On Fri, 20 Aug 1999 09:59:31 -0500, Chris Huff wrote:
>This has bothered me for a while now, why hasn't that noise function
>been fixed? Is there a maximum value to it, so the value of the pattern
>could be scaled down to fit in the correct range? Is the overrun because
>of numeric inaccuracies?
>
>I haven't done any tests with crackle, though. I wonder if there are any
>other patterns with this problem.
>

It can be scaled.  Crackle is, in fact, though the scaling factor seems a
bit arbitrary.  But part of the problem is that the overruns are usually
outliers, and scaling the pattern to fit the 0..1 range perfectly would
decrease the range of the far-more-prevalent middle ground.  Also, of 
course, making such a change at this late date would likely affect many
scenes - if you spent hours tuning the cloud density in your bozo sky, you
might not appreciate it if it were to increase or decrease significantly
in the next version of POV.  Personally, I think the best solution is to 
just let maps contain values outside the range 0..1.

Post a reply to this message