Head Tracking with Cliffs of Dover

[CharveL]

Good luck with that.

I think it's cool and all what the MS kinect promises, and eventually will offer but don't expect it to work all that well for flight sims....yet anyway.

First off, I would wonder what the refresh rate on the kinect is? Even at 60fps that's only half of what a TIR offers, and yes it makes a LOT of difference in game.

Secondly, with all of that processing to recognize your head and how it's moving I would also wonder what the lag will be like. TIR is almost instantaneous the lower you keep the smoothing (processing) and I would imagine its much worse on an interface that has to figure out where your nose is in relation to the rest of your face. Then again, FT users are probably used to having more input lag than a dedicated TIR unit but I will concede that it might not be that bad having never tried FT and contend that it will likely suck on a kinect.

Thirdly, the kinect is made for console gamers. Which means it's probably not built with fidelity and accuracy in mind. C'mon, we're talking about Call of Duty type gamers used to having the console autoaim for a 'close enough' satisfaction appeal.

Still, I really do like the idea of having a Kinect-like device that eschews the need for hats/clips and look forward to the day a good dedicated company like NP come out with one.

And NP is a small company with a great group of guys from my experience, who are always willing to personally go out of their way to help, not some cliched "corporate bigwig" stereotype. They put a great deal of effort into getting games to open up to head-tracking for a flight/racing sim market that is infinitesimal.

[LoBiSoMeM]

Originally Posted by CharveL

Good luck with that.

I think it's cool and all what the MS kinect promises, and eventually will offer but don't expect it to work all that well for flight sims....yet anyway.

First off, I would wonder what the refresh rate on the kinect is? Even at 60fps that's only half of what a TIR offers, and yes it makes a LOT of difference in game.

Secondly, with all of that processing to recognize your head and how it's moving I would also wonder what the lag will be like. TIR is almost instantaneous the lower you keep the smoothing (processing) and I would imagine its much worse on an interface that has to figure out where your nose is in relation to the rest of your face. Then again, FT users are probably used to having more input lag than a dedicated TIR unit but I will concede that it might not be that bad having never tried FT and contend that it will likely suck on a kinect.

Thirdly, the kinect is made for console gamers. Which means it's probably not built with fidelity and accuracy in mind. C'mon, we're talking about Call of Duty type gamers used to having the console autoaim for a 'close enough' satisfaction appeal.

You really don't understand that Kinetic-like solutions will not do the HT trick just scanning your head, but can do that scanning shoulders, arms, etc. And if you use some "markers", the 3-point calculation using Kinetic can be MORE accurate and fast as TIR, FT, etc.

In a flight sim it will be great in 6DoF HT, much more natural, with much more realistic movement. It's a totaly new approach, and we will see how it will work in near future. Isn't just for "Call Of Duty" games. Open your mind and think about possibilities: with Kinetic you can have the same 3-point HT PLUS really 3D motion capture.

[brando]

"...And NP is a small company with a great group of guys from my experience, who are always willing to personally go out of their way to help, not some cliched "corporate bigwig" stereotype. They put a great deal of effort into getting games to open up to head-tracking for a flight/racing sim market that is infinitesimal...."

Spot on. I first contacted NaturalPoint after I lost my arm, what, fifteen years ago? The support from NP was great as I struggled to learn how to use SmartNav and cope with my disability. If I wrote to them they always replied immediately and positively, and, although it might sound trite, I felt they were really 'there for me'.

Later, when I got into flight games to pass the time, I wrote and asked how to use my SmartNav with Warbirds and Dawn of Aces. I guess I wasn't the first to ask, because TrackIR came out about a month later. I bought it and off we went. I bought the 3Pro when it hit the shelves - Forgotten Battles was around by then and I really enjoyed the extra sensitivity and the Vector Clip. About a year ago I bought TrackIR 5 and gave my 6-year-old 3Pro to a new-to-FB pilot. He's loving it. I'm loving the 5 too, and the Track-clip's accuracy & convenience...

Watching this thread I've been surprised at the animus that seems to lurk in the argument of the anti-NP crowd. I don't give a flying f*** if someone wants to use a freetrack solution to achieve head-tracking, and I hope they will be able to in CloD. But why on earth do some people want NP to "go bankrupt"? NP exists by creating a kit that works 'straight out of the box' - useful to disabled people like me (and many others) and the sort of people who haven't the time or the aptitude to either assemble the equipment or write code.

Along with the NP hardware and software comes that same level of support that was there 'for me' fifteen years ago. When I had a problem I wrote an e-mail; and when America had woken up, I received a reply that put me straight. Somehow I don't think I could expect the same level of dedication from the freetrack crowd.

brando

[sigur_ros]

The true measure of a man is not how he treats his friends, but how he treats his enemies.

[Wolf_Rider]

Originally Posted by Blackdog_kt

Personally speaking, the bigger picture for me is not spending 150$ right now but when i feel like it, maybe a couple of months post release. What FT does with NP is between them and the longer we focus our energy on it, the longer a generic headtracking method will take to implement.

by the sounds of it, you've already spent $150 approx.
A consesus was reached here very early in the piece that games should be available for alternative forms of headtracking... the hack method of doing so is in question though. Solution - develop a clean product

Originally Posted by Blackdog_kt

I'm not hating on naturalpoint, i'm one of their customers in fact.

see point 1

Originally Posted by Blackdog_kt

However, i'm not going to take up their legal defence pro bono when
a) i'm not sufficiently qualified and
b) they are a big company making good money and they can afford the lawyers, seeing as how they didn't do anything up till now makes me think they don't have a reasonable chance of winning such a case

see second last paragraph

Originally Posted by Blackdog_kt

If we want the developer to have freedom in providing us with alternatives, then the developer must have a way of washing their hands clean. My solution was very simple and effective.

They give us the aforementioned interface and we decide what to do with it. They are not responsible for the way we use it.

see point 1

Originally Posted by Blackdog_kt

Saying that they are is like saying western digital should make sure their customers are not using their hard disks to store illegally dowloaded content: it's out of the scope of the business and totally non-enforceable.

subject for a different thread, not the one in hand

Originally Posted by Blackdog_kt

If a developer provided a specific, custom tailored interface for one particular head tracking method that was dubious then yes, they could face problems.

we have agreement on that

Originally Posted by Blackdog_kt

If they provide a generic instruction set that lets the end user take it from there, then they have no responsibility whatsoever...the user has it.

you may have mised the fact, there was a consensus reached very early on

Originally Posted by Blackdog_kt

I think it's the best solution either way we look at it. I don't want to have to wait for FT to settle their disputes with NP,

Thier disputes with NP, or NP's disputes with FT?
From you saying, it seems FT doesn't have much of a case either... how long has it all been going on for now? 2 - 3 years?

Originally Posted by Blackdog_kt

when i can get my buddy to code me an alternative in 2 evenings worth of time that will make use of a generic interface, plus in favor of community spirit i would gladly distribute it to the rest of the community as well.

How much support would be offered with your bud's ap?

* Edit

Originally Posted by LoBiSoMeM

You really don't understand that Kinetic-like solutions will not do the HT trick just scanning your head, but can do that scanning shoulders, arms, etc. And if you use some "markers", the 3-point calculation using Kinetic can be MORE accurate and fast as TIR, FT, etc.

In a flight sim it will be great in 6DoF HT, much more natural, with much more realistic movement. It's a totaly new approach, and we will see how it will work in near future. Isn't just for "Call Of Duty" games. Open your mind and think about possibilities: with Kinetic you can have the same 3-point HT PLUS really 3D motion capture.

.... you whinge about the cost of TIR (the whole argument for FT you present) but jump at paying for Kinect?

[CharveL]

Originally Posted by LoBiSoMeM

You really don't understand that Kinetic-like solutions will not do the HT trick just scanning your head, but can do that scanning shoulders, arms, etc. And if you use some "markers", the 3-point calculation using Kinetic can be MORE accurate and fast as TIR, FT, etc.

In a flight sim it will be great in 6DoF HT, much more natural, with much more realistic movement. It's a totaly new approach, and we will see how it will work in near future. Isn't just for "Call Of Duty" games. Open your mind and think about possibilities: with Kinetic you can have the same 3-point HT PLUS really 3D motion capture.

I would argue that I do, in fact, understand the "Kinetic-like" solutions better than you likely do. It takes a lot more processing power to figure out tracking points on a face or body than it does for 3 solid dots, and even more still to do so with any accuracy. I'm not saying it isn't possible just that I'm very skeptical for the first generation of these devices to be able to even come close to competing with TIR for the purpose of tracking head movement.

I do think this is the direction we're going though and it will be a good solution. Just not yet.

[Wolf_Rider]

its the kind of gizmo (after more development) which would work nice with a virtual reality helmet.
might be a bit hard to mount on top of the monitor for use as a headtracker, with its current size though

[LoBiSoMeM]

Originally Posted by Wolf_Rider

.... you whinge about the cost of TIR (the whole argument for FT you present) but jump at paying for Kinect?

YES!

Kinetic is some unique hardware unit that can do what a cheap wecam can't do, isn't TIR overpriced crap!

[Wolf_Rider]

lol

(if TIR is crap, why do you go all out to emulate it?)

[LoBiSoMeM]

Originally Posted by CharveL

I would argue that I do, in fact, understand the "Kinetic-like" solutions better than you likely do. It takes a lot more processing power to figure out tracking points on a face or body than it does for 3 solid dots, and even more still to do so with any accuracy. I'm not saying it isn't possible just that I'm very skeptical for the first generation of these devices to be able to even come close to competing with TIR for the purpose of tracking head movement.

I do think this is the direction we're going though and it will be a good solution. Just not yet.

"Kinetic-like" solutions can traking just 3 dots like our today HT solutions. They use a normal camera that can trace the "3 dots" easily. The WAY they track these dots is that make these kind of solution much more flexible, with the TOF camera. If someone want great response just tracking 3 dots, it's easy to do, it's just a small part of the solution, isn't?