Hi everybody.
I really miss Shape Keys based on how they are implemented in different 3d programs. For example, in Blender 3D, as a sample.
I.e. there is a basic mesh. And you can record several samples of the distorted grid. And independently mix them in the animation with each other. For example, closing the eyes, different facial expressions. Here is a description of the process: https://docs.blender.org/manual/en/latest/animation/shape_keys/shape_keys_panel.html
Shape Keys
Spine does not currently support shape keys in the same way as Blender or other 3D programs. In Spine, vertex animation is achieved through deform keys, which are similar to blend shapes or morph targets but are not as flexible for mixing different deformations independently. Deform keys in Spine store the position of all vertices for a mesh, and all vertices are keyed together, which can make it challenging to animate different parts of the same attachment independently. However, you can use weights to animate bones that deform the mesh, which can be combined with deform keys for more complex animations. This approach can reduce the number of deform keys needed compared to using deform keys alone.
Not sure if this will work the way I have in mind.
On the video, as I understand it, you can use the slider to mix in a certain animation. Or not? The point of Shape Keys is that sometimes it is difficult to adjust the weights on the bones to separate parts, especially when using pseudo-3d. Then you can move individual parts of the mesh using the sliders. And animate parts of the mesh asynchronously. Which is hard to do on complex meshes right now. The approach with separate animations is also quite good, but it probably won't work. For example, you need to cover one eye, and the mouth should articulate speech. Shape Keys have indexes of individual points. When mixing individual animations, this is most likely not the case.
You're right that even with sliders, deform keys are for the whole mesh. We don't currently have a way to key a deform for only part of the mesh.
In Spine you are usually better off using weights then moving bones to adjust a portion of the mesh. Using deform keys has some disadvantages listed here:
https://esotericsoftware.com/spine-keys#Deform-keys
While only one of those reasons is "all vertices for an attachment are keyed together", we could definitely improve keying deforms so only a portion of the mesh is affected. It would still be more efficient to use weights and bones, when that can achieve your goal.
Spine is more sensitive to performance that most animation apps, as we have to handle the runtime part and keep that efficient. We could do such cool things super easily if we were only outputting images or video!