Change rig after animation has already started?
Hi ppl,
I just wonder what solutions there might be out there to solve this case, if you wanna change a character rig (esp. mgear rig) after the animation has already started and animators have already produced animations that you don't want to put in the trash bin...
Since (as far as I know) iterative work is quite popular and systems like mgear are pretty much built to enable quick changes, I can't imagine that this is an uncommon problem in studio pipelines and probably most studios have some kind of solution for that!?
At least, I know that there are some custom solutions for that Ubisoft Montreal and Toronto with their custom IK Rig, but of course, that's months of custom development (and not for Unreal Engine). Maybe there is a simpler solution, maybe even a standard one for Unreal Engine? I cannot imagine that Epic Games doesn't face this issue in their pipeline, so surely they provide something to solve this in Unreal Engine?
Any thought and tip in a direction is appreciated.
Thanks a lot and have a nice day,
Micha
Replies
"after the animation has already started and animators have already produced animations that you don't want to put in the trash bin..."
Well, this whole premise is flawed anyways : A rig consists of a set of controls driving an armature ; animations for a given armature can be produced by different rigs. For instance, animators making animations for the Unreal Marketplace don't need to use the same rig - they just need to drive the same base skeleton armature.
So no, none of these already made animations need to go to the trash bin, since the game only cares about the exported animations and never sees the rig(s) used to produce them.
Of course the above only applies if what you mean by rig is actually the rig. Now if you actually mean a change in armature/joints/skeleton, then that's a whole different story.
in that case unreal has retargetting built in - as does motion builder and I'm sure there's something for Maya
Thank you for your answer. Ok, fair. I forgot to be more precise by what I mean. I was talking about changing the skeleton, like adding joints or sth like that. So what would be the workflow to keep old animations while updating the skeleton? Is retargeting the option to choose? So, I guess, retargeting would be kinda daily business for tech animators or riggers in a game studio, if you want to iterate over the skeleton often? Or how should I imagine this in a game studio?
In a game studio (or any CG production really) something as drastic as a skeleton refactor is not supposed to happen, because it would break a huge amount of things. The preproduction period is precisely here to avoid that. And then these things get locked in for the duration of the project.
If such a change is really needed, then everything related to that side of the production pretty much stops for a week or more so that the assets can be ported over - and then things can go back to normal IF things didn't break too much.
Now a change of rig (as opposed to a change of skeleton) only affects authoring but not the game assets themselves, so that's much, much less of an issue (if any).
"Since (as far as I know) iterative work is quite popular ..."
Breaking a skeleton mid-production is 100000% not a popular thing to do.
Thank you for clearing this up :)
Shouldn't and can't are different things
Can you come up with a way to make this viable?
yes you can - provided you plan it into the way you work from the very beginning. I've built several rigging systems where the the bind (exported) skeleton is 100% independent from the control rigs (the ones you animate) - in those cases your animation work remains valid unless you change something about the proportions of the skeleton/add legs etc. and you just have to account for anything you've added. This sounds great but there's still a lot of other work associated with the changes (like rebinding all the meshes to the new skeletons, integrating them with the control rigs again, patching up anything that breaks at the game end etc.
Should you ?
That depends. If you feel like you're going to be working on the same sort of things for the next 2-3 game projects then investing a few months in tooling up for this sort of thing makes a lot of sense. If you're likely to move onto something completely different then it could well be considered a waste of time