On Mon, 12 Jul 1999 17:57:49 GMT, pet### [at] usanet (Peter Popov)
wrote:


>Suppose I have a sphere that is electrostatically charged. The
>intensity of its field at any point in space is proportional to the
>reciprocal squared distance, and the vector points away from the
>sphere. For a plane, it's easy. I think I can manage torii, cylinders,
>and cones. Even non-smoothed meshes.
>
>How about a smoothed mesh? OK, I have all the vertices and
>corresponding normals stored in two arrays, but then what? How should
>I proceed to calculate the field vector and intensity at an arbitrary
>point in space?
>
>And yeah, I am talking about POV script here, not C :)

Take your existing vertices and apply Catmull-Rom spline
interpolation.  (Use Catmull-Rom because it makes determining normals
trivial).  Use points from the resulting spline to refine the
triangulation until it is accurate enough to suit you.  Then use your
non-smoothed mesh method.  Just a suggestion off the top of my head.

Jerry Anning
clem "at" dhol "dot" com

Post a reply to this message