Is anyone using/considering using simple bitmap scaling from a smaller
rendered image to accelerate walking/flying through worlds? Presumably
rendering fewer points would reduce the render processing pretty much
linearly with number of pixels. When the user stopped moving, the scene
could be rendered at full resolution for clarity.
For any given window size, allow the user to select a 'quality vs. frame
rate factor', maybe from 1 to 5. This would specify the size of the image
to be rendered prior to scaling to full window size. '1' means 'render
at full window size', 2 - render at 1/2 the height/width, 3 - one third,
etc.. The choice of integral scaling factors should make scale-up to full
window size clean and fast.
For example, I might want a near-full-screen VRML browsing window --
perhaps 900x700. I might find that a speed setting of 3 gives me an
adequate 300x234 rendering resolution, so that I don't really notice
much quality difference as I move around.
It might also be feasible to consider variations on this. For example,
if you stand in one place and turn, a QT-VR-like background texture
might be scrolled at full resolution quality. Or if you are manipulating
an object over a fixed backdrop -- say you're examining a coffee maker in
a Home Shopping Web page -- the backdrop could remain at full resolution.
Behavior generated animation might use this technique for a sub-window.
-Tom Craver-
Cable Products Division
Intel Corporation
eMail: [email protected]
All opinions expressed are my own, and do not represent those of
Intel Corporation.