Attached is the code to calculate the node’s postition matrix.  The two calls of interest are:

GetLocalPositionFromTake - Computes the node position at a specific time in a given take
GetDefaultLocalPosition - Computes the node’s default position (set bUseDefaultPosition to true)

	Attachment
17479_HohkbfqDiCuE4LykNPGe.zip

I’ve been trying to get the local matrix for a while as well, and I ended up solving it by multiplying the parent node global matrix inverse by the child node’s global matrix.

clevijoki 23 March 2009 02:55 PM

I’ve been trying to get the local matrix for a while as well, and I ended up solving it by multiplying the parent node global matrix inverse by the child node’s global matrix.

That is a great solution.  Unfortuntely, this is far slower than the code I posted. The humanoid models takes several times longer to load.
There really should be a function that returns this. But this works and should be always correct.

hi there,

i am writing my first importer using fbx sdk and i am still a bit confused about a couple of issues.

After struggling a bit around transformations matrices, i came across this post,
and it solved a lot of problems. I have implemented a couple of functions,
following the example of

Whenever i convert a static scene, or an animated scene where the animations are attached directly to the geometries, the code works like a charm.

Eventually i tried an fbx exported from 3dmax. It’s a model that uses helpers to translate/rotate different geometries. The scene has it’s own hierarchy, something like:

mainHelper (translation, rotation)
|
|- groupHelper01 (translation, rotation)
| |
| |- geo01 (small translation)
| |- geo02
|
|- groupHelper02
| |

Now, with this file I can’t retrieve any correct transformation. My best guess is that i am missing something important about transformation combination. When i call
GetLocalPositionFromTake do i retrieve the tranformation matrix relative to the node parent? Or i retrieve the global tranform (and so, if i push all the transformations toghether i end up appling multiple times the same transformations).

Thanks for any help

edited:
solved it was just a trivial quaternion conversion problem. My fault....