"Eriban" <pov### [at] spamgourmetcom> wrote:
> No, that should not be too difficult indeed. However, I don't think it would
> look too good, or at least, that it won't look that biologically plausible.
> Scales to me suggest that they have grown independently, which would make a
> continuous pattern across the scales seem unnatural.

Maybe there's a way to get the colors confined to the same structures that make
up the cells?

Hum... what one need would be a "crackle warp", that takes a crackle pattern to
warp a normal 3d co-ordinate to the corresponding "crackle cell's" center
point, and pick the color from there - resulting in the whole cell being
colored the same. So one could use the same crackle pattern both for normals
and to warp the current pigment. There, your scales.

To achieve this, you could conjure up a function to take a few samples of the
crackle "potential" to estimate the gradient vector, use that to estimate the
location of the cell's center, and pull the color value from the original
pigment at that point.

It would be a bit cumbersome because POV doesn't allow vector functions, so
you'd have to call the function separately on each dimension, but somehow it
should be possible.

Post a reply to this message