Hi everybody!
So recently I've been interested in 3d animation after watching this video:
Which shows some really cool stuff, like animation retargeting to entirely different non-human skeletons from human mocap.
They've shown it at GDC and the talk is available on GDC Vault but I don't have access :( They released the presentation though, which I didn't save and now it's unavailable. But even after reading it it's a little difficult to understand.
How does it all work?
What I already understand
- They call it the "IK Rig" while everybody else calls it a "Runtime rig" :P
- Instead of baking all the animation data to bone translations they export the IK goal positions directly and that's all the prebaked data they have. Everything else like (skeletons and their constraints) is calculated based on these IK goals at runtime.
- They use a bunch of constraints on bones, like position, rotation, look at etc, but I don't understand how they're supposed to work...
Can anybody maybe point me in the right direction, like some articles, books, anything?
Thank you.
Edit:
So I experimented a bit in blender and got this result:
Which is using 2 positions: begginning and end of the leg, and one rotation for foot angle, everything else is interpolated from that, it has 4 bones for the actual skeleton.
So basically I kind of sort of figured it out, except I still don't understand constraints they were using.