Yes it's not the carving , it's the PVS sets. I can see regions of the map that DO NOT have any point to point relation
(i.e.: two rooms separated by a solid wall.) VIS does the PVS stuffing and that is where the problem is. When I walk the map with gl_wireframe 1 the BSP carved leafs look fine. Using hint brushes to carve them differently would not make things any better. Again it's not the carving it's the PVS sets created for each leaf. In many cases I would like to fiddle with the PVS lists to limit far away leafs. -----Original Message----- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Jeffrey "botman" Broome Sent: Friday, January 09, 2004 11:34 AM To: [EMAIL PROTECTED] Subject: Re: [hlcoders] VIS issues K. Mike Bradley wrote: > I have had an issue with vis for 2 years now. It doesn't work very > efficient. Maybe, but it would be difficult to programmatically do much better. > It has nothing to do with hint brushes or how BSP carves up the vis leafs. Sure it does. That's what hint brushes are for, to allow the level designer to give hints to the compile tools about how the world should be carved up. Do you understand HOW the PVS of nodes are built? If ANY point in one node can see ANY point in another node then ALL points in the first node are considered visible from ALL points in the second node. That's just the way the visibility algorithm works. See Michael Abrash's chapter on "Quake's Visible-Surface Determination" (chapter 64) to understand why this works the way it does... http://public.planetmirror.com/pub/gpbb/ -- Jeffrey "botman" Broome _______________________________________________ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlcoders _______________________________________________ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlcoders
