Just to add to what Jimmy nicely rounded up: the facial animation rig was pretty much all done with envelope deformers rather than blend shapes. Chimp lips are surprisingly flexible so having the deformers being driven by curves was certainly the way to go with that.
The eyelids were also controlled with nulls looking at curves so the animators could define pretty much exactly what kind of shape they wanted - there's a lot of nice, subtle movement in them which is pretty much all down to their keen eyes and decision-making, really. Not much of it was automated so they did a really nice job - pretty much everything you see was something they chose to move, I just gave them the option! Throw on a few sticky controls in key areas for secondary movement and you've pretty much got it as far as the control rig went. Then off it went to Vince's muscle system (though by that stage I unfortunately had to head back to the UK office, so I never got to see the cool stuff in action). On 21 May 2013 22:05, jimmy gass <ji...@nervegass.net> wrote: > No strand caching. We wanted to save our self the extra step, since there > were already like 5 stages of cache and sim. So the strands were left live. > To get Motion blur to behave properly, I just made a compound at the front > of the entire system, that calculated the strand velocity and stored that > every frame, but then set it back to what it was the frame before at the > beginning of the sim to keep the behavior right. Basically just forcing the > velocities for the rendered to do what it needs to do. That node has become > quite popular here for that purpose. > >
