Hey Louis - great post. You can get a similar effect like this:
- Have your animated blendshape mesh as a Ziva bone
- Duplicate the base mesh and make it a fairly hi-res tissue
- Have multiple attachments which attach the tissue to the bone, with different maps painted for different regions of the face.
- For each blendshape target, you'd have a corresponding mapping to the envelope on each of attachments. For example when the r_smile blendshape target is dialed in, this would set the 'lips' attachment strength to 1, l_cheek attachment to 0.0, r_cheek attachment to 0.75.. etc etc.
This way, when the input mesh (Ziva bone) is in motion, different regions of the tissue face are trying to hit the input mesh at varying intensities.
You could do a similar thing with material layers.