"Gregory D Abram" <[EMAIL PROTECTED]> writes:

> 
> If you are dealing with surface elements - eg. quads and triangles,  and
> you are rendering smooth surfaces,  I'd give serious thought to Lloyd's
> solution - the overhead of converting quads to triangles is probably worth
> the C1 continuity that you get.   Visible partition boundaries suck.
> 
> Greg

When I used DX some time ago, it appeared to me that DX does a
decomposition of quadrangles in triangles anyhow, i.e. they didn't use
a bilinear interpolation on quadrangles.  Is that correct?  And can
that behaviour be enhanced?

Nicolas.

Reply via email to