...IDK, I haven't tried it out yet (I'm hoping I'll have time soon), but I think if you rotate the offset first, then it will work better. Hard to explain:

var offset[3];
vec_diff(offset, you.x, my.x);
vec_iRotate(offset, my.pan);

...the offset variable should now have the "local difference", instead of world. So now:

vec_to_angle(temp, offset);

...temp now holds the angle it should point in relative to the model.

c_rotate(..., temp, ... + USE_AXISR);

...I don't know, something like that. I never really use c_rotate().


Of course all of this is just theory, and I'll just have to test it later