It would be nice to have the option to use the webcam to drive mouth shapes. Sometimes I want to hold on to a certain mouth shape, but since there is no audio, it reverts back to the default shape. I don't want to rely only on keyboard shortcuts. Also, there are occasions when the room is too noisy for audio lip-syncing.
Currently the only way to do webcam mouths are with the Nutcracker Jaw behavior. Name something "jaw," add the behavior, and it will move up and down with your own chin. I've seen some really smart uses of this.
Thanks for the idea, we can check it out. I'd imagine audio is always going to be more accurate, but who knows. Currently, triggers really are the best way to hold a mouth - there is an option under lip sync to control them with keys, just by using the first letter of each viseme. Or you can make custom mouths/triggers. If you're recording, you can also extend or delete shapes while editing with the "Split Lip Sync Into Visemes" option.