[Algorithms] (no subject)
Brought to you by:
vexxed72
From: Sander v. R. <mv_...@ho...> - 2003-07-14 09:11:45
|
>I think a pretty good solution was posted by Per >http://www.flipcode.com/cgi-bin/msg.cgi?sh= >owThread=3D00008625&forum=3D3dtheory&id=3D-1 > >You started that thread, in fact. Is there a reason you don't want to use >his linear programming >method, or his make-a-polygon method? Maybe you could elaborate on what >you're looking for... Sorry for the crosspost. I'm just experimenting with some math, for fun, to see if i can make building brushes and CSG more precise (less fp errors) and more efficient Tough precision is more important to me than efficiency, since this is for preprocessing. Right now i'm working an algorithm that builds a brush by calculating the intersection points of 3 planes together with edge connectivity (well actually plane connectivity), and use that to build my polygons. Basically i store my vertices as a matrix (which contains the three planes), and i never have to clip my polygons. Since all vertices are always calculated with the exact same process and with the exact same number of multiplications (no divisions), numerical errors are at the very least more predictable.. I haven't put much tought into the csg part yet, but i have a couple of ideas.. I need a plane/brush intersection determination algo to determine if during a csg operation a plane actually intersects a brush.. _________________________________________________________________ Add photos to your e-mail with MSN 8. Get 2 months FREE*. http://join.msn.com/?page=features/featuredemail |