I am in the process of integrating post processing effects into my 3D engine. I have hit a roadblock with capturing depth data using a FrameBuffer Object with WebGL. There has been no issue capturing color data into the FrameBuffer. However, I am not able to get any Depth data even with the Depth Extension enabled for Chrome/Firefox.
var DepthEXT = webGLContext.getExtension( "WEBKIT_WEBGL_depth_texture" ) ||
webGLContext.getExtension( "MOZ_WEBGL_depth_texture" );
I have tried many different settings to see if it was a configuration issue, but no matter what I try I just see a white texture. The screenshot at the end of the post shows the color attachment rendered as a texture and the depth attachment rendered as a texture. Is there a problem with my initialization of the FrameBuffer or should I be looking elsewhere to solve this issue?
Below is my code for initializing the FrameBuffer Object:
// Frame Buffer
this.m_frameBuffer = gl.createFramebuffer();
gl.bindFramebuffer( gl.FRAMEBUFFER, this.m_frameBuffer );
// Render Buffer
this.m_renderBuffer = gl.createRenderbuffer();
gl.bindRenderbuffer( gl.RENDERBUFFER, this.m_renderBuffer );
gl.renderbufferStorage(
gl.RENDERBUFFER,
gl.DEPTH_COMPONENT16,
sharedRenderer.canvasDOMElement.width,
sharedRenderer.canvasDOMElement.height );
// Diffuse Component
this.m_diffuseComponentTexture = gl.createTexture();
gl.bindTexture( gl.TEXTURE_2D, this.m_diffuseComponentTexture );
gl.texParameteri(
gl.TEXTURE_2D,
gl.TEXTURE_MAG_FILTER,
gl.LINEAR );
gl.texParameteri(
gl.TEXTURE_2D,
gl.TEXTURE_MIN_FILTER,
gl.LINEAR );
gl.texParameteri( gl.TEXTURE_2D, gl.TEXTURE_WRAP_S, gl.CLAMP_TO_EDGE);
gl.texParameteri( gl.TEXTURE_2D, gl.TEXTURE_WRAP_T, gl.CLAMP_TO_EDGE);
gl.texImage2D(
gl.TEXTURE_2D,
0,
gl.RGBA,
sharedRenderer.canvasDOMElement.width,
sharedRenderer.canvasDOMElement.height,
0,
gl.RGBA,
gl.UNSIGNED_BYTE,
null );
// Depth
this.m_depthComponentTexture = gl.createTexture();
gl.bindTexture( gl.TEXTURE_2D, this.m_depthComponentTexture );
gl.texParameteri(
gl.TEXTURE_2D,
gl.TEXTURE_MAG_FILTER,
gl.NEAREST );
gl.texParameteri(
gl.TEXTURE_2D,
gl.TEXTURE_MIN_FILTER,
gl.NEAREST );
gl.texParameteri( gl.TEXTURE_2D, gl.TEXTURE_WRAP_S, gl.CLAMP_TO_EDGE);
gl.texParameteri( gl.TEXTURE_2D, gl.TEXTURE_WRAP_T, gl.CLAMP_TO_EDGE);
gl.texImage2D(
gl.TEXTURE_2D,
0,
gl.DEPTH_COMPONENT,
sharedRenderer.canvasDOMElement.width,
sharedRenderer.canvasDOMElement.height,
0,
gl.DEPTH_COMPONENT,
gl.UNSIGNED_SHORT,
null );
// FrameBuffer
// Diffuse
gl.framebufferTexture2D(
gl.FRAMEBUFFER,
gl.COLOR_ATTACHMENT0,
gl.TEXTURE_2D,
this.m_diffuseComponentTexture,
0 );
// Depth
gl.framebufferTexture2D(
gl.FRAMEBUFFER,
gl.DEPTH_ATTACHMENT,
gl.TEXTURE_2D,
this.m_depthComponentTexture,
0 );
// RenderBuffer
gl.framebufferRenderbuffer(
gl.FRAMEBUFFER,
gl.DEPTH_ATTACHMENT,
gl.RENDERBUFFER,
this.m_renderBuffer );
// Unbind buffers and textures
gl.bindTexture( gl.TEXTURE_2D, null );
gl.bindRenderbuffer( gl.RENDERBUFFER, null );
gl.bindFramebuffer( gl.FRAMEBUFFER, null );
Here is the code where I render the current scene to the FrameBuffer.
CBRenderer.prototype.renderSceneToGBuffer = function( sceneToRender, GBufferTarget, deltaSeconds )
{
CBMatrixStack.clearMatrixStackAndPushIdentityMatrix();
this.applyProjectionMatrix();
GBufferTarget.bindGBufferFrameBuffer();
this.renderer.enable( this.renderer.DEPTH_TEST );
this.renderer.depthMask( true );
this.renderer.clearDepth( 1.0 );
this.renderer.clearColor( 0.1, 0.1, 0.1, 0.0 );
this.renderer.clear( this.renderer.COLOR_BUFFER_BIT | this.renderer.DEPTH_BUFFER_BIT );
sceneToRender.render( deltaSeconds );
this.renderer.flush();
GBufferTarget.m_dirty = false;
GBufferTarget.unbindGBufferFrameBuffer();
this.renderer.clearColor( 0.0, 0.0, 0.0, 0.0 );
this.renderer.clear( this.renderer.COLOR_BUFFER_BIT | this.renderer.DEPTH_BUFFER_BIT );
this.renderer.bindTexture( this.renderer.TEXTURE_2D, null );
}
You don't have the depth texture attached to your FBO at the end of setting it up. You do set the texture as the depth attachment at some point:
// Depth
gl.framebufferTexture2D(
gl.FRAMEBUFFER,
gl.DEPTH_ATTACHMENT,
gl.TEXTURE_2D,
this.m_depthComponentTexture,
0 );
But immediately following that, you set a renderbuffer as the depth attachment instead:
// RenderBuffer
gl.framebufferRenderbuffer(
gl.FRAMEBUFFER,
gl.DEPTH_ATTACHMENT,
gl.RENDERBUFFER,
this.m_renderBuffer );
You can have only one target attached to each FBO attachment point. So the second call sets a new depth attachment, replacing the one you just set in the first call. So at the end of this, m_depthComponentTexture
is not attached to the FBO anymore.
If you want to use a depth texture, I don't see why you would want to create a depth renderbuffer as well. You should only need one of the two. You would normally use a renderbuffer for the depth if you don't need it for sampling later. If you want to sample from the result, you need a texture instead, and there's no need for a renderbuffer anymore.