Just a note on my workflow. I am still planning to go with the joint-based facial setup, but first, I will be modeling all the defined shapes, and then match up the performance of the joint-based facial setup to achieve the modeled shapes. This is to emulate the actual production situation whereby all the facial setups of the character have to be "in-model", that is, approved and defined by the director (That is the reason why most production facial setups are still blendshape-based). To summarise:
- Joint-based facial setup to simulate "skin-sliding" effect of the controls
- Model actual blendshape targets to define the look and feel of the character
- Make use of the blendshape targets as additional corrective shapes of the facial setup
Next step: Rough blocking of the body. Just wanted to roughly define the shape of the shoulders and neck first to lock down my face geometry.