Yes, you need to instantiate a new AnimClip and TransformTrack.
Not sure why @Nehon designed it like this in the new animation system.
an example code for retargeting an AnimClip from one model to another model supposing both have the same rig :
Spatial model = assetManager.loadModel("models/character/model-1/model.gltf");
AnimComposer ac = getAnimComposer(model);
AnimClip clip = ac.getAnimClip("walk");
rootNode.attachChild(model);
model = assetManager.loadModel("models/character/model-3/model.gltf");
ac = getAnimComposer(model);
SkinningControl sc = ac.getSpatial().getControl(SkinningControl.class);
SafeArrayList<AnimTrack> tracks = new SafeArrayList<>(AnimTrack.class);
for (AnimTrack track : clip.getTracks()) {
TransformTrack tt = (TransformTrack) track;
HasLocalTransform target = null;
if (tt.getTarget() instanceof Node) {
target = getAnimRoot(model);
} else if (tt.getTarget() instanceof Joint) {
Joint joint = (Joint) tt.getTarget();
target = sc.getArmature().getJoint(joint.getName());
}
TransformTrack newTrack = new TransformTrack(target, tt.getTimes(), tt.getTranslations(), tt.getRotations(), tt.getScales());
tracks.add(newTrack);
}
clip = new AnimClip("walk");
clip.setTracks(tracks.getArray());
ac.addAnimClip(clip);
ac.setCurrentAction("walk");
model.move(2, 0, 0);
rootNode.attachChild(model);
Maybe we can introduce a new constructor and getter/setter for directly setting :
CompactVector3Array translations,
CompactQuaternionArray rotations,
CompactVector3Array scales
in TransformTrack:
so that we can directly share the compact translation, rotation, scales between Tracks without unpacking /repacking them. @pspeed, @sgold any thought?
Edit:
Okay, never mind. Using TransformTrack.jmeClone() does already share those compact transforms and times. We just need to reset the target to reference new Joint after cloneing. Sorry for my ignorance 