Open ginkxo opened 4 years ago
I don't think that's expected for the example. Maybe there's something wrong with the implementation.
The per corner normal should have sharper edges like the example in the README.
Make sure you pay attention to the units of what's passed in
So it appears my per-corner normals implementation gives a cup that looks exactly identical to the per-vertex normals implementation (including some blurriness), mostly because it seems that every adjacent face's dot product is less than the provided dihedral 20. is this expected for the example?
I tried to see if there was a difference when not including the normal of the face whose current corner it belonged to in the calculation of the weighted normal sum, and there were clear artifacts: