I have my TrackIR3 and yes at first it's very disorienting and takes some getting used to. Also head tracking works the best being very close to the camera which is another plus for PC. (notice how close the man is from the screen/PSeye in the video) TrackIR sensitivity can be adjusted so you are not completely turning your head to the left to see left of your plane. It's more like you aim your head (nose) to the left side of the monitor to look left of the plane.

I doubt a developer would make a game where the player is running around their whole living room just to get a certain angle as there could get reports of people getting hurt.

I'm almost certain the average Wii owner has 2 Wii-motes and since head tracking is only a viable feature for a single user my statement is correct.  The point is that not many PS3 owners have the eye and most would have to buy it and almost exclusively for that purpose in most cases..a second Wii-Mote is something most Wii owners already have and probably want anyways if they don't have two.  My point is that if we are trying to figure out how much extra it cost my analysis is correct, if you want to know how much it cost from scratch then we can add in console prices etc...

@Fury

Just noticing that something is getting larger and smaller can be used to detect if it is getting closer or farther but doesn't actually provide depth.  With that said I do believe that for this application it can mimic it well enough to make the effect look good.  Both solutions are capable of this mimicry.  Also note that facial recognitiion is not perfect and it refines itself from frame to frame (as seen in the video), the result is that something like turning your head slightly could be sorely missinterpreted.  Needless to say the implementation of the Eye method is much more taxing in terms of development and CPU time, or in short in order to avoid missinterpretations (which is very doable) the software needs to be fairly robust which impacts development time and CPU load.

I want to be clear that I'm not going out of my way to favor one over the other here.  It is a simple matter of fact that two permanently afixed points on someone's head are easier to interpret in this application than attempting to accomplish the same thing through facial recognition.  I actually think the facial recognition method is likely how it will be done in the future as good software can allow for much more than just positioning. Things like facial expression could be used as input for example, so in games like LBP your SackBoy's facial expressions could mimic your own.  From that standpoint the technology is extremely promising and I want to be clear that I'm not implying its a bad way to go about it.  Just that this incarnation is not up to that quality in terms of hardware and in addition to that problem the current consoles do not have the horsepower to handle great graphics, with great AI, great physics, with facial recognition, and expression interpretation, etc, etc... even high end PCs would likely be extremely taxed to pull it off.

As for locking onto a face, see my point about a robust software suite being required.  Much of the problems are certainly solvable but again this is at the expense of CPU load and development time. 

And finally on the subject of purchasing the Eye by itself, I couldn't find it with my search at target, amazon, and walmart but I do remember hearing about it.   What is its price by itself?  I'm sure quite a few people would get it by itself rather than the whole package.

Unless you're thinking wearing a headband is fun...

2 point head traking on the wiimote is very limited for depth perception - it is incapable of determining if you turned you head sideways (which gives the appearance of the 2 points becomming closer) or actually moved forwards or-backwards (which also gives the same effect ), - the software would have to make a guess - it's a cheap solution and not even in the same league as real camera based head tracking systems >>>.
Real camera based head tracking can do good Six Axis tracking of head position - take a look at Toshibas work with their digital mirror software running on the Cell this is remarkable. The PS3's Cell processor is over 3* faster than that used by toshiba. PS3 could do this with processing power to spare !.

PS3 Eye is just a high speed camera but with the CELL it's capable of far more than just taking photos or video clips.

 rofl obviously your 9 years of age and still play mario kart racing for n64

I'm not entirely sure you read the thread before posting ...

As I've said both of these solutions have limited depth perception, and both face issues with the head turning. I also specifically mentioned that the camera methodology was the way the technology will work in the future as it has significant advantages, however I also pointed out the significant roadblocks currently preventing the facial recognition methodology from being viable in gaming.

Now it seems you've diverged from the topic a bit, you go off and talk about Toshiba's Digital Mirror technology and not the Eye. The Toshiba solution is extremely nice and does a whole hell of a lot more than either of these solutions but also happens to cost around $4,000 making it somewhat untennable as a gamming product until the price is dropped dramatically. Also note that this extreme price tag is indicative not only of the higher quality hardware used but I'm sure also includes covering the cost for the software required to make it work.

If I'm being honest I take exception to the way you wrote your post in an extremely missleading fashion. First you pointed out some of the flaws of both systems and only bothered to attribut them to the Wii-Mote solution when it's clearly true of the Eye solution as well. Next, you proceeded to refer to the Eye as "real camera based head tracking", by which you try to suggest that the Eye is somehow of equal quality to Toshibas technology. If you want to compare the Eye against the Wii-Mote that is one thing but surely you recognize that the Toshiba solution is far and away better than the Eye and doesn't even belong in this discussion as it is completely unuseable as a gaming product.

I wonder if the missleading nature of your post wasn't intentional. It certainly comes off as if you would like people to think that the Eye is capable of everything the Toshiba solution is. Perhaps I've missread this though and you can clear it up.

Finally, I think you've severely overestimated the capability of the PS3. It simply is not capable of running the robust Facial recognition software required to do half the things you think it can do while also rendering a game's graphics, calculating its phsyics, and handling AI. If facial recognition was just a matter of some cameras and a lightweight software package you would have already seen them in widespread use long ago. The software side of things is still a relatively new science and none of the existing practical applications involve any sort of rendered 3D environment much less a game. The closest I've seen is Toshibas solution which makes some real time modifications to a 2D image, but this is a world of difference from actually rendering a 3D scene.

If you think about it you are literally asking the box to both Render and Interpret two completely different 3D environments simultaneously. In the future this will be possible at a reasonable price, and as I said it will have some profound impacts on gaming but it simply isn't an economic possibility right now and that is precisely why the "cheap" solution of a Wii-Mote is viable right now...because its cheap easy and works on par with what would be possible through facial recognition system even on a significantly more powerful piece of hardware.

Actually, depth perception is quite easy to do on the Wii. The wiimote's IR camera can track up to four points, so adding two more LEDs to your head tracking cap can provide enough data to distinguish between a turned head and one which moves closer to the screen.

Even with two on either side of the head if you turn your head so that they appear to be very close together it could be interpretted as you being very far away.  Adding LEDs does help but then you can get into issues of setting standard arrangements etc..etc.. and with people making their own IR glasses for use it could get to be an issue in terms of programming as well.  I think its far easier to just use two and honestly now that I think about it how often do you turn your head away from the screen in a game that would actually benefit from HT?

The toshiba spurrs engine is a cut down CELL with 4 SPU's at 1.5Ghz to keep power low for laptop's while the CELL in PS3 is 7 SPU's at 3Ghz and the RSX GPU is probably also more powerfull than the GPU in the toshiba laptop too. Yes the PS3 is cheap but the CELL is quite capable of this application with plenty of processing power to spare ( One researcher claims voice recognition of 100's people simultaneously on just one CELL processor!).

FYI the digital mirror it's far more than a simple 2d coloriser. It does a complete 3D capture/reconstruction of a persons face and renders it as a 3D model, it also allows it to be modified it's complete real time 3D face capture and render system.

Software wise the digital mirror /and other research projects may be far from being packaged in a form which would allow developers to use them in games but the hardware in the PS3 is fully capable of supporting it if they wanted to right now. ( PS3 CELL is a popular with researchers who do intensive calculations )

Why dont we see this in games then ?? good question - my answer >>>

Games developers do not seem to take advantage of new research like this. Voice recognition for example is rarly if ever used in console games ?? yet many companies use voice recognition /synthesizer systems to answer customer phone calls etc. Games developers are very slow to recognise or take advantage of new software techniques like this even if they are quite possible to do.

The most innovative thing i have seen from a game developer lately was UT3 on PS3 allows the use of a mouse. WOW they must have worked overtime to get that feature to work and gave themselves a pat on the back - talk about innovative NOT.

Regarding using 4 points for tracking on the Wii -actually would need 4 points ON 3 independent planes for full 3D tracking and if any became obscured or mixed up it would loose track so such a helmet would be rather awkward looking and probably need at least 6 or more points.

Sorry i gave the impression that Wii head tracking is useless this is not my intention - it's is a usefull idea even with 2 points, it's simple and usable for some things.

However what ps3 eye could do for gaming on the PS3 is far more interesting and far greater posibilities than what an IR point tracker can do.

Thumbstick controllers are a very poor way to interact with a computer game IMHO and game developers are very slow to innovate. I am all for better ways to interact with computers like face recognition, voice recognition , head traking , motion sensing controllers , etc

It's pretty sad when games developers are so backwards they would rather the player remember various buttons to have their character say something when they could instead do Voice recogniton on machines like the PS3 , Xbox360 and mabey even the Wii.

 Actually, I'm thinking of a third point placed high on the forehead, forming a triangle. If the two points at the base get closer together while the height of the triangle remains constant, the user is turning his head. If the proportion of the triangle stays constant, but the distance between all points changes, the user is moving closer or further away. It wouldn't detect the direction of a change in pitch or yaw, but it could correct for such a change so it doesn't misinterpret it as depth change.

 I don't see this arrangement being hard to implement. I'm pretty sure any game with head tracking implemented could use a two-point system or a three-point system, the three-point system is just less subject to confusion about depth.