Lattices vs shape keys for facial rigging?

For the past few months, I’ve been digging pretty deeply into rigging.

I’ve read discussions on using bones vs. shape keys…but then I came across this rig, which uses a combination of shape keys (for the eyebrows) and lattices that interact with vertex groups for the mouth shapes.

There is a lattice modifier on the face. The greater the strength of the modifier, the more a specific vertex group on the face gets deformed so that it matches the shape of the lattice.

As a side note, this can have a sort of weird effect if you move the face mesh without also moving the lattice. The shape of where the lattice is remains the same, but the vertex group will continue to be affected, albeit in a different spot:

Anyway, my question is, what advantages/disadvantages do lattices have over using shape keys?

Depends on how you animate the lattice. If you animate via shapekeys on the lattice, it’s going to interpolate pretty similarly. If you animate via an armature, it’s going to interpolate differently.

Beyond interpolation, lattices are just non-destructive modelling tools (with the understanding that animation is “just” modelling over time.) You could write your lattice modifier, in several positions, to shapekeys, and use shapekeys if you’d prefer. But once you have a lattice set up, you might want to leave it, to demonstrate the technique, or to more easily add more expressions later.

Lattice is pretty similar to a mesh deform. The main advantage to either over moving things manually is that there are fewer verts to control. That’s also the main disadvantage. (Lattices are also a pain to edit. Mesh deformers are easier to edit.)

1 Like

Great – thanks for the insights, @bandages!