> It's not that hopeless, you could store all function values evaluated and
> use them later for new rays, together with the known max_gradient you can
> reduce the number of necessary new function evaluations this way, but you
> would have to design efficient methods and data structures for accessing
> the stored values and the memory requirements can be enormous.

True.

Chances are, unfortunately, that using these remembered sample points would
only make the render time longer.

- Slime
[ http://www.slimeland.com/ ]
[ http://www.slimeland.com/images/ ]

Post a reply to this message