Morph Targets
Shape keys and blendshapes
Morph targets (blend shapes) store alternative vertex positions that are blended at runtime to produce deformations like facial expressions or corrective poses.
How it works
- Each target defines per-vertex deltas relative to the base mesh.
- A weight (0–1) scales each target’s contribution; multiple targets can combine additively.
- Exporters may bake targets per mesh or per-shape, and index them by name.
Use cases
- Facial animation (phonemes, expressions)
- Corrective shapes for joint-based skinning (elbow/knee fixes)
- Stylized deformation and secondary motion
Interoperability notes
- Vertex order must be stable between base and targets; topology changes invalidate targets.
- Some tools cap the number of simultaneous targets or precision of weights.
- Naming conventions for targets vary; keep names consistent for pipelines.
Limitations
- Memory grows with number of targets and vertex count.
- Large target sets can impact export/import times and runtime blending cost.
