@Quillist can we get some resolution criteria?

Here's an example of what diffusion models can do now: https://civitai.com/images/108363698

@ProjectVictory I don't have a resolution benchmark yet, but I can clarify better what freehand animation means.

I am refereeing to freehand movement. A style that is only really achieved when each frame is hand drawn and where no interpolation / tweening tools are used. One characteristic effect is smear frames

Quick movements could produce distortions that are nonsensical when looked individually, but make sense in quick sequence.

@Quillist I see what vibes you are going for, but you should consider coming up with stricter criteria early on, because it's very common for market creators to no longer be on the site by the time resolution date rolls in and the mods will have to interpret what you meant and decide if it was achieved.

@ProjectVictory The resolution will look something like the following:

8 prompts will be given to the video generative model.
Something like:

• “A dancer performs an improvised freestyle spin with loose arm flow and shifting balance.”

• “An artist sketching on a large paper pad, hand making quick circular strokes.”

• “A boxer shadow-boxing with dynamic torso rotation and fast arm recovery.”

• “A skateboarder losing balance, then regaining control.”

• “A person slowly turns their hand to examine their palm in curiosity, fingers flexing subtly.”
  
  
  I would want to sample a movement that was already executed by an actual artists of varying skill levels (beginner to professional) for comparison, and then do a direct comparison followed by a blind competition between real vs generative works.