deformers - blendshapes
Blend what?
Think of it as a transition from A to B. We have a sad face, and a happy face, when the face goes from happy to sad we’re blending one shape to the other.
The most familiar application of blendshapes is usually setting up rigs for
facial animation, though blendshapes can be used to solve a lot of other problems
too.
Characters can have hundreds of these shapes all defining little areas of movement in the face that combine to create complete expressions. It is then up to the rigger to create an intuitive ctrl system to drive the blending.
Here is a look at a blendshape created for the cube, with 2 shapes it can blend to.
Side Note: Each cube used as a different shape share the same structure as the shape we want to deform. If we don’t do this the computer won’t know how to blend things that don’t exist on the base model.
So as you can imagine, if you have hundreds of these shapes for your 3D asset it can take some time to setup a good control rig. Most riggers learn some basic coding skills to help reduce the time it takes to set things like this up.
Corrective shapes is a term you’ll hear being used a bit of too.
It is a slightly more advanced topic, but essentially these are blendshapes too.
But they’re really not human readable (that means if you looked at the shape you’d think it was just a mess) and that is because they are special shapes that are used to fix tricky deformation issues in, for eaxmple leg/arm FK/IK setups, and the corrective shapes get driven by rotation values / psd drivers(see below) from the rig automatically, animators would never need to worry about the corrective shapes.
Correctives come with their own setup implications and things such as the braveRabbit weightDriver tools get written to help reduce the pain involved in managing these kinds of workflows.
If you’re starting out, it will take some time before you really start needing corrective shapes, but it’s important to know of the concept.