What is skeletal animation?
Skeletal animation is a technique for animating a 3D mesh by binding it to a hierarchy of virtual bones and posing those bones over time. Each vertex on the mesh is influenced by one or more bones according to a set of skin weights, so when a bone moves the surrounding mesh deforms with it. The approach decouples the animation data (bone poses) from the geometry (mesh and weights), which typically makes characters and other deformable assets cheaper to animate, retarget, and stream than baking per-vertex animation directly.
The technique dominates real-time character animation in games, virtual production, AR/VR experiences, and web-based 3D. It is also the format most commonly used to ship animated content in interchange formats like glTF, FBX, and USD.
How skeletal animation works
A skeletal animation pipeline typically moves through four stages, from authoring to runtime playback:
- Rigging. An author builds a hierarchy of bones (sometimes called joints) that approximates the structure of the character or object. The hierarchy may include hundreds of bones for facial detail, fingers, hair, and clothing.
- Skinning. The mesh is bound to the skeleton, and each vertex receives a small set of skin weights describing how much each nearby bone influences it. Smooth skinning (linear blend skinning) is the most common approach; dual-quaternion skinning and corrective blend shapes are used when twists and shoulders need cleaner deformation.
- Posing and animation. Animators set keyframes on bone transforms over time, or motion capture data is retargeted onto the rig. Inverse kinematics (IK) solvers can derive bone poses from end-effector targets, which is useful for foot planting and hand reach.
- Runtime evaluation. At each frame, the renderer evaluates the current pose, computes the world-space transform of every bone, and applies the skinning math on the GPU (typically in a vertex shader) to produce deformed vertex positions. Blend trees and state machines combine multiple animations smoothly.
Where skeletal animation is used
Skeletal animation appears wherever a 3D asset needs to move organically. Common deployments include:
- Characters in games and virtual production. The dominant use case — humanoid and creature rigs animated by hand, motion capture, or procedural systems.
- Real-time avatars. Telepresence, social VR, and live-streamed avatars all rely on skeletal rigs driven by tracking data from headsets, controllers, or camera-based pose estimation.
- Animated product configurators. A configurator might use a skeletal rig to open and close a laptop, articulate a chair, or flex a shoe; the same rig drives both the marketing render and the interactive 3D viewer.
- XR and embodied agents. Spatial-computing experiences animate hands, controllers, and full-body avatars through skeletal hierarchies that react to user input in real time.
Skeletal animation in streaming and interchange
Because skeletal animation separates motion data from geometry, it is well-suited to streaming. Bone curves typically compress well, especially when sampled at lower rates than the render frame rate and reconstructed with interpolation on the client. glTF and USD both define standard ways to carry skin weights, bind poses, and animation tracks, which lets authoring tools and runtime engines exchange animated assets with minimal manual rework.
For large character libraries, runtime engines can also retarget a single animation across many skeletons that share a common bone layout. That makes it possible to ship one walk cycle that drives an entire roster of characters, rather than authoring separate animations per asset.
See also
Polygon mesh — The surface representation that skeletal animation deforms each frame.
Inverse kinematics (IK) — A solver that derives bone poses from end-effector targets such as foot or hand positions.
Motion capture — A common source of animation data that gets retargeted onto skeletal rigs.
Shader — The GPU program that applies skinning math each frame in the vertex stage.
glTF — The interchange format that carries skinned meshes, bind poses, and animation tracks for web and runtime delivery.
Blend shape — A complementary technique for facial and corrective deformation, often layered on top of skeletal skinning.
Additional resources
