OptionaldriverPose-driven weight; null or omission leaves the weight to expressions. A
driven target cannot share its name with an expression preset or ARKit
channel, ignoring case, because the viewer writes those names from the
expression after the pose.
Nonempty name, unique within the mesh. Expressions write ARKit channel and preset names, matching exactly first and then ignoring case.
Flat normal deltas [dx,dy,dz,...], one triple per mesh vertex, or null
to leave normals unchanged by this target.
Flat position deltas [dx,dy,dz,...] in metres, one triple per mesh
vertex, in the mesh's own frame.
One morph target of a mesh: per-vertex displacements added to the rest geometry in proportion to the target's weight, before skinning (the glTF morph target model with relative deltas).
The weight comes from an expression, which writes it by name, or from
driver, which follows the pose. A mesh's deformed position isp + sum(w_t * positions_t)and its normaln + sum(w_t * normals_t), then skinned.