AutoMovie API
    Preparing search index...

    Group each leg's stance mask into maximal inclusive runs and assign one fixed model-frame contact per run. A run beginning at frame zero resumes an authoritative opening target; otherwise its first foot supplies XZ and sampled ground supplies Y. Parallel times/FK arrays are caller-admitted.

    type IProps = {
        groundAt: (x: number, z: number) => number;
        legs: readonly IAutoMovieFootLeg[];
        openingTargets?: ReadonlyMap<AutoMovieHumanoidBone, IAutoMovieVector3>;
        resolved: ReadonlyArray<
            ReadonlyMap<AutoMovieHumanoidBone, IAutoMovieResolvedBone>,
        >;
        support?: IAutoMovieTimedGround;
        times: readonly number[];
        tolerance: number;
    }
    Index
    groundAt: (x: number, z: number) => number

    Ground height at a plan position.

    legs: readonly IAutoMovieFootLeg[]

    Legs whose feet are planted.

    openingTargets?: ReadonlyMap<AutoMovieHumanoidBone, IAutoMovieVector3>

    Prior beat's authoritative opening pin per foot, in model space.

    resolved: ReadonlyArray<
        ReadonlyMap<AutoMovieHumanoidBone, IAutoMovieResolvedBone>,
    >

    Per-frame FK bone lookups.

    Static terrain scoped on the local clock, when authored.

    times: readonly number[]

    Frame times, parallel to resolved.

    tolerance: number

    Contact tolerance above the ground counted as stance.