As a suggestion to hep the new users get better acquainted, how about a wizard to help them with workflow?

I think this is possibly the most awkward as there are a number of things that might be forgotten.

Doesn't have to be complex, just guide them through the steps

1. Calibration (load video), calibrate,
2. Load main video,
3. Load saved calibration scene,
4. Set up actor properties
5. Centre coordinates
7. Set up in/out region
8. Track

etc etc.

I know that order of some of the above isn't important as long as they're all done before pressing the "Track" button but it's because of that, I think it would help those testing to see quicker results and for you ideally another sale from a successful trial.

Plus I wouldn't mind if it said "open AVI", then popped a dialog for calibration xml, then popped a dialog for actor

I really like the idea of a wizard for processing mocap and I hope you'll do it.

The only thing to add is that I'm still bogged down at the calibration stage and could use a similar wizard for that.

As I mentioned, I have nice green calibration noodle but now I need to define the ground plane in the video.

When I try to process ground plane points, the DMC just goes into some indeterminate state during that process and I don't know what to do about it.

There's not even an error report, which would help.

I don't really have a lot of time to do intricate troubleshooting, so I get stymied.

It sounds to me like maybe you don't have suitably placed "light" marks on the ground? In one of greenlaws videos, he does a very simple method which I can best describe as like one of those 4 legged gazebo frames.

In other words draw your calibration light down one (imaginary) leg of the frame and back up again, move forward, down to the floor and up again, move across, down and up again until you've marked the 4 corners of the capture volume. What you then do after that by doing spirals up and down or just waving around or whatever I suppose is up to you with whatever results but I always do this now, up and down at each corner and I have no problem finding ground. In my early calibrations I didn't get the floor marking at the corners right at all and just ended up with a front ground part and that wasn't enough for the software to deal with.

Since then, i've had no probems with the above method.

Hello!

Good point! I hope we'll have opportunity to implement it. Meanwhile we are concentrated on tracking algorithm for Kinect and on performance and accuracy issues of current tracking algorithm for multi-camera video.

Thank you for your valuable suggestions.