This topic is now archived and is closed to further replies.


Optimal Terrain Rendering

Recommended Posts

I''m currently storing my entire terrain in a single VB (+/- 200,000 polygons). Then, when I render, I am selecting strips of the VB and rendering them with the drawprimitive command. So far so good. This method means that I can draw 10 strips of 10 ''squares'' to cover 10x10 ground. If I can increase the strip width to 100, I could in theory cover 100x100 with just 10 calls as well. But now, I''m wondering, would it be more efficient to switch from strips in a single VB to using squares split over multiple VB''s? I could easilly incorporate a ROAM type of system if I went this route (where a VB gets loaded as you move around - much harder to do with a single VB), plus the LOD would be much simpler (squares further away use a low-detail VB, ones closer use a high-detail VB). But, how much would it impact on performance to split the terrain up into multiple VB''s? Also, with the LOD, would it be more efficient to store a high & low detail VB, or just the low-detail and n-patch it? Learning to fly is easy, but as a tortoise, the landings are really rough.

Share this post

Link to post
Share on other sites