Perhaps it would be a good idea to send at least one extra trace() which
is just slightly off the original direction. You can do this one only if the probable
rebounce position is already in direct vicinity. If the two gathered normals
differ too much, perhaps you have an idea on what's going on?
It's just an idea. Not too parsing intensive, I guess, but perhaps some more thinking
would do a better trick.
Tim
Post a reply to this message
|