RE: [Algorithms] Terrain performance comparisons
Brought to you by:
vexxed72
From: Charles B. <cb...@cb...> - 2003-07-27 21:26:28
|
At 09:03 PM 7/25/2003 +0100, you wrote: >just that simulation time always move forward by a constant increment. Real >life does, games might as well! Um, are you saying that your real life is a series of snapshots with a constant time interval between them? That's a rather bizarre thing to say. Other than that, I definitely believe that optimizations based on coherence of viewer *position* are much better than those based on viewer *direction*. The old ROAM and view-dependent progressive meshes used view direction, which is pretty much junk for most games because you can turn as much as 180 in a single frame in some cases. Position is usually reasonably coherent, set by the maximum velocity at the largest time step, in most games this means somewhere around 1 meter max change per frame, usually less. -------------------------------------------------------------------------------------------- Charles Bloom email "cb" http://www.cbloom.com |