Yeah, I have a function which will rotate a vector to any given quaternion orientation. Your description of covering the hemisphere line by line, rotating up a little at each stage does make sense. That only leaves me with the question of orienting all that to the surface. However, I guess the best way to do that would be afterward. Precalculate a set of vectors for a 0,1,0 normal oriented hemisphere and then rotate them all wholesale to the orientation of each normal as I go.
Yeah, it sounds ok in theory. I'll have a go at coding it up. Thanks for the idea.
Yeah, it sounds ok in theory. I'll have a go at coding it up. Thanks for the idea.