Hi and happy new year,
I use a float pixel buffer and render it in a texture. But I also need to add and substract in that buffer. Unfortunately , blending is not allowed (I read it somewhere) for float pixel buffer
So the pb is : How to add and substract in a float pixel buffer (I use an ATI card 9800))?
You can’t blend in floating point surfaces, because there is no hardware that supports it. The work-around, which works unless you need self-blending within a specific mesh, is to map the previous render target to a texture, and add the value of that texture when rendering the next pass into a new float buffer.
Double-buffered pbuffers are said to make this fairly easy, assuming render-to-texture is supported. And, sad to say, DirectX makes this quite easy. We’re all waiting for superbuffers in OpenGL…