forums.naturalpoint.com

Hi!

I�ve come across what may be an error in the "faceData.c3d" file from the latest Motionbuilder goodypack (v2).

The optical marker "C3D:Unlabeled3" is moving around in the scene, across all three of its axes. Meanwhile, no marker contains any rotation data, and so I am assuming that this marker is in fact the rotation marker, but with the rotation data erroneously stored as translations.

Is this indeed the case?

I'm trying to do the facial motion capture tutorial without access to an optitrack license at the moment, so instead of streaming the data from Arena using the plugin, I just opened the C3D file directly in Motionbuilder, where I saw this.

I'm a new user of Arena Expression and Motionbuilder 2010, so this may be trivial, but I thought it would be prudent to post here and verify.

Thank you!

Just an update, I now got access to our Optitrack Arena workstation, and so tested exporting the data in the C3D file from Expression again.

The same issue persists when opening the file into the 1.6.300 version, and then exporting into a new C3D file. What is more, checking and un-checking the option to disable head rotation makes no difference.

Thankfully, exporting the file as FBX produced data that is correctly interpreted in Motionbuilder 2010, so at least the incompatibility can be circumvented!

It turns out the tutorial cannot really be completed this way I realize, as the python script produces several errors about this and the other not being found, both with the original and updated C3D files, and the FBX created in Arena from the C3D as described above.

So off to understand what the script is doing

Hey Onar - any progress on this? I'm just encountering the same set of issues. Thanks for posting your findings

Hi!

Not really, I just reverted to not using the script at the moment.

I am new to facial animation so there is a lot to learn, python scripting is further down the list. Of course at one point I will need to understand that too, but now is not the time

What I do is simply constrain the opticals to the facial bones/clusters manually. It works, but it could be much improved of course.