OptionalchannelsThe channels whose joints the layer writes: the head and the trunk for gaze, breath and an affect display, an arm for a mood that lets it hang. A layer scheduled in a window during which one of the person's actions holds a limb channel it names by a contact (a grip on that hand, a planted foot of that leg) is refused when the performance is built; the head and the trunk bear no contact window, so naming them refuses nothing. Absent for a layer that does not say what it writes.
Stable layer name, unique per performer; keys the layer's draws.
The posture with this layer's contribution added.
The active action's posture, after earlier layers.
Film seconds.
The person and the performance's queries.
The layered posture.
OptionalblendEvaluate the contribution at the scheduled window weight, where zero returns the incoming posture and one applies the full layer. A dependent target can follow the joint geometry the provider produces at that weight.
An additive layer over whatever action a person performs: gaze, breathing, an affect display. A layer never starts, cuts or ends an action; it takes the active action's posture at a film second and returns it with its own contribution added, the way layered animation adds an additive pose on top of a base locomotion (Unreal's layered blend per bone, Unity's additive animator layers).
A layer must keep every planted or gripping limb goal unchanged, because contacts are re-solved from those goals and must not slide; it should change joints a body action leaves free (the eyes, the head, the chest's breathing) or add small offsets to the root and the trunk. Its draws come from
random, keyed by the performer and the layer's name.A layer runs for the whole film, or inside a window the performance is given with it (IAutoMovieMotionHumanLayer.IScheduled): from and to a film second or an instant of an action or of a cue track, entered and left over a fade so neither end steps.