"Slime" <fak### [at] emailaddress> wrote:

> If you're not taking enough media samples, you get incorrect results, but
> sometimes it looks good enough that you don't know or worry about it.
> Unfortunately, when you intersect another transparent object with the media,
> you get *different* incorrect results depending on whether the ray passes
> through the second object or not. You can tell if this is the problem by
> seeing if it goes away when you raise the samples a lot ("samples 1000").
> Sometimes you can find an ideal number of samples that looks good and
> renders fast enough... sometimes you can't.
>

That has been my experience as well, and raising the samples value does work, as
you mentioned. Took me a long time to figure that out.  :-(

Something else to note: If you have two media container objects where one is
totally enclosed in the other, the enclosing object's samples *override* the
values for the interior media object. They seem to be totally ignored. In fact,
it seems that samples count, method and intervals for that interior object are
all ignored, and can be left out. That has its pros and cons: If for some
reason you actually *want* the interior object's media to have a different
samples count (for its visual results, for example), it can't be done. I assume
that POV-Ray just "works that way" when dealing with multiple overlapping
medias.

Ken W.

Post a reply to this message