Ability to see all expressions as a shape key grid and batch re-import them

[Theking1313]

Just caught some of the livestream and was thinking about a shape key grid / facial pose library like the ones I’ve seen online and in rigging videos.

Example from Blender studio,

The idea would be to generate a separate mesh object for every expression and arrange them as a gallery or grid inside the Blender scene. You could then go through them visually and use Sculpt Mode to edit each expression. It would be useful to include the eyes, teeth, tongue, and other relevant meshes with each one so you can see the complete expression as you sculpt.

After editing them, you could press a button to batch re-import the objects back maybe via it's name, possibly using the bone matching algorithm that was demonstrated in the livestream.

My main thought is that there could be a few different options depending on how much you want to edit:

• A smaller grid containing only the main expressions
• A full grid containing every shape key and corrective
• A filtered grid for specific areas or expression groups, such as smiles, eyes, mouth shapes, etc.

As a lot of the time, you may only want to tweak how the character smiles or how a certain expression looks, rather than fine-tuning every individual corrective. Having all the expressions laid out visually in the scene could make that workflow much faster and easier to understand.

[jack-yao91]

@Theking1313 thanks for filling this out.

Would this just be for the initial isolated Raw Control activated to 1? Or all the target Poses in the rig definition, like shown in the Behavior viewer? Once you get into these, there are shape keys that are left or right sided or upper and lower etc. I can see how this would be useful to paste over different parts of an expression, but what are your thoughts on expressions with combined correctives and which part of the expression belongs to which shape key?

Theres only 798 shape keys but I believe your looking at well past a thousand unique expression combinations. My initial thought is that we have to find a subset of expressions that would be automatically decomposable and have minimal or no shape keys down stream of them.

In the stream, some of the basic ones like mouthLeft, mouthRight, JawOpen etc. should be easy to do this and batch apply back.

What are your thoughts?

[Theking1313]

I'll admit I'm looking at this more from an artist-driven perspective, so yeah, having it work on some of the basic controls first would probably be best.

I can see the issue with certain presets activating many controls at once. For example, a preset like Anger-04 Rage, sculpting it closer to what you want for the face, and then mapping those changes back would make it difficult to know which shape keys or controls should receive the edits for each area.

I was naively looking at this more like an expression editor, rather than tweaking all the smaller shapes and correctives that activate the final expression.

So I agree that starting with the easiest method is probably best, just to see how well it works and the basic workflow.