Full body mocap / facial mocap with Blender ?

I watched some vids on YouTube where folks put dots on their faces, recorded vids (to my understanding) and then captured motion to facial rig using object tracking (I think that’s what it was). There are no updated tutorials unfortunately on the subject. So I am wondering if full body mocap is doable following the same technique, namely attaching some kind of markers to the body, recording video, tracking markers, etc.

How would one set up rig for face and body of the character in Blender for such mocap? Also, what would be the general workflow for this technique ?


If we can get multi-cam solving then perhaps full-body motion capture could work. I doubt in real-time, or very easily at that though.

I can’t find the subscription link here right now, but a few months ago someone had an awesome rig they put together for facial mocap they put up here and on facebook. I knew I should’ve saved the link to my toolbar.

I’ll see if I can find it later. But for now, here is one I found right away

@@Daniel8488: It doesn’t have to be real-time. Just as long as it’s captured and applied to a rig.

@@Craig_Jones: Yeah, that’s the vid I watched. However I don’t recall seeing any .blend supplied :frowning:

That’s the thing, I need to find the one that was here a few months ago, because he had a rig available for the facial capture


I think this thread is the one I remembered, maybe it can help you out towards your end goal, motorsep

Thanks! I kinda grasp the idea how, just need to figure out tracking :slight_smile: (haven’t tried it yet)

Hi Motorsep,

your idea is great. I am also trying the similar thing. How about your progress?

No progress unfortunately. Not only I need 2 Kinect’s (and my suspicion is that it’s better to use more HD webcams instead of 2 Kinects), but also commercial software that stabilizes the input, otherwise there will be a lot of cleanups to do.

As for facial capture, there is http://www.faceshift.com/ and I think it worth a shot.