Thanks, I didn’t know that reference; I will have a look. ]]>

Also, there is one small error: when you solve for the quadratic form for a, b, and c, in the equation for b, everything to the right of 2 should be inside a bracket. I.e., 2 multiplies everything there not just the first term. Thankfully this was easy to spot as you were very thorough in your derivation.

Thanks again!

]]>Very nice write-up ! I’ve implemented that six years ago in the context of raytracing curves.

But the best way to raytrace curves is found in the paper from Koji Nakamaru and Yoshio Ohno which I believe is widely used nowadays. This way you don’t need to discretized the curve in small segments (facing camera) or small cones, etc.

Cheers!

]]>Sorry, this is a major pet peeve of mine, one sees it everywhere! :”D

]]>