
{
glGenTextures(1, &m_normaltexture);
glTexImage2D(GL_TEXTURE_2D, 0, 4, static_cast<int>(g_window->GetWidth() * m_screenWidth), static_cast<int>(g_window->GetHeight() * m_screenHeight), 0, GL_RGBA, GL_UNSIGNED_BYTE, buffer);
// Bind Normal Texture to FBO
glFramebufferTexture2DEXT(GL_FRAMEBUFFER_EXT, GL_COLOR_ATTACHMENT1_EXT, GL_TEXTURE_2D, m_normaltexture, 0);
}
// UnBind FBO and cleanup default buffer
delete [] buffer;
Clear();
}
}
And the code I use to set the current render target:
void RenderTarget_GL::Set (void)
{
if (s_supportfbo && g_glgraphics->GetShaderEnabled())
{
static const GLenum buffer1[] = {GL_COLOR_ATTACHMENT0_EXT};
static const GLenum buffer2[] = {GL_COLOR_ATTACHMENT0_EXT, GL_COLOR_ATTACHMENT1_EXT};
glBindFramebufferEXT(GL_FRAMEBUFFER_EXT, m_fbo);
if (m_hasnormal)
glDrawBuffers(2, buffer2);
else
glDrawBuffers(1, buffer1);
}
}
And finally, my actual drawing code:
glClear(GL_COLOR_BUFFER_BIT | GL_DEPTH_BUFFER_BIT | GL_STENCIL_BUFFER_BIT);
// Setup the camera transformation
glMatrixMode(GL_MODELVIEW);
glPushMatrix();
if (m_camera)
m_camera->GLMatrix();
else
m_defaultCam.GLMatrix();
// Setup Render Target
if (m_shaderenabled)
{
glPushAttrib(GL_VIEWPORT_BIT);
glViewport(0,0,g_window->GetWidth(),g_window->GetHeight());
m_initialpass->Set();
}
// Draw All Objects with their per-object shaders
// Clear render target and shader bindings
if (m_shaderenabled)
{
glPopAttrib();
RenderTarget_GL::Clear();
Shader_GL::ClearShaderBinding();
}
glPopMatrix();
glMatrixMode(GL_PROJECTION);
glPushMatrix();
glLoadIdentity();
// Draw Scene
glActiveTexture(GL_TEXTURE0);
glBindTexture(GL_TEXTURE_2D, m_initialpass->GetColorTexture());
glBegin(GL_QUADS);
glTexCoord2f(0.0f, 0.0f);
glVertex3f(-1.0f, -1.0f, 0.0f);
glTexCoord2f(1.0f, 0.0f);
glVertex3f(1.0f, -1.0f, 0.0f);
glTexCoord2f(1.0f, 1.0f);
glVertex3f(1.0f, 1.0f, 0.0f);
glTexCoord2f(0.0f, 1.0f);
glVertex3f(-1.0f, 1.0f, 0.0f);
glEnd();
Texture_GL::ClearTextureBinding();
glPopMatrix();
// Swap Buffers
GL_TEXTURE_MIN_FILTER is GL_NEAREST_MIPMAP_LINEAR by default. Supply mipmaps or switch to GL_LINEAR or GL_NEAREST.
The OpenGL Wiki has more.
I'm learning to use OpenGL in my free time, and I'm working on deferred rendering right now. I've been stuck on this problem for a few weeks. When I bind my diffuse texture to GL_TEXTURE0, each pixel in the texture is (0.0f, 0.0f, 0.0f, 0.0f). When I switch it to GL_TEXTURE3 it works. Where should I be looking to debug this problem?
I will post the draw call for the rendering below.
The entire source code is in this repository:
https://github.com/DanielSmithMichigan/OpenGLExperimentation
The act of writing and reading from different textures happens in here:
https://github.com/DanielSmithMichigan/OpenGLExperimentation/blob/master/OpenGlTest6/Defer.h
The draw call:
glClear(GL_COLOR_BUFFER_BIT | GL_DEPTH_BUFFER_BIT);
// if I switch GL_TEXTURE0 to GL_TEXTURE3
// and diffuseUniform << 0 to diffuseUniform << 3
// then it will work
glActiveTexture(GL_TEXTURE0);
glBindTexture(GL_TEXTURE_2D, diffuseTexture);
*diffuseUniform << 0;
glActiveTexture(GL_TEXTURE1);
glBindTexture(GL_TEXTURE_2D, positionTexture);
*positionUniform << 1;
glActiveTexture(GL_TEXTURE2);
glBindTexture(GL_TEXTURE_2D, normalsTexture);
*normalsUniform << 2;
*ambientColorUniform << objects->ambientLight->color;
*lightColorUniform << objects->sun->color;
*lightDirectionUniform << objects->sun->direction;
*cameraPositionUniform << objects->camera->position;
*specularColorUniform << objects->sun->color;
*specularPowerUniform << 5.0f;
glBegin(GL_QUADS);
glTexCoord2f(0, 0);
glVertex3f(0.0f, 0.0f, 0.0f);
glTexCoord2f(1, 0);
glVertex3f((float)width, 0.0f, 0.0f);
glTexCoord2f(1, 1);
glVertex3f((float)width, (float)height, 0.0f);
glTexCoord2f(0, 1);
glVertex3f(0.0f, (float)height, 0.0f);
glEnd();
I was trying to say GL_TEXTURE1 with uniform = 1
Because I doubted a texture you made before the diffuse texture using glGenTextures(1, &depthBuffer) but I don't know why you made its name depthBuffer depthbuffer must be one of RBOs not a texture.
Anyway,
the default value of glactivetexture() is GL_TEXTURE0
You didn't call glactivetexture(GL_TEXTURE0) before you called glbindtexture(depthBuffer) although because the default value is GL_TEXTURE0, you are using GL_TEXTURE0 already for the depthBuffer .
The issue was that in the files Mesh.h and Grid.h, I was binding a sampler to index 0 (GL_TEXTURE0). This sampler was still bound by the time I was trying to bind/read my diffuse texture on index 0.
The solution was to use glActiveTexture(GL_TEXTURE0) and glEnable(GL_TEXTURE_2D) before I started my geometry pass, and the opposite glActiveTexture(GL_TEXTURE0) and glDisable(GL_TEXTURE_2D) after it ended.
Then in the draw call for my deferred rendering object, I use glBindSampler(0, 0).
As far as I can tell, all 5 functions mentioned above were necessary.
The fix for this problem is in commit hash d60fec483e896971304f5f9f24c4bd0f10fe48fe for anyone who needs more information.