THREE.JS / FOUNDATIONS 98

One rig. Three kinds of motion.

Compare action crossfades, simultaneous normal weights, and a converted additive clip on the same RobotExpressive rig.

Initializing -

Preparing three independent mixers and their animation actions.

Clip inputsshared source / cloned actions

Base clip Target clip Additive clip

Each lane owns a mixer. Only the additive lane converts its overlay clip into relative keyframes.

speed

Crossfade lane-

Duration1.00s

Warp time scales to match clip lengths Wait for base loop (Idle transitions immediately)

The destination action is reset, played, then blended in with AnimationAction.crossFadeTo().

Weighted lane-

Target weight0.35

Both normal actions play at once; the visible pose is the weighted result of their shared property bindings.

Additive lane-

Overlay weight0.65

Static target pose weight0.00

No base action (retain both overlays)

Reference frame0

Animation action lifecycle

clone clips->clipAction->set weights->crossFade / update->render

AnimationAction weights are observable at runtime; additive conversion changes the clip blend mode.

Show skeletons

The three characters share the same model source but use separate mixer state. Add ?forceWebGL=1 for a WebGL2 comparison.

LAB 98 / ANIMATION BLENDING