JOGL Shader Questions

OpenGL Development
#1

I’m trying to figure out how to start using GLSL with JOGL to do pixel shading on images drawn to GLAutoDrawable via glDrawPixels.

I’m confused as to how a few things work in GLSL:

  1. How can I access the color of the current pixel in the shader call? What is the form of that pixel (RGB, BGR, etc.) so that I can unpack it?
  2. Do I just assign gl_FragColor to change the pixel?
  3. How do I get coordinates of the current pixel?
  4. Does it generally make more sense to perform lighting calculations in the GLSL frag shader or maybe in OpenCL?

I’ve trying really hard to muddle through learning JOGL by myself, but I’m really stuck on these fundamental problems with GLSL, and I haven’t been able to find consistent help through Google.

Any help is appreciated, and I thank you in advance for your time.

#2

  1. The “current pixel” is not known. If you want post-processing, then typically you would render the screen to an FBO texture, and then in your post shader sample from the texture in order to “get the screen pixel.”

  2. Yep; unless you are using GLSL 130+ in which case you would use a custom out attribute.

  3. As gouessej said, gl_FragCoord.xy will do the trick. However, it gives you the screen coordinates, not the coordinates of, say, your quad. If you’re trying to (for example) make a horizontal gradient from 0.0 to 1.0, you would be better off using texture coordinates than relying on gl_FragCoord.

  4. OpenCL is tricky to interact with OpenGL/GLSL and probably won’t give you much of an improvement, unless your lighting calculations are extremely expensive.

It sounds like you are having trouble understanding the role of fragment and vertex shaders. You can read up on them here: