On 02/22/2011 08:40 PM, clipka wrote:
> The die is actually an isosurface I had modelled earlier for a different
> scene; DON'T TRY THIS AT HOME WITH SSLT! Kudos to Kevin Loney, Jaap
> Frank & Tor Olav Kristensen for their isosurface approximator include
> file, which cut down the rendering time by a guesstimated 98-99% (!) to
> "only" 3.5 hours.

no kidding ... why is that with isosurfaces the render times skyrocket
(besides the obvious isosurface slow downs) ... i tried to make coral
material for my mandel goodie and after overnight on 13% progress ...my
guess is that up until now that fairly simple objects have been used as
demo. you guys are getting fairly decent times with mesh objects am i
correct? iso's are just a more complicated surface to consider
(sslt-wise) is that the reason? btw: did that test with RC3 bits ...
should i expect similar performance with updated method?

Post a reply to this message