Kenneth wrote:

> Each rand is a different value--but those three different values will always be
> the 'same', since they use the same seed() value. Change seed() and you get
> three *new* and different rand() values.

I actually thought I'd be clever and trick pov into providing
a seed based on the pixel value of a jittered 2x2 pixel image
rendered in a first pass ;) However, it seems the internally
used random numbers are not seeded by a timer either, so when
rendered after a restart the results were again predictable.

Post a reply to this message