I exported a skinned mesh with Max 2011, using the Autodesk collada exporter and there is no skeleton node at all in the file. Is it possible to have a skinned mesh without a skeleton node?
Yes it’s possible to have such a document although in your case it doesn’t seem like what you are trying to do.
Check your exporter options and make sure you have the right settings.
It’s also possible that the rigging you have made is not right as yet.
I don’t know how you would find the root joint without the skeleton node. There is definately nothing I’m doing wrong.
Anyways, I discovered that OpenCollada supports Max 2011, so I’m going to use that instead.
What do you mean by skeleton node?
If there is no skeleton element in the instance_controller then the source for the JOINT input, containing IDREF_array values, could still be resolved. It’s just less reusable with other instances.
I don’t follow you. What are “IDREF_array” values?
Well, looking into it a bit more closely, you could assume that the first bone in the “skin joints” array is the root bone, but this would be just an assumption and it wouldn’t work in the case when the skinned mesh has more than one root joint.
Content that omits the <skeleton> element from <instance_controller> is going to be less flexible since the binding is to IDs that unique within the document rather then to SIDs that are relative.
Also note that each instance can specify zero or more skeletons. Roots of the node hierarchy are self evident within the document structure.
If by chance you are saying that the exporter is not exporting any node hierarchy at all, in your case, then sounds like a question to ask at Autodesk.
It does export the joints in the visual_scene. All I’m saying is that there is no way of knowing which one is the root. Some exporters export a skeleton where the root joint is a child of another joint.
Anyways, I’m just going to use the OpenCollada exporter. Thanks for you help.
Just give the joint nodes name attribute values that are meaningful to you like “skeleton-root” or “arm-ik-root”.
Yes it’s true, in terms of COLLADA data model, that all <node> elements are equal in semantics. There are no special cases.