Blend Shape

Definition

Blend shapes—also called morph targets—are predefined variations of a base mesh where specific vertices are repositioned to create expressions, phonemes or shape changes. Animators interpolate between these targets via slider controls, enabling smooth transitions from neutral poses to smiles, frowns or other deformations.

Origins & Adoption

Blend shapes date back to 1990s character animation, when studios sought more control over facial performance than skeletal rigs alone could provide. By storing multiple mesh snapshots—e.g., jaw open, eyebrows raised—artists could mix and match expressions in real time. Today’s pipelines integrate blend shapes with rig controls and corrective shapes for dynamic realism.

Workflow

1. Pose Modeling: Modelers duplicate the base mesh and sculpt target shapes for each desired expression or phoneme.
2. Shape Connections: Targets are linked to the original mesh, with each assigned an attribute slider (0–100%).
3. Animation Curves: Animators keyframe blend-shape sliders to choreograph facial dialogue and emotional beats.
4. Driver Systems: Automated drivers sync lip-sync shapes to audio waveforms, reducing manual keying.

Applications

• Lip-Sync Animation: Mapping phoneme targets to dialogue tracks for accurate speech.
• Emotive Performance: Layering multiple blend shapes for nuanced character acting.
• Non-Human Morphs: Transforming creatures, vehicles or abstract forms by blending shape sets.

Trivia

• Some advanced rigs combine blend shapes with muscle-based simulation for ultra-realistic skin deformation.
• Game engines like Unreal support GPU-accelerated blend-shape interpolation for high-fidelity real-time characters.