"KalleK" <kal### [at] gmxde> wrote in message news:3a152b70$1@news.povray.org...
>
> (Thanks! I'll keep that rainbow and skysphere.)

YW

> > // problem with this part?
> > plane {y,-.01 pigment {checker color rgb .67 color rgb .75 scale 10}
> >  clipped_by {box {<-100,-1,-100>,<100,1,100>}}
> > }
> It's exactly that part. Your plane is too close to the x-z-plane. (if you
> change -.01 into -.02 there are no speckles)
> Guess:
> The checker-pattern changes in the height of 0. If you get too close to this
> change, this effect appaers. And because of the 'scale 10' in the
> pigmentstatement you are "in reality" 0.001 povunits appart from zero.

I hadn't thought of that.  Thanks much.

> I'm a bit astonished at the fact, that the precission problem is that big
> (0.001 is a huge number in this case)!

Well, yeah, I knew that number tends to be at a lower limit for some things.  It is
something I try to keep in
mind but I still often use such low scales/numbers in defiance of such limitations. 
Texture is not the usual
thing I consider though, out of habit I guess.

> I came across this checkered-precission-problem in former times. When I
> layered a checker and a scale-2-checker, the changes where not the same
> place. (I made an animation in which I cut a box with a moving plane and the
> different scaled checkers jumped in different times.
> Once more shall we say: is it a bug?
> But my effect was without speckles, maybe the errors got nothing todo with
> eachother.

That is good to know about.  Bad, but good to know.

Bob

Post a reply to this message