I had some problems with power outages and surges recently that resulted in my power supply getting fried. Thankfully, the rest of my PC survived and the PSU was replaced for free thanks to a 3-year warranty, but the whole thing ended with my TrackIR 4 camera going dead.
I was of course considerably annoyed (to put it mildly
) because i didn't want to buy a new set, but on the other hand head tracking is one of those things that once you get used to you can't go back.
So i installed freetrack and after some tweaking i can use it with a cheap microsoft lifecam vx-5000 and my battered trackclip Pro (that thing is so notoriously flimsy that i've had to use duct-tape a mere 2 months after i bought my TrackIR+trackclip Pro set).
I use my webcam to place skype videocalls to relatives abroad, so i change the camera's configuration back and forth whenever i want to use freetrack.
Also for the same reason i haven't modified it (didn't remove the IR filter), but the IR leds from the trackcklip still register if the room is dark.
Granted, my initial observations are a bit tinted by the fact that i haven't tweaked freetrack to my taste just yet, but truth be told if i could spare the money right now i'd buy a new TrackIR set.
Don't get me wrong, Freetrack is a very good solution and i hope it's supported in CoD. However, coming from 2-3 years of TrackIR use it seems to me that TrackIR is definitely better (or just something i'm used to).
My comparison will focus mainly on the things that i think trackIR does better, not because i want to diss freetrack but because if it was equally good i'd be just saying "omg, it's as good as trackIR and free, definitely prefer it!" and be done with it
For me, the points i personally disliked were:
1) There's a noticeable delay between moving my head and freetrack doing its job. I guess this is not an issue for someone who's new to head tracking or only used freetrack in the past, but for someone who's used TrackIR and got spoiled by its near instantaneous response the delay is noticeable.
2) The freetrack software is definitely more complicated than TrackIR version 5 software, but it's more or less on par with the older trackIR version 4 versions. I guess that's the price one pays for the ability to customize, more parameters and so on.
3) The keys i use to control freetrack sometimes have trouble registering. I use F9 to pause it and F12 to recenter. Sometimes i have to keep pressing 2-4 times before it does what it's supposed to do. The biggest drawback in this is that if you pause it, it seems to also deactivate the camera so when you resume it you have to wait for the webcam to also do its own "start-up". It's not a century of delay of course, but it's a noticeable and often crucial few seconds.
With trackIR, i could point my head into a weird angle, pause the device at that spot and turn my head straight on the screen to keep looking that way without straining my neck, then i would press F9 again and it would instantly start following my movement again.
With freetrack, i find that it will probably take 2-3 attempts to do it and then i'll also have to wait a few seconds for the webcam to come back to life.
4) Trying to set a proper curve for one axis in Freetrack is driving me nuts. The trackIR method (especially in the version 5 software) is definitely superior by leaps and bounds.
5) Accuracy is not the same, even when using 640x480 webcam video resolution. I also have a steady 30FPS with 0 jitter and the freetrack software clearly shows 3 dots being tracked. I won't hold this last one against it however, because i think it could be improved with some tweaking on my part.
I get a lot of "blind" angles, so for example if i want to look to my 5 o'clock the movement is smooth until a certain point, i get a blind spot where it loses contact and then it picks it up again. I think this could be solved by making my own custom IR leds and sanding them down a bit so they are easier to pick up, plus taping some photo film in front of my webcam to act as a visible light filter.
6) I get motion in an extra axis during certain movements. If i'm looking straight up the in-game view also rotates to clockwise some, if i'm looking down into the cockpit it rotates counter-clockwise. TrackIR had a bit of this too in the version4 software but definitely not so pronounced and it was totally eradicated for me with the version 5 drivers.
I think this is also a point that can be solved with some tweaking, as i think it has to do with setting the distances of the reference point in the freetrack interface.
Finally, my thoughts on a few improvements that could make freetrack even better.
1) Don't turn off the webcam when pause key is pressed, leave it running, keep tracking it but just stop the data stream towards the game. This would solve the wasted seconds between pause and resume.
2) A revamped interface for setting up the sensitivity curves. Keep the old one for those that might prefer it, plus copy the trackIR method as an alternative way to do it.
3) Have a way to automatically calibrate the reference point so we don't have to take a measuring tape to our heads
A 3-step calibration could be used, where the player needs to look straight ahead, all the way up and all the way down. The software could then extrapolate the reference point datum by measuring the rotational offset that's caused when looking up/down, because it knows it shouldn't be there.
My final assesment is that freetrack is not as good as trackIR in an absolute comparison, but it's better on a cost-effectiveness basis. If you want to get head tracking on the cheap definitely try it out, just be prepared to spend time configuring it and having to deal with a few frustrating aspects.