Toolkit model loading only loads two bones

[Darklynx46]

I have tried using the toolkit for model loading and i have noticed the following:whatever the model is,the bones collection in the model class contains two bones only.I browsed the source code trying to find the one responsible for omitting the proper hierarchy but i couldn't spot the one responsible.So my question is:
-is what i have noticed the intended behaviour?Is the toolkit supposed to only use 2 bones for all models?
-if it is so,can you help me discover the culprit?I have checked the modelcompiler and the modelreader classes but they seem to load all the bones and not discard any of them.

[ArtiomCiumac]

SharpDX uses Assimp.NET to read model data during compilation - so you could check if your model loads correctly there (they have a demo application, afaik).

Also I am not sure about the correctness of the line 445 in ModelCompiler.cs - but I need a demo model to check if this is working correclty. Can you post one?

[Darklynx46]

I was using the models from the model rendering sample of SharpDX as reference.The dude model for example has 58 bones in its skeleton hierarchy and the Bones collection of the toolkit sample reports only two.

[ArtiomCiumac]

Thanks for pointing out, I will have a look at it.

2014-06-02 16:14 GMT+01:00 Darklynx46 notifications@github.com:

I was using the models from the model rendering sample of SharpDX as
reference.The dude model for example has 58 bones in its skeleton hierarchy
and the Bones collection of the toolkit sample reports only two.

—
Reply to this email directly or view it on GitHub
#405 (comment).

[jwollen]

Right now the compiler only creates bones for nodes that contain meshes, and their ancestors (see line 436). All other bones, including those used in skinning, are thrown away, as animation isn't yet supported.

If you are interested, I did a very rough implementation of skinning and animation in this fork. I'm hoping to do some improvements and refactoring in the near future.

@ArtiomCiumac: I think line 445 only collects nodes and assumes that the actual mapping is done separately (which happens for mesh nodes in line 480). For nodes used in skinning, this is currently not done, probably because skinning isn't fully implemented anyway.

[Darklynx46]

Thank you both,i have a clear understanding of where the discarding occurs now.Actually i don't need anything more besides the full skeleton hierarchy itself so i should get away with minor modifications.

[ArtiomCiumac]

@jwollen, your code looks nice. Why don't you create a pull request? Even it is a work in progress - it would be easier to track and help on what you are doing ;)

[jwollen]

Yep, I will do so shortly.

[xoofx]

Follow on Issue #407