Layer Mixer

Overview

Combines two recording layers (RecordedLayer) into one. Merge motion recorded in separate takes, such as body sway and arm movement, before turning the result into a clip. This is the recording-layer counterpart of clip_mixer.

Usage tips

• When both inputs contain a layer for the same target, B replaces A; this is not motion blending. The behavior matches clip_mixer.
• This enables a divided workflow where each part is recorded in a separate take and combined later. Only the relevant take needs to be rerecorded.

Related nodes

• clip_mixer — equivalent processing at the clip stage
• record_take / pose_step_track — supply recordings to combine
• keyframe_clip — turns the combined result into a clip

Notes

• When layers with the same id overlap, the one from B takes priority.
• The current implementation combines layer documents; it does not perform additive composition per target.
• A collision means the exact same layer ID in JSON, not a matching display name. Only that ID from B replaces A, and output duration is the larger input duration.
• The only evaluation input is progress (normalized 0–1); the legacy absolute-seconds time pin has been removed. Connect Time Source to progress.