okay, this is kinda off topic, but, what IS FSAA, and does it really mean full SCENE or SCREEN. I’m sure I used to remember it being full screen, but then I might be thinking of another antialiasing technique.
what I take to be FSAA is camera jittering. You render the sceen mutliple times with the camera perturbed within a circle describing the resolution of the screen. Each image is accumulated together.
Alternatively, you can anti-alias by doing this kind of thing in hardware at the pixel level by having multiple buckets per pixel. So a point not only projects to a pixel, but projects to a bucket within a pixel (and then, when it comes to rastering, the final colour of the pixel is the average of all its buckets).
So, which is FSAA? Since I thought it was the first approach, then I’d argue that it really IS full SCREEN antialiasing, for the following reason: since you’re jittering the camera, the entire viewport is offset. However, there is nothing to say that the entire scene must be sent to OpenGL while accumulating/jittering the images. (You could easily isolate part of the scene and render that without jittering the camera, and thus the entire scene is NOT antialiased).
What’s the story?