Assignment 4 – Happy Blendshapes

/ˈhæpi/

hap-ee

I determined that the ‘hæ’ sound would be the same viseme as they are combined in the movement when I say ‘Happy’ in the mirror, and the ‘p’ and the ‘i’ sounds would be their own visemes.  It is possible to make the ‘h’ and the ‘a’ sounds separate, but that sounds incorrect when added to the rest of the word.

Volumetric Facial Models

Rafiki rafiki1_persp

rafiki1_front rafiki1_side

Rafiki2 rafiki2_persp

rafiki2_front rafiki2_side

 

The main parameters that would be needed to easily deform Rafiki’s face are the jaw, the brows, and the nose.  The brows are very expressive, going up to show the whole eye for the excited face in the first example, and going down very far for the look of deep contemplation in the second example.  The jaw morphs drastically from the first image to the next, going from two distinct sections of the upper and lower mouth (the skull and the mandible) to a small, almost unseen slit in the second.  It would be hard in a 3D model to hide the teeth in from the first pose to the second due to the wild variation in the shape of the jaws.  The nose stretches a lot from the first pose to the second as well, due to the exaggerated squash resulting form the open mouth and the exaggerated stretch from the lip going down.  The lower eyelids are also a consideration because they define the shape of the eye in the expression, and the pupils would also need to be able to move.  The hair could probably just be an afterthought if modelled accordingly- since I used modified volumetric cones, cubes, cylinders, and spheres it was harder to get them in the right place according to the picture.  The hardest part to work with would probably be the closing of the lips around the teeth, transitioning into the smaller mouth shape of the second posture.