Hi,

If i have two xbox one kinect and i plugged them into each computer, How do i record at the same time?

thanks

Hi
You should use the distributed recording feature.

Oh i see, thanks for the info.

Actually I need to understand one more thing. If my computer has two USB 3.0 slots, i can just plug my two xbox one kinect into each slot?

No, you can't. This is a limitation of Microsoft Kinect SDK 2. It allows to work with a single Kinect only.

While we're on this subject, I see how to correctly (I hope) position each Kinect in opposite diagonal corners of the room. Do they need to be at the same height?

I ask because it would be easier for me to put one a bit higher than the other (due to other things in the room) although I *can* put them at the same height if I need to. Also, it occurs to me that even if they are at the same height they can be angled up or down -- how important is it for the angle of each one to be the same?

I always position my Kinect devices at about 3 to 4 feet from the floor. For me, it's been better to make them the same height so I can cover the same area from different vantage points.

When you consider that each device is generating a point cloud that represents half of of a 'shell' of the performer, I think it's better to capture this shell 'head-on' so the software can try to assemble a whole 'shell' (or as much as possible anyway.)

That said, the angle and height you use may depend on the motion you want to capture. For example, if you're lying down, setting a Kinect at a higher angle may give you better overhead coverage than a lower angle would.

But in general, I think it's better to keep the height and pitch angles similar. The important thing is that the Kinect devices do not face each other head-on. This crossing rays can confuse the data.

The best way to know what works best for you is to try it out different configurations. With the Kinect system, it's quick and easy to experiment with different setups since there are fewer devices to deal with, and calibration can be easier. (With v1 I mean; calibration with multiple v2 is the same as with PS Eye cameras.)

Hope this helps.

G.

_________________
Greenlaw
Artist/Partner - Little Green Dog | My Demo Reels (2013,) (2015,) (2017,) and (2019)

Watch a one minute excerpt on Vimeo now!

As for angles, there's no set rule. Use the angle that gives you the most coverage of the performance. For example, tilting up may give you more coverage if you're jumping or raising your arms. Tilting down may give you wider coverage for your feet if you're walking around.

Sometimes, I may use switch to different configurations during a mocap recording session depending on the action I need to record.

Just be sure to document the changes so you know which calibration data to use. I keep notes in a Google Docs Sheet or a Open Office.org spreadsheet when recording a session. Having these notes have saved me a ton of time during tracking and animation, especially on projects with multiple characters.

G.

_________________
Greenlaw
Artist/Partner - Little Green Dog | My Demo Reels (2013,) (2015,) (2017,) and (2019)

Watch a one minute excerpt on Vimeo now!

Thanks, guys, for the info.

I have two Kinect V2 -- they can't face "head on", but you do want them (at the 180 degree setup) to face directly opposite, right? So if I imagine a diagonal line crossing between them, I want them to be parallel to that line, but "offset", right?

Also, am I right in assuming that the 180 degree setup would be less likely to get occlusion than to position them at 90 degrees? And, if so, is there any compelling reason to use them in any other configuration.

I'm still setting up my multi-Kinect v2 so I have yet to fully test this yet. However, I do have a few years of experience with multi-Kinect V1 setups.

As such, I'm not sure if this is true for v2 but with v1 you want to avoid facing the devices perfectly head on (i.e., exactly 180 degrees from each other.) The reason for this is that the IR rays from the far Kinect will get mixed in with the rays bouncing back from the near Kinect, which can cause errors in the depth data. If they are slightly off angle--say, up to 175 degrees, the errors are less likely to happen, and you're coverage can still be more or less complete.

Because of the size of my recording space, I'm usually shooting at 70 - 90 degrees, which works well for many situations. Occasionally, I will set up a '180' capture but the actual angle is more like 160 - 170 degrees.

Sometimes I will also set up for a triple v1 capture--with three, you can have the devices at narrower angles and still have full coverage, and lessen the likelihood of IR errors.

I'm guessing this is just as true for v2 but I don't have that experience yet to be sure. My understanding is that v2 is better at determining it's own data but the rules probably still apply somewhat.

G.

_________________
Greenlaw
Artist/Partner - Little Green Dog | My Demo Reels (2013,) (2015,) (2017,) and (2019)

Watch a one minute excerpt on Vimeo now!

Thanks again.

Based on what I'm reading, the main differences between V1 and V2 Kinect is the angle of view (so you can work closer, essentially).

I wish someone had a video somewhere showing the setting up of the light marker, as the explanation isn't quite clear. I understand I need to have the light visible to both sensors at the same time, and I understand I need to cover the area, but the whole "spiral down" thing and then the directions don't seem to line up. Do I start at the front top right, move over to the front top left, and then... what? How exactly do I spiral through the area? Or is it more like I start at the front top right, to to the front top left and circle around to the back top left, and then back top right and then keep spiraling down through the area until I reach the bottom? (This is one place where a video would be worth a million words).