Trackmate Experiments

Noteput by Jonas Friedemann Heuer

Wed, 17 Feb 2010 12:23:00 GMT

Just ran across Noteput, a great idea for a tangible musical learning tool. Jonas' table uses the vvvv toolkit and fiducial markers rather than Trackmate, but it's an excellent practical use of similar technology.

I may also steal some of his lighting and reflector ideas (see the images above)... :)

Hiatus Over: Back On “Track”

Tue, 02 Feb 2010 07:35:00 GMT

After a bit of a break due to other commitments, I'm jump-starting my TrackMate project again. Quite a bit has happened since then; most notably a huge amount of progress on the Windows PS3Eye driver by Alex Popovich.

Having switched my OS to Windows 7 since the last time I worked on this, I downloaded and installed the latest driver this morning, and it functioned perfectly with no issues (see Trackmate screen shot above).

So, a good re-start then; watch this space for more progress.

P.S: I am aware that the post archive isn't functioning correctly, hopefully I'll get some time to fix it this week.

Lighting

Wed, 29 Jul 2009 13:08:00 GMT

Just a quick update: I have had a huge workload lately, but the project is still underway!

I've been experimenting with different lighting techniques. I'm currently looking into LED strips and some kind of diffuser, but it's largely just a process of trial and error.

Anyway, I will post more pics soon, and hopefully some video too.

Focusing Solved

Wed, 15 Jul 2009 19:57:00 GMT

After some more research and advice, I have managed to get the system working much more smoothly and consistently.

The first issue was the focusing: after some good advice from Adam (comment here) I flipped the frosted glass table top over so that the frosting was closer to the tag patterns. The difference was instantly noticeable: no more blurry tags!

The other mistake I had been making was assuming that the green dots marking the tag centres in debug view were a good thing, so I was adjusting the threshold control until all these dots turned green. It turns out that the brighter the green, the less certain the software is of the tag's position - once I realised this and adjusted the threshold the other way (until all the centres were dark), the tracking accuracy improved dramatically.

I am also running the table without its internal lighting at the moment; just boosting the gain on the webcam slightly and relying on the room lighting coming through the table top. This seems to work well, and the tracking works much better over the whole surface, which I think is due to the lighting being much more even. I can now experiment with some different ways of diffusing the internal lighting, or perhaps use a number of LEDs, to make sure the surface is receiving roughly the same amount of light all over.

I will try and get some video up as soon I can get hold of a video camera!

Focusing Issues?

Sat, 11 Jul 2009 22:17:00 GMT

One thing I've noticed is that in the screenshots in the walkthrough, the tags appear much, much sharper than in my setup. Here's a screenshot from the walkthrough:

And here's mine:

As you can see, the tags in my screenshot are quite blurred and dull, and I think this is harming the software's ability to track them.

There are a few possible causes for this: it could be that I have the camera at the wrong distance from the glass and so it can't focus properly, or the frosted glass I have used to make the table top might be too thick for this purpose. A bit more tweaking lies ahead!

Setting Up the PS3Eye

Sat, 11 Jul 2009 21:44:00 GMT

Well, I managed to turn off the auto-exposure on the PS3Eye, with help from Adam Kumpf, the author of the Trackmate software; there's a useful thread relating to inconsistent tracking on the Trackmate discussion forum.

After this, I discovered that the PS3EyeTest application is not the best app to use for setting up the camera - when you change settings (exposure, gain etc) they aren't immediately reflected in the preview window; you have to start and stop capture. Also the app is a tad flaky, so sometimes when you restart capturing you just get a black screen. All of this makes it near impossible to get the settings right.

However, using AMCap (which is also actually bundled with Alex P's Windows driver), you can adjust the settings with a real-time preview, then apply them once you're happy. So that's the way to go in future.

After a bit of tweaking, all the tags on the layout guide sheet are being recognised well. Here's a screen grab from Trackmate:

All those green circles with larger green dots in the middle indicate that the tag info is being picked up across the surface. And here's one with just a few tags placed on the table, which also seems to be working fine:

One small issue, though: the Trackmate software is only reporting a 5fps capture rate, whereas the camera is capable of 60. I will need to find out what's happening here and try to get it running at full speed.

PlayStation 3 Eye Camera

Thu, 09 Jul 2009 21:07:00 GMT

Having read up on some of the example Trackmate projects, it seems that many people are using the PS3 Eye webcam as their preferred camera for this application. I've picked one up from eBay and I can certainly say the quality is a massive step up from the Logitech. This camera will capture 640x480 video at 60 frames per second, which gives a very smooth and impressive image (the old camera got about 15fps - pretty jerky).

Obviously it's a Playstation 3 accessory, so to use it with Windows you need special driver software, which luckily, a clever chap called Alex Popovich has written for us.

I installed the driver, opened the supplied PS3Eye test app, and it worked perfectly. Until I tried to open the Trackmate application, which refused to recognise it and told me I didn't have a webcam installed...

This turned out to be due to the installer not correctly registering the PS3Eye.ax DirectShow Video Source component - I manually registered this using regsvr32, and Trackmate could find the camera with no problems.

I've only had a little time to play with the camera tonight, but I ran into a couple of issues. Although the picture quality is hugely improved and the tracking is much faster, it doesn't seem possible to turn off the automatic exposure, which means that other light sources (like the main light in my room) affect it badly. I couldn't get it to recognise individual tags properly at all, because as soon as I remove the test sheet the exposure dramatically changes and ruins the setup. Hopefully I'll solve this problem with a bit more research.

02/02/2010: Edited the driver link to point to Alex P's new site.

First Tests

Thu, 09 Jul 2009 20:53:00 GMT

The first attempts went well, but I discovered that the Logitech QuickCam E2500 webcam I'm using (see above) is nowhere near well-specified enough to keep up with rapid object movement over the surface. It also seems very difficult to get it in focus and to get the lighting right.

However, surprisingly (despite the fact that the Trackmate tags look like tiny blurry blobs on screen), the tracking is working, and I have managed to successfully control some of the example apps that come with the LusidOSC Processing Library download.

A good first step!

Hardware Build

Thu, 09 Jul 2009 19:59:00 GMT

The first stage was to build a basic enclosure for my table, which I did with a lot of help from my dad and his extensive selection of power tools (thanks, Dad!). I had a custom piece of frosted glass made, which cost about £12, and them we measured and built the wooden enclosure around that.

The inner pieces are pretty basic: just a standard mains bulb holder and a 60w energy saving bulb (I wired two of these in just in case I needed more light), a mains junction box, an inline power switch on the cable, and a Logitech Quickcam E2500 webcam, mounted on a cross member about half way up the sides (see the picture above for details).

Trackmate + LusidOSC Sequencer

Wed, 08 Jul 2009 09:47:00 GMT

This is the video that first got me interested in the Trackmate system: here the table is controlling a basic sequencer application written with Processing.