I would like some advise from you guys about geometry culling. I''m trying to decide which method i should learn. BSP,Octrees,Quadtrees etc. Which one do you guys think is the best and easiest to learn for a newbie on this subject. PS: I think that Geometry culling is the right word but i''m not sure, so if you have no idea what the hell i rambling about let me now

Anyone?

Nobody?

I guess this guestion doesn''t belong in
the OpenGL section. Ok, i''ll post it
in the Game Programming section instead.

Uh, make that, Graphics programming and theory.

Umm, it depends on what you want to do. As with all things, there are tradeoffs to each method. My preference goes with Octrees, but thats just because thats what I''m working with right now. =) Bsp trees are generally the fastest. They are best suited to static geometry which, for most games isn''t much of a problem. Bsp trees also provide perfect front to back sorting so thats a little benefit. Quake 1 - 3 and Doom used them for culling so that''s proof that they are very effective. Octrees tend to be a little more conservative, but that depends on how you build them. They don''t provide sorting at all though. It seems like most people are using octrees these days anyway. Quadtrees are best suited for use with landscapes because they are a 2-d data structure and can be used very well to partition height maps. Well, that''s a general rundown of the algorithms you mentioned. There are also many occlusion and image-based methods that are becoming more popular as hardware gets better. A general trend these days is towards more conservative visibility methods because video cards can handle more tris and the time it takes you to figure out which polygons are visible is longer than it takes to draw them. Hope I helped a little. See ya and good luck and God-bless.

God Bless America.

I quess i should try them all and see which
one is easier and works best for me.

Anyway, thank you for your response!