ss7m+@andrew.cmu.edu (Steven M. Stadnicki) (06/29/91)
I'm working on a small ray tracer that uses a grid for its accelleration
technique, and I can understand all the bits of traversing to the next grid
cell and such, but I'm having problems scanning my primitives (spheres,
polygons, and possibly cylinders and cones) into the grid... for spheres I've
been trying an approximate technique, saying that the sphere lies in all cells
whose centers are less than r+(sqrt(3)/2) from the center of the sphere, the
sqrt(3)/2 factor being the maximum diagonal distance from the center of the
cell to one of its edges; this may give a few false hits, but it seems like it
ought to be okay. On the other hand, I have no good ideas on how to do
polygons (just triangles and parallelograms) or cones of whatever nature...
does anyone have any experience with this that they could share? It doesn't
seem to get discussed in any of the papers I've looked at...
Steven Stadnicki
ss7m@andrew.cmu.edu