Relative FK Spine (RFK spine)

Hello-

Retargeting animation has become important on the studios projects, and I want to make a spine that distributes twist from the top and bottom to the middle. I want to find a solution that doesnt use spline iks. I currently have a system that uses world matricies, but it flips after 90 since it converts quats back to eulers. Not converting back causes issues when the spine folds forwards.

David Hunt describes what im trying to make here.

Any ideas how hes achieving this?

Thanks,

-Thomas

Let’s start with the “is it plugged in” question: Could you just use an orient constriant?

I fully recognize that the answer could be “no” for any number of reasons. But whatever that reason is could help me come up with a usable solution.

I found a video made by the guy they talk about in the bungie video. https://youtu.be/wjjzM1Vy344?si=tmI9BE2-lhsu2qzN&t=1000. Im looking into orient constraints and hwo tehy respond to forward bending. theres a strong chance i may ahve been overcomplicating.

Figured out what I want–

I used a flat hierarchy for the twisting controls and orient constrained an offset for the middle segment to the top and the bottom so it averages between (only constrained to the rotation axis), and then connected the main FK controllers to the offsets of the twister controls. Connecting the twisting controls to the deform joints gets it all hooked up, and its working without cycle errors. With a 3 joint spine (which im happy with) it works up to 180.