PS eye vs Kinect

madmike1 · **Joined:** Mon Jan 12, 2015 9:52 am **Posts:** 1

I am wondering what usually gets the better results?

I been told Kinects don't require as much space, but I see so many videos involving the PS eyes, is that the most accurate? It looks like the eyes require more set up, but it is the most smooth and require less clean up. But I have not tried, at the moment I bought 2 kinects to try motion capture for the 1st time intill the PrioVR comes out.

But I am just wondering the advantages and disadvantages of the PS eye and Kinect. I personally want to do fighting animations, which requires me to hit the ground a lot.

Greenlaw · **Posted:** Mon Jan 12, 2015 5:37 pm

Each system has it's own strengths and weaknesses. Here are the main considerations:

- PS3 Eye allows for much larger capture space. However, it also requires more space for setup.
- Kinect is easier to setup as it doesn't require special clothing or lighting.
- 2 Kinect sensors produce about the same quality as 4 PS3 Eye cameras.
- 3 Kinect sensors produce about the same quality as 6 PS3 Eye cameras. You still have a much smaller capture space though.
- Kinect capture can be sensitive to IR pollution so you generally want to avoid a room that's brightly lit by sunlight or has a lot of reflective surfaces (the floor especially.) I usually record at night to avoid sunlight altogether, and when I need to record during the day, I will cover the windows with black out material.
- PS3 Eye needs fairly flat lighting, so you want to avoid harsh directional lights that case dark shadows on the subject. When I use PS3 Eyes, I like to light the subject using at least two soft boxes. If you shoot outdoors, you want to shoot on a overcast day or very early in the morning before the sun is too high.

Personally, I like the Kinect setup because it's so convenient for me that I can use it in our home studio. That said, I've kept my PS3 Eye setup just in case I ever need to record motions that cover a large area. In the latter case, I shoot outdoors because our indoor space is too small for PS3 Eye.

This question comes up often so if you do a search in the forums, you'll find more detailed info info for comparison.

Hope this helps.

G.

Greenlaw · **Posted:** Mon Jan 12, 2015 5:48 pm

A little more info: For fast moving fighting motions, PS3 Eye might produce more accurate results because you're recording at 60fps. Also, having full coverage is probably necessary if you're going to be falling down a lot.

In general, however, I haven't had too much trouble with fast motions using Kinect. Here's a capture of my daughter when she was five that was done a couple of years ago--I think it shows what's possible with dual Kinect recording even if it isn't as accurate as using six or more PS3 Eyes.

Mocap Test with Small Child

The iPi system is really quite good regardless of which configuration you go with. If you can only have one setup, I'd got with the one that gives you the most flexibility. (In my case, that would be 2 to 3 Kinect sensors.)

Hope this helps.

G.

Greenlaw · **Posted:** Mon Jan 12, 2015 5:58 pm

Some final notes, if you go with Kinect:

I only recently started using the triple Kinect setup. The main advantage, of course, is that three Kinects will give you full coverage of the action, which means less chance of occlusion. This could be very important if there is a lot of turning around.

I'm finding that triple Kinect is a little more difficult to calibrate though. It requires using a maglite, which is harder to see in my recording space during daylight hours. It seems to work fine at night though. With dual Kinect, I find using the flat board for calibration is much easier to use but you can't use the board method for triple Kinect.

For many actions, I'm finding that dual Kinect is often 'good enough'. It's easier to set up, and the data is easy to track when the action generally favors one side of the room (although it's certainly possible to record complex 360 motions with dual Kinect as shown in the above video.) For most actions, I stay with dual, especially if I need more capture space. (When using dual Kinect sensors, I usually have both Kinects at the front of the room facing the rear of the room.)

I'll switch to triple Kinect when a particular motion is giving me a lot of occlusion issues. However, because this requires the third Kinect to be placed across the room, it does reduce my capture space.

TBH, I haven't used the PS3 Eye setup for quite some time so I really can't comment much further about it. Hopefully somebody with more recent experience will share their tips. :)

Hope this helps.

G.

PS eye vs Kinect

Who is online