Chris Huff wrote:
> 
> I don't think it's that bad...no more so than any other curved surface.
> Recursively subdividing the sides of a tetrahedron or octahedron seems
> to perform pretty well. You don't have as much control over the
> resolution, but the triangles are fairly even in size. Add in smoothing
> of normals, and there isn't much of a problem.

I think tesselating a smoothly curved surface is always bad to some
extend.

> 
> > Newly generating the mesh with every change of viewpoint, ...
> 
> Why? The viewpoint doesn't affect the geometry of the scene...
> 

Because the needed accuracy of tesselation depends on the viewpoint, if
this is not taken into account, visible corners will occur.  

You can of course create a quite 'foolproof' tesselation of a sphere with
- let's say - 1000 triangles, but in certain situations also this won't be
enough and in other situations it would be far less efficient to use a
1000 triangle mesh instead of a simple sphere.

Christoph

-- 
Christoph Hormann <chr### [at] gmxde>
IsoWood include, radiosity tutorial, TransSkin and other 
things on: http://www.schunter.etc.tu-bs.de/~chris/

Post a reply to this message