This little project showcases an approach you can take to combine different automatic generative processes together to quickly create generative animation. Everything is derived from a single static source image.The animation above was generated using the random walk feature of the new QuadMixer. So i'm randomly walking automatically through a morph parameter space plane and recording all of the output in the History sequence. I then run an Action menu command to convert the entire history sequence into a single key-framed Vectorizer action step in my paint action sequence.
I then used my paint action sequence to drive the latent input to the U-Net in the generative ai context. I'm working with a static story and style text prompt, so no key-framing of the prompt.
You can extend this approach using additional Studio Artist features to slow down the animation movement if you want that. We'll cover that in a later post.
No comments:
Post a Comment