To be fair, stop motion animating is fucking time consuming and such an art. Let alone for first person camera views AND third person
Having an AI quickly generate a hand wave motion etc and convert to first person and vice versa would save hours and hours of capping one frame at a time
Most digital animation done by hand isn’t done in the stop motion/frame-by-frame way, It’s done with key frames that are interpolated between. That’s not to say it isn’t still super time consuming (especially when done well).
A lot of motions still need to be scrubbed frame by frame and adjusted if any of the key frames dont interpolate cleanly
It’s one of those menial tasks that AI could actually specialize and hugely assist in for simple repetitive animations. Or imagine have a custom animation that the AI/ML uniquely generates the key frames on the fly. Talking distant future but every animation could be potentially be entirely unique everytime you see it
I could see a system where you very loosely define a handful of keyframes, give it some kind of description for what your looking for, and then the model fills in the rest for you. It almost certainly would still require some tweaking on the artists part, but I think a huge benefit is that AI can be non-deterministic. If you don’t like what you got you can just run it again without changing inputs. Maybe save multiple runs and manually combine all the best elements. Lots of opportunity.
That’s not true, with motion capture you are recording an actor do all the motions and fixing whatever is necessary. It’s different to pose to pose animation.
To be fair, stop motion animating is fucking time consuming and such an art. Let alone for first person camera views AND third person
Having an AI quickly generate a hand wave motion etc and convert to first person and vice versa would save hours and hours of capping one frame at a time
Most digital animation done by hand isn’t done in the stop motion/frame-by-frame way, It’s done with key frames that are interpolated between. That’s not to say it isn’t still super time consuming (especially when done well).
Yes! You nailed it
A lot of motions still need to be scrubbed frame by frame and adjusted if any of the key frames dont interpolate cleanly
It’s one of those menial tasks that AI could actually specialize and hugely assist in for simple repetitive animations. Or imagine have a custom animation that the AI/ML uniquely generates the key frames on the fly. Talking distant future but every animation could be potentially be entirely unique everytime you see it
I could see a system where you very loosely define a handful of keyframes, give it some kind of description for what your looking for, and then the model fills in the rest for you. It almost certainly would still require some tweaking on the artists part, but I think a huge benefit is that AI can be non-deterministic. If you don’t like what you got you can just run it again without changing inputs. Maybe save multiple runs and manually combine all the best elements. Lots of opportunity.
That’s not true, with motion capture you are recording an actor do all the motions and fixing whatever is necessary. It’s different to pose to pose animation.Removed by mod
Turns out I can’t read