Gail Shaw wrote:
> 
> A reasonable approximation of a sphere isn't hard. A sphere's geometry
> is well defined. 

A reasonable tesselation of a sphere is quite a difficult thing IMO. 
Enough triangles that no corners are visible with the current camera
position and angle and not to much that it does not need too much memory.
Newly generating the mesh with every change of viewpoint, ...

Of course these things also apply for other objects, but most other things
do not need to be that totally smooth.  

> The problem I can see with isos expecially is that they
> don't
> have a well defined geometry and can look like anything.
> 
> I don't even want to think about writing code to tesselate any isosurface.
> 

I don't see that much principal difference between a sphere and an
arbitrary isosurface apart from the fact that you can parametrize a sphere
and thereby have a way for tesselating it.  

I recently read an article about tessellation of isosurfaces (in
March/April issue of IEEE Computer Graphics and Applications) and it's not
that difficult, the method Warp uses in his tesselation patch for example
works equally good (or bad) on conventional objects and isosurfaces.  

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