Author Topic:   ROAM "wedgies" ?
New Member
posted January 14, 2000 04:40 AM            
Okay guys, the ROAM paper is losing me when they begin to talk about the wedgies.

Is a "wedgie" a ficticious triangle w/a pseudo-3rd dimension (height/depth) that covers many vertices / height-mapped points (i.e. is a large "coarse" LOD triangle), which are then used for splitting into the smaller triangles (thus creating newer, smaller wedgies)??

If not, can anyone take a stab at explaining them to me in a manner I might understand better?

BTW, when the triangles are split - does a ROAM implementation split down to a finer resolution than the vertices given by a pre-defined height-map by using some algorithm for estimating these new "in-between" points; or do you have to create a height-map with the smallest LOD implemented into your defined vertices?




posted January 14, 2000 06:22 AM            
a wedgie is that uncomfortable feeling when your undershorts get caught inside your gutamas maximus....

heh sorry i really dont know i am just playing arround


posted January 14, 2000 08:47 AM            
"Wedgie" is really not a very descriptive term. As I understand it, it really just means the maximum difference in height among the height-map data encompassed within a given triangle at some LOD coarser than the finest level. The diagram in the article (Fig. 6) is better than the text at explaining the concept.

And no, ROAM doesn't try to interpolate heights at a finer level than the data in the height map; indeed, of the dozen or so LOD algorithms I've investigated, none do this. The height map data is always the finest-grained level. How you *interpret* the height map data, of course, is a different matter. Seumas interprets his data as 1-meter intervals; you could use the same data and call it 100-meter intervals, or 0.1 meter. It's a simple matter of scaling.

Hope that helps.


posted January 14, 2000 03:49 PM            
i like my answer better :-P