From: Allen A. <ak...@po...> - 2000-10-26 06:17:46
|
On Wed, Oct 25, 2000 at 11:08:37PM -0400, luc-eric rousseau wrote: | ... and if there are roadblocks I would hit. ... One thing that I would expect to cause trouble is the precision with which you'll need to represent intermediate results when interpolating vertex attributes. 32-bit fixed point is marginal for 16-bit interpolants. You might need to go to 64-bit fixed or double-precision float. Check out the math in tritemp.h. Allen |