I'll have to post up some of the work I did with a research group at USC. It's full motion capture like this too. We had markers on the face and body that captured obviously the body and face animations/models but also got diffuse, normals, specular, and subsurface scattering diffuse. The lips should be more accurate if…
It's interesting they went for the easiest haircut they could find, and left it out for most of the video. For all the blockiness the graphics of 15 years ago had, they at least could make curly hair. Also, a question. Isn't it possible to make a special texture to darken teeth that depends on the animation, like a wrinkle…