DTS vs. DIF - looking for pointer to thread
by Ray Depew · in Technical Issues · 03/28/2005 (4:06 pm) · 5 replies
I'm trying to tell somebody why DIF is preferred over DTS for buildings, and I'm failing miserably. I know this has been the topic of multiple threads; unfortunately, I'm also failing miserably at finding one of those threads.
Not trying to restart the discussion, just looking for a pointer to a thread. Thanks in advance for any help.
Not trying to restart the discussion, just looking for a pointer to a thread. Thanks in advance for any help.
#3
03/29/2005 (10:00 am)
I just added that answer to the FAQ. 
#4
Ah, so that's the reason... but since the collision meshes are convexes (just like brushes), wouldn't it be possible to parse them into a bsp, just like interiors do?
05/25/2005 (12:22 pm)
Quote:* Collision searches are linear...*lot* slower than bsp searches (roughly 15-30%)
Ah, so that's the reason... but since the collision meshes are convexes (just like brushes), wouldn't it be possible to parse them into a bsp, just like interiors do?
#5
05/25/2005 (2:47 pm)
Yes and no...with the current system where you can't animate collision meshes then there is nothing stopping you from doing this (in fact I have ;) but in the future it would be good to able to animate the collision meshes which'll mean you can't use a bsp. 
Associate Matt Fairfax
PopCap
* Only convex brushes so your visible mesh exactly matches your collision mesh
* Automatic collision mesh generation
* Collision meshes are parsed into a bsp structure for searching...very fast collision
* Self shadowing since every surface has a unique ligthmap
* Precalced lighting (potentially much nicer)
DTS:
* Arbitrary polygons - concave, convex, closed, or open
* Separate hand generated collision meshes that must be convex so rarely match visible gemometry
* Collision searches are linear...*lot* slower than bsp searches (roughly 15-30%)
* No lightmaps so no pre-calced lighting or self shadowing