Originally posted by jwatte:
[b]You have to render to framebuffer, make framebuffer into texture, and then render into new/re-used framebuffer, using the texture as one input.
This means that individual triangles within the same geometry won’t blend with each other.[/b]
So you mean I should use the first rendered frame buffer contents as a texture input and other data as another texture input and blend them before the blending part in the pipeline using glTexEnv or something similar?