Converting euler to rotation matrices

[areiner222]

Hi,

Thanks for your great work!

I noticed in the smpl dynamics tooling (dnd.models.smpl.lbs.euler_to_rot_mat3) it seems like you convert from euler angles to rotation matrices around x, y and z axes but then you rotate the joint unit axes with a ZYZ rotation. Please correct me if I'm wrong, but if you convert the euler angles to xyz rotations then shouldn't we need to do a XYZ rotation for the axes?

I was investigating the ZYZ rotation, and I see that the resulting rotated "axes" do not preserve orthogonality in the column vectors. Is this expected? Am I missing something?

Thanks, Alex