Wasn't it Greg M. Johnson who wrote:
>Wow great!
>
>Thanks for your efforts.  Apologies if I "cast dispersions" on your macro-- 
>it's great work.   Perhaps I'm just doing something unnatural with the
>spline.
>
>The circular section version doesn't show this problem-- but my spline curve
>is lumpy so perhaps that's part of the problemo.

No, the spline lumpiness isn't the problem, nor is the fact that the
spline contains a mixture of vectors with different dimensions. 
(I didn't know you could do that, but POV apparently happily considers
them to be colour vectors and won't let you say things like 
  sphere {VSpline(I),0.1}
because VSpline(I) is a colour instead of a position.)

I think you're right that there's a normal problem. It's quite subtle. I
make my normals always perpendicular to the USpline, but they really
shouldn't be. They should tilt slightly as the diameter changes with the
WSpline. I've just uploaded a fix for this. I kept the old version as
SweepSPlineV2.inc because it seems to work slightly better on some
surfaces - I don't know why.

I think that there might be a general situation with mesh objects in
places where the light falls almost tangential to one of the triangles.
If the situation is such that a flat triangle would be in shadow
(because the light is just below the tangent) but a real smooth triangle
would be partly out of shadow where it curves above the plane of the
flat triangle, then the shadow may not look natural. In such situations
you might get round it by increasing the number of triangles.

-- 
Mike Williams
Gentleman of Leisure

Post a reply to this message