Skip to content

Dynamic Weight Dynamic Weight

The joint is frozen; only the painted field is animated.

Node type RigExecDynamicWeight
Example dynamic_weight.usda

On this page:

Overview

Dynamic Weight effect

Paint that moves: a dynamic weight takes a static base field and reshapes it every frame from an animated driver — a muscle that engages with effort, a corrective that fades with a pose, a knee influence that breathes with the stride. Same target contract as the field it wraps: it declares the same rigExec:weightTarget and is bound in the mover's place of the field it modulates.

Fixed-shape weight field recomputed from statically declared inputs, optionally modulating a base weight object: r_i = (b_i * d) * s + a; w_i = RangePolicy(r_i). Without rigExec:baseWeight, b_i = 1 and the field must be constant.

How it works

The compiler binds the mover's rigExec:weightObject to one computeWeightPacket for the whole compiled rig, and every generation re-evaluates that packet from inputs:driver, inputs:scale, and inputs:bias as ordinary dynamic inputs before the mover walk reads it at the consuming revision. The base field resolves first — the rigExec:baseWeight relationship pulls that object's own packet — then each element becomes (b × inputs:driver) × inputs:scale + inputs:bias — inputs:scale and inputs:bias apply after the driver is combined in, not to the driver — and rigExec:rangePolicy either rejects a result outside [0, 1] (strict) or clamps it (clamp). The resolved envelope is handed to the mover and, when the influence overlay is armed, published as the field a rigger is shown; nothing is written back to the stage.

Wiring

Relationship Points to Required
rigExec:baseWeight Weight field this one modulates; required unless the representation is constant, and at most one target. yes
rigExec:weightTarget Exact points property this field covers; must match the target of the mover that binds it. yes
(bound by) A mover's rigExec:weightObject applies this field. -

Parameters

Weight field

guide:weightDisplay

Type: uniform token. Default: "tint".

Valid values: tint, heatmap, bands.

Selected influence overlay: tint blends gray to red; heatmap interpolates dark blue, blue, teal, yellow and red at weights 0, .25, .5, .75 and 1. bands assigns five colors in .2-wide bins. Colors use resolved vertex weights, clamped to [0,1] for display; faces interpolate them, so bands are approximate on coarse meshes. Does not change weights, materials or opacity on the source stage.

rigExec:weightTarget

Relationship.

Canonical prim or exact property carrying the weighted domain. For a constant envelope over an atomic multi-target mover, this is the mover prim itself.

rigExec:representation

Type: uniform token. Default: "constant".

Valid values: constant, dense, sparse.

rigExec:rangePolicy

Type: uniform token. Default: "strict".

Valid values: strict, clamp.

Node parameters

rigExec:baseWeight

Relationship.

rigExec:operation

Type: uniform token. Default: "multiply".

Valid values: multiply.

inputs:driver

Type: float. Default: 1.

inputs:scale

Type: float. Default: 1.

inputs:bias

Type: float. Default: 0.

Example

A 24 × 6 strip is skinned to one joint that holds a fixed 35-degree bend, and the only animation in the file is inputs:driver sweeping 0.15 → 1 → 0.15. The base paint is a straight ramp from 0 at the one-sixth mark to 1 at the tip, and with inputs:scale = 1 the driver simply scales that whole field at once: every painted point follows the same fraction of the bend at every moment, so the grey-to-red ramp brightens and dims — and the strip curls further and relaxes — while the control and the joint never move. The docs renderer tints the strip by the field the mover actually consumed — grey at weight 0, red at weight 1 — so the ramp and the curl that follows it are the same event.

Open it live with:

bin\launch_usdview.bat docs\examples\dynamic_weight.usda

Re-render the GIF above with:

python docs/render_media.py --page dynamic_weight

Tips

  • rangePolicy: clamp is what lets inputs:scale overdrive the paint: a result outside [0, 1] is clamped instead of invalidating the field and passing the mover's points through, so a driver can push part of a painted ramp to fully followed while the rest of it still fades.
  • The influence overlay paints the field a mover consumed, so point it at the dynamic weight, not the base it wraps: an unbound base has no resolved field of its own to show.
  • Drive the driver from another channel (a float math mover or a connection) to tie corrective strength to posing; each of driver, scale, and bias takes at most one float connection. A connection to a channel that float math movers revise reads its authored base unless the input declares rigExecReadPhase = "final".

See also


UsdRig