Hi. We are trying to figure out our workflow for modular character animation mixing and hope someone could help steer us in the right direction. We have a lot of characters talking in our game so we want to ideally mix and stack ( additively ) 4 animations:

1. 1) idle loop of character standing in place, breathing, blinking. This always plays.
   2) a fixed facial expression that changes the eyebrow position for things like: neutral, happy, sad.
   3) a fixed body pose that takes on the idle animation without having to reanimate it. say arms up, arms down, leaning forward, leaning back.
   4) Talking animation loop that rotates the neck and moves the head position and animates the mouth whenever a character speaks.

The problem we have is with any shared keyframes ( like neck and head position and rotation), the next track just overrides the other. We are currently doing mixing with one animation at a lower percentage but the results are often wrong. We've also considered doing double parenting of bones so theres a neck bone for the idle animation and then a child of that for the talking animation, but you can get weird results with that as well.

Do we just need to make a ton of animations for every combination? Can someone suggest a better way to do this or any workarounds? Thanks!

I think we worked out a method based on secondary bones. Sorry for the trouble!