Search Results
6/18/2025, 6:19:59 PM
>>11313933
>I've been trying to do this stuff every few months for over a year now and came to the conclusion that you need to learn to be a good animator first and be able to sketch your own keyframes before you can get any cool results.
As an (amateur) animator I always suspected this would be the case. Much like an artist has to think about composition and make deliberate conscious choices, so does the animator in how each movement is interpreted and represented. How fast or slow, how exaggerated or muted, how much anticipation, rebound, stretch etc
AI struggles immensely because it does not make conscious decisions. I don't know precisely how it does things, but I know how it doesn't do things and this leaves animations feeling weightless and generic. It doesn't have a vision for what it is it's making and describing movement into words is even harder than visual elements.
Maybe one day it will improve to the point where it can add some nuance but it will still require a lot of human guidance to tell it how to do each motion if you want a proper result. Much like today with image generation, it may be hours upon hours of inpainting, splicing and refining, at which point I'm not so sure you wouldn't be better off learning how to animate. Even the best AI generated animations are still quite far from even basic live2D at the moment, so it's wait and see for now, for me at least.
>I've been trying to do this stuff every few months for over a year now and came to the conclusion that you need to learn to be a good animator first and be able to sketch your own keyframes before you can get any cool results.
As an (amateur) animator I always suspected this would be the case. Much like an artist has to think about composition and make deliberate conscious choices, so does the animator in how each movement is interpreted and represented. How fast or slow, how exaggerated or muted, how much anticipation, rebound, stretch etc
AI struggles immensely because it does not make conscious decisions. I don't know precisely how it does things, but I know how it doesn't do things and this leaves animations feeling weightless and generic. It doesn't have a vision for what it is it's making and describing movement into words is even harder than visual elements.
Maybe one day it will improve to the point where it can add some nuance but it will still require a lot of human guidance to tell it how to do each motion if you want a proper result. Much like today with image generation, it may be hours upon hours of inpainting, splicing and refining, at which point I'm not so sure you wouldn't be better off learning how to animate. Even the best AI generated animations are still quite far from even basic live2D at the moment, so it's wait and see for now, for me at least.
Page 1