Hello!

A few months ago I wrote a DAE/Collada parser that I used for creating skeletal animations which worked great! However for this project I decided to use Assimp's parser to avoid the pain of having to use my own.

When I load the bones it seems to work fine with smaller models like my test model with only two bones and 6 vertices. Although when I attempt to load my chicken model with 11 bones assimp only detects 9 of them as actual bones while the other missing two are still in the aiScene hierarchy as aiNodes without any bone data. They are leaf bones and in this case they are feet attached to the legs. The .blender file contains an animation and the two missing bones do have animation data and transformations applied to them so I don't think they are ignored for being ‘idle’, but they've still somehow gone afoot.

If anyone has any experience with this I will be very grateful if you can share your findings!

From my own experience with assimp I had a lot of issues with animation and collada.

Also, in general assimp tends to add its own bones during the import, so what you have is a bit strange.

Finally, don't forget that all animation is referred to by their names, therefore you find bones by names, not strictly by hierarchy.

_Silence_ said:

Also, in general assimp tends to add its own bones during the import, so what you have is a bit strange.

Far more likely it's the exporter adding the leaf bones (Blender has some weird settings by default), not the Assimp importer. I would reimport it into your modeling tool and check for yourself.