Kyle wrote:
> I'm still curious as to why the artefacts appear, though, even with a higher "n"
parameter of 0.25.

  It's hard to say the exact reason offhand, but it's most probably
related to floating point inaccuracy. Maybe 64-bit IEEE floating point
numbers simply aren't accurate enough in this case.

  (Of course there's a slight possibility that there's some other
reason, eg. a bug or an ill-written implementation.)

Post a reply to this message