Collada animation data

Well, first the number and order of the input tags is important:




This tells you, that each vertex data consists of a pair of indicies, a index of the joint and the index of an according weight.

Next you need to know if a vertex depends on one or more bones. This is encoded in the vcount-array. In your example vertex 1 to 12 depends on just one bone(=joint), the rest on two bones.

Well, with this knowledge you could decode the v-array in a sequential manner.

0 1 0 2 0 3 0 4 1 5 1 6 1 7 1 8 0 9 0 10 0 11 0 12 0 13 1 14 0 15 1 16 0 17 1 18 0 19 1 20
=> divide it into (joint,weight) pairs
(0 1) (0 2) (0 3) (0 4) (1 5) (1 6) (1 7) (1 8) (0 9) (0 10) (0 11) (0 12) (0 13) (1 14) (0 15) (1 16) (0 17) (1 18) (0 19) (1 20)
=> consider number of bone dependencies
[(0 1)] [(0 2)] [(0 3)] [(0 4)] [(1 5)] [(1 6)] [(1 7)] [(1 8)] [(0 9)] [(0 10)] [(0 11)] [(0 12)] [(0 13) (1 14)] [(0 15) (1 16)] [(0 17) (1 18)] [(0 19) (1 20)]

As you can see, you got 16 sets marked with [], for each vertex one(vertex count=16). These are all indicies, you have to look em up in the according joint/weight arrays to get the concret values. Take a look at the first and the last vertex.

The first vertex depends only on joint 0 and on weight 1, which will be 1.0. Whereas the last vertex depends on joint 0 and 1 with the according weights 19 and 20, which will sum up to 1.0.

Thats all ^^

--
Ashaman

Quote:

-Warning: link Bone01's mode is not supported. It will be treated as mode eTOTAL1. -Warning: link Bone02's mode is not supported. It will be treated as mode eTOTAL1.

Can't help you with that. Maybe ask the export plugin's developer?

[/quote]
One of the first things i need to parse is

<vertex_weights count="16"><input semantic="JOINT" offset="0" source="#Box01Controller-Joints"/><input semantic="WEIGHT" offset="1" source="#Box01Controller-Weights"/><vcount>1 1 1 1 1 1 1 1 1 1 1 1 2 2 2 2</vcount><v>0 1 0 2 0 3 0 4 1 5 1 6 1 7 1 8 0 9 0 10 0 11 0 12 0 13 1 14 0 15 1 16 0 17 1 18 0 19 1 20</v></vertex_weights>

[/quote]

vcount tells you the number of bones affecting each vertex. The last four vertices are influences by 2 bones each, the other are only moved by a single bone.

v now tells you which bones affect each vertex. The JOINT semantic has offset 0, so the first number is always the bone index. The WEIGHT semantic has offset 1, so the second number is an index into the weights table. The v element holds a number of such index pairs for each vertex, with the count of pairs stated in the vcount element. Read it like this:

Vertex 1 is affected by 1 bone, namely bone 0 with the weight from weight_table[1]. Vertex 2 is affected by 1 bone, namely bone 0 with weight 2. And so on... Vertex 12 is influenced by 2 bones (see vcount), the first one being bone 0 with weight 13, the second one being bone 1 with weight 14.

HTH

Bye, Thomas

That's not animation data, that's skinning data, if you're not doing so already check the COLLADA spec document, it should tell you how to interpret all the data as well as what to expect where.

If you are able to read C#, Remi from the Collada team (i suspect) did a good job on a parser, allthough it doesn't fully work as the specs say it should. Gives you a good indication though:

link