Shape-keyed mouth movement test

I’m working on a new head, and, before I get too much further, I was wondering if I could get some feedback on how it’s working so far in terms of the model, the movment, etc.:

Any comments and suggestions are welcome.

There is no nasolabial fold!

Can you show wires?

Hi, nice work. It looks like the lips are gaining mass as they’re being stretched. I appreciate that they’re being flattened against the teeth, and maybe making them a little fuller in the rest position may help.

Thanks for the suggestions! There actually was a nasolabial fold, but it was shallow and the lighting was washing it out. I made it deeper and changed the lighting to show it better.

I adjusted the smile shape so that the mass didn’t look to increase on either lip–good spot on that one.

Here’s my new render:

It also features a bunch of new shapes: narrow, upper lip up/down, lower lip up/down and a bunch of combinations with mouth open and closed. I also moved the teeth up Z so that they actually made sense with the lip position.

Any new suggestions are much appreciated.

I see the need for a ‘snarl’ crease above the nostrils when the upper lip is drawn up. Almost the whole lower nose - at least the fleshy bit - would rise with the upper lip (though not as much). Likewise (in reverse) when the upper lip stretches down covering the teeth. Minor crits though, can’t wait to see a movie with a little dialogue!

Hmm, that’s interesting. I’ve been working pretty faithfully from Adam Osipa’s Stop Staring book–which is just flat out excellent for this stuff. At any rate, he has two different shapes he mentions–one for upper lip up, which only move the upper lip up, and one for “sneer,” which is closer to producing the snarl crease you mentioned. His upper lip up shapes seem more geared toward producing phonemes (like “sh”) and the sneer/snarl more for emotions.

But I just noticed that you’re absolutely right–there IS a definite movement of the nostrils based on the contraction of the Levator labii superioris’ lateral part. It’s not as pronounced as a full sneer/snarl, which seems to involve the LLS’s nasal part (and includes “scrunching” the nose), put there is movement there. Ditto with the “upper lip down” shape–slight downward movement on the outer nostrils.

Since a lot of Osipa’s shapes are based on working nicely with other shapes for blending, I’ve have to test out the effects of adding creases and slightly raising the nostrils when mixed with other shapes (particularly with “narrow” in order to make the “sh” shape.) Still, it’s an excellent observation and one I’ll definitely experiment with.

I’m dabbling right now with CMU Sphinx library in an attempt to make a super simple speech recognition app. As soon as I figure out how I can get it to spit out timed phonemes, I intend to convert the output to something compatible with the speech Python scripts I’ve seen available for Blender. Then I’ll tie it in to the model–and we’ll REALLY see some dialogue! :wink: