Re: Approximating SubDiv mesh with NURBS

Nicholas,

>   Given a subdivmesh polygonal cage, is there a way to approximate the
> underlying limit surface with NURBS surface?

if you are talking about Catmull-Clark subdivs w/o creases and
corners, then yes.
The easiest way I'm aware of is to subdivide once (so your mesh is all
quads). Now any connected submesh that consists entirely of vertices
with valence 4 can be used as the control polyhedron of a cubic b-
spline surface.
Problems usually manifest themselves where such meshes meet at
vertices with a valence != 4.

If you need something quickly, I believe Hal Betram has written a C-C
subdiv evaluation lib that he licenses to third parties. :)

..mm