I did Google it and found nothing definitive on latency. Hence, my asking.
-
Hey, guest user. Hope you're enjoying NeoGAF! Have you considered registering for an account? Come join us and add your take to the daily discourse.
With VR coming soon/soonish... Was MS investment on Kinect worth it?
- Thread starter Miroslav
- Start date
I did Google it and found nothing definitive on latency. Hence, my asking.
With Kinect 1 it's both frame rate and latency. With Kinect 2 latency is better but the frame rate is still low.
Nightengale
Member
Augmented reality.
Fortaleza is basically Google Glass + PSVita/3DS/smartphone AR tech without marker cards.
Kinect hasn't lost value from the VR angle. People are interested in solving the control problems for VR and see Kinect as a viable option.
Beyond VR Kinect has applications in robotics which is getting a ton of investment these days.
Kinect may lose out big time from being a tool in the living room though which was the primary point. TV manufacturers are beginning to install their own cameras into their displays making the Kinect seemingly superfluous. MS hasn't convinced people Xbox is good set top box+DVR replacement.
By losing it's primary value the investment doesn't seem worthwhile but Kinect has value in other industries that will help MS write off their living room mistakes.
Beyond VR Kinect has applications in robotics which is getting a ton of investment these days.
Kinect may lose out big time from being a tool in the living room though which was the primary point. TV manufacturers are beginning to install their own cameras into their displays making the Kinect seemingly superfluous. MS hasn't convinced people Xbox is good set top box+DVR replacement.
By losing it's primary value the investment doesn't seem worthwhile but Kinect has value in other industries that will help MS write off their living room mistakes.
I believe you're thinking of Fortaleza. That roadmap showed a path to Augmented Reality, not Virtual Reality. There's a big difference there.
AllIsOneIsNone
Banned
I think we know what the answer will be here. Microsoft can use the number of sales the original Kinect sold to warrant it's ("necessary") return all they want, the fact is they alienated the core fans who wanted a gaming machine as well as distancing the casuals because of the price point. You start basing your moves purely on stats and you're doing your company a huge disservice.
Reminds me of the movie Big when the douchey antagonist is giving all the statistical data at the toy company and Tom Hanks is just like "what's fun about that though?".
In regards to VR; a Kinect-like device would be fantastic for it, but it sounds like even the Xbox One version doesn't have the right tech to work around it, so there goes the only potential I would be interested in.
Reminds me of the movie Big when the douchey antagonist is giving all the statistical data at the toy company and Tom Hanks is just like "what's fun about that though?".
In regards to VR; a Kinect-like device would be fantastic for it, but it sounds like even the Xbox One version doesn't have the right tech to work around it, so there goes the only potential I would be interested in.
No, the processing can take longer than 66ms which causes latency. In Kinect 1 there is hundreds of miliseconds of latency.
Edit: especially if you do temporal smoothing to eliminate jitter which is absolutely necessary otherwise your skeleton throws down like in that old fatboy slim video.
ZehDon
Gold Member
Kinect itself does nothing of the sort - though full motion body tracking in and of itself could be very powerful with VR. As has already been accurately stated in this very thread, Kinect's latency is simply too high. 60ms is over double the maximum amount of latency that is generally agreed as tolerable for a VR experience. Kinect 2.0 was not future-proofed - it was built to power the experiences that Microsoft promised Kinect 1.0 would deliver, and little else. Software improvements could be made, sure - but decreasing latency by 66% on just software? Microsoft aren't that incompetent, regardless of what we might like to think.
It's evident that Microsoft intended for the every-man to believe that Kinect was magically powering the Xbone's "amazing" voice command system, and not question the fact that the vast majority of Kinect's function is being entirely ignored not only by developers, but Microsoft itself. Full motion body tracking is amazing technology - it really is. However, it's applications in the console gaming industry are so few as to be not worth mentioning. The utter mistake of its inclusion is Don Mattrick's legacy.
To answer the OP: is Kinect worth the investment for Microsoft? Not really. Kinect 1.0 had a $500m marketing budget. Its sales probably paid for its R&D and marketing. The cost of 2.0, however, is unknown. Microsoft spent $100m on the controller, and what it produced has only one agreed upon improvement: trigger feedback. It's safe to say they spent more than that on Kinect 2.0, which has only slowed down the Xbone's adoption rate due to price inflation, costing Microsoft valuable early adopters and losing market-share gained by the Xbox 360.
Coupled with a near-total lack of titles for Kinect, and it's evident that Microsoft's R&D produced nothing more than an expensive anchor that drags the Xbone down by preventing competitive pricing, and pushing a focus on entertainment that too few people actually want. Kinect 2.0 is an interesting piece of technology, no question, but its not a consumer product, and it brings nothing of any real value to the Xbone that couldn't be replaced with a pack-in microphone.
I suspect Elop performs a daily "Curse you Mattrick" ritual before beginning each business day. If I'd been left with the Xbone and the problem of Kinect, I know I would.
It doesn't matter which measurement people use, 30fps means 33ms interval image sampling, it's too high even if the image processing was literally zero ms and it's not even close.
This isn't a serious roadblock to MS supporting VR anyway, just throw a cheap camera in the box with the headset, this isn't a problem.
Eh. It can send data in different streams; it needn't send everything at once.
i.e. it takes additional time to process the raw data into human motion capture data, but it can send the position of IR or LED points almost instantaneously without processing.
Send? Send it where? Either way the console has to process the positional data, obviously you can bring that processing time down dramatically if you're only tracking something like LEDs on a headset, but as long as they can't significantly improve the sampling rate of the camera, it's not going to solve the problem at all.
Maybe it is possible to have the camera capture at 120fps, but if it were, I'd think they'd allow developers to do that for normal Kinect interactions which could be more crisp if they don't require compete body tracking.
Neuromancer
Member
So why is the PS camera better suited to VR/head tracking than the Kinect?
test_account
XP-39C²
I think its more about what they believe could be a success or not. VR is not really a new concept, it have excisted for 20+ years or something. Its just being revisited now, and the result is better than we saw before because the technology is more advance now
Allforce
Member
Is the new brass at MS ready to invest in R&D for a brand new proprietary VR headset? There's already clamoring from major shareholders and board members that they need to stay out of the hardware field.
I mean they bragged that it cost almost 100 million dollars in R&D to design the XBox One controller, and there's virtually nothing different about it from the old one save for a few small features.
If MS jumps into the VR field it just feels like history repeating itself, they're constantly a day late and a dollar short with getting into the game with hardware. Zune, Windows Phone, Surface, all decent products but all released far too late to have any impact on the market.
I mean they bragged that it cost almost 100 million dollars in R&D to design the XBox One controller, and there's virtually nothing different about it from the old one save for a few small features.
If MS jumps into the VR field it just feels like history repeating itself, they're constantly a day late and a dollar short with getting into the game with hardware. Zune, Windows Phone, Surface, all decent products but all released far too late to have any impact on the market.
Very high framerate.
DeviantBoi
Member
AR allows them to display ads.
There's ad money to be made in AR. That's why MS is interested in it more than VR.
There's ad money to be made in AR. That's why MS is interested in it more than VR.
iceatcs
Junior Member
High framerate, cheaper and safer.
AR will be a bigger deal long term, they should be investing in it, I'm sure Sony are too.
Why do people often claim that Kinect's full body tracking would be great for VR? To me that just sounds like accidents waiting to happen.
AR makes far more sense in conjunction with Kinect, as you'd still be able to actually see the environment you're flailing about in.
I also think it has a better chance at mass market appeal, as it doesn't isolate you from social experiences locally. Having multiplayer games where the information isn't confined to a single screen could be a big deal in the future.
AR makes far more sense in conjunction with Kinect, as you'd still be able to actually see the environment you're flailing about in.
I also think it has a better chance at mass market appeal, as it doesn't isolate you from social experiences locally. Having multiplayer games where the information isn't confined to a single screen could be a big deal in the future.
Woah. It's easy to be an analyst with the benefit of hindsight, eh?
I'd be interested in how the Kinect can work with AR.
This. Making people move all their limbs while their vision is completely impaired isn't a good thing at all.
Imo, Kinect still hasn't proven that it was ever worth the investment. Outside of voice (which could have been done with a microphone, given the software libraries), there has been [almost] nothing that warranted such a huge investment in the technology.
Kinect, on a conceptual level, is great for VR because it allows the tracking of your hands without a specific tool. If you take what the Move does, the PS4 will produce a virtual representation of your body, it will take the position of the headset and contrive an arm movement model based on your head, and the controller in your hand. With Kinect, you wouldn't need to fake that data, if you bend your knees to lower your body without tilting, you could do that with your knees facing forward, or off to the side, Kinect would know that. It's all just about producing a better model of your position and movement to be represented virtually.
PoloBlue28
Banned
So it's impossible for MS to drop the Kinect resolution from 1080p through a firmware update? I really doubt it.
So, we're talking games like what we currently have with Kinect (that most people seem to hate), except with less movement range, because you'd need to pretty much remain in place?
I'm not really seeing how this has much application, if the current Kinect supposedly doesn't.
Dee Dah Dave
Member
The known latency figures of Kinect 2 are for the full skeletal tracking of 6 people at 1080p / 30 fps. There's nothing to say that the camera couldn't be used in a different way for faster / lower latency results (a la Sony).
Still, leaks would suggest that Microsoft are looking at AR not VR and may even have a product on the market this year if we go by their leaked roadmap from 2010.
Still, leaks would suggest that Microsoft are looking at AR not VR and may even have a product on the market this year if we go by their leaked roadmap from 2010.
AllIsOneIsNone
Banned
I think you're missing the potential for free hand/arm controls in a 3d environment.
Okay, imagine you put a headset on, and you're in a virtual environment, you look down, and you can move your legs, and they move, and you can hold your virtual hand up to your face and it moves and articulates exactly as your hand does, that's the kind of things we'll get to in time, which is lightyears away from the Kinect applications we've seen so far.
I don't think people think Kinect is bad technology, they think it's had bad implementation, the concept has incredible potential.
With appropriately made proximity sensors you'll be able to avoid mishaps like that. But I doubt we'll see anything like that outside of the next generation of arcades.
ricardokenshin
Member
then you bump into a table, fall down and break your 3d headset or something else.
I think that vr + kinect would only work with things where you are sitting down. Say a cockpit for a spaceship or a car. But, without the feedback from the steering wheel or the joystick I don't know if it would be very immersive.
virtual porn?
How? The original Kinect made quite a bit of profit, and the research that has gone in to both projects can be used for future projects (some of the ideas from IllumiRoom will likely get implemented in their AR glasses for example) . It's not a waste of time or money.
It's also pretty far off and requires a lot of technology currently being researched for VR.
SpookyFries
Member
Name me ONE Kinect game that isn't a glitchy ass mess.
No I'm not talking about voice commands in games, I am talking full Kinect games. The technology was neat when announced but its execution is shit. I don't believe it will ever be a viable solution for VR
No I'm not talking about voice commands in games, I am talking full Kinect games. The technology was neat when announced but its execution is shit. I don't believe it will ever be a viable solution for VR
Senor Kami
Banned
That's what I'd assume. In like every first use video I've seen, one of the first things people try to do is reach out and grab something or move their body only to find out that their head is the only thing in the virtual world. MS already has a product that solves that. Kinect might not be 100% perfect but it's a way better option imo than say Occulus, where it's 0% perfect for using your body to interact with a virtual world.
PlayStationboy
Banned
I definitely do think I'm missing something here, or at least have a failing imagination. I'm only seeing a difference of perspective. The Kinect can already track your hand/arms in a 3D environment, the only difference I see is that you would be able to see your how virtual body reacts from a first person perspective (looking down at your in-game body, or reaching up to the sky).
I can see it from a technical standpoint, I'm just not sure what additional gaming experiences this can really lead to. You could get higher FOV versions something like Child of Eden I guess.. or maybe a puzzle game where you hold the solution in your hands?
Overall though it seems like it would be far less versatile than the Kinect of today already is. How many gaming uses can there really be for looking at your own body, when any sort of wide ranging movement can't reliably be incorporated? It seems like it would only be able to provide a small subset of the games the Kinect does today, whilst ruling out many others (such as Kinect Sports, Dance Central, Kinectimals etc) and not adding many new ones.
Futhermore, the use of your actual body's movements seems like it would be better served by AR in most cases anyway, and this doesn't require removing any of the currently viable Kinect experiences.
It's not so much about what applications those abilities have, as how it impacts the users sense of immersion within the environment.
VR is about creating the sense of being within the game, and the richer your physical articulation is conveyed within the game, the more impactful the sense of self within the game should be.
Right now, the Move might be a very broad approximation of that experience, like here:
But long term, a motion tracking camera can be responsible for this.
I don't understand why the Kinect's viability/success is being compared to VR. Kinect as/is a $100 device that is easy for everyone to understand and use, and has one of the highest console attach rates (1.0 and 2.0). VR is much more expensive, much more niche, and I can't imagine it will be anywhere near as successful (in the short run). Not even a comparison.
WhiskeyKnight
Member
Goldenroad
Member
I think MS will probably let other companies invest in the R&D and trying to bring VR to market and then, if and when it proves to be a success, MS will iterate on it. I don't get why all these companies are diving head first into VR when it has yet to be proven to be a product that the general public is demanding. It could end up being much like 3DTV, where everyone thought they wanted it until they realized that they were more comfortable playing games and watching movies without the 3D.
ScepticMatt
Member
The latency of Kinect 2 is way, way too high for VR.
See my earlier post for research papers on latency requirements for VR and Kinect:
http://www.neogaf.com/forum/showpost.php?p=106134287&postcount=192
See my earlier post for research papers on latency requirements for VR and Kinect:
http://www.neogaf.com/forum/showpost.php?p=106134287&postcount=192
Neuromancer
Member
Interesting, thanks. I just assumed the PS camera was worse than Kinect 2, but apparently that's not the case, in all aspects anyway.
It's not so much about what applications those abilities have, as how it impacts the users sense of immersion within the environment.
VR is about creating the sense of being within the game, and the richer your physical articulation is conveyed within the game, the more impactful the sense of self within the game should be.
Right now, the Move might be a very broad approximation of that experience, like here:
But long term, a motion tracking camera can be responsible for this.
Yea, but there's no point of having an increased sense of immersion unless there's actually something worth playing with it.
Not being able to sidestep, move forwards and backwards etc, because you don't have any idea of your surroundings goes a long way towards undoing your sense of presence, because the game needs to confine you, or you need to make the conscious effort to confine yourself. By taking away the Move, you've removed your ability to navigate, and have confined your in game character to stationary or on-rails experiences.
You can't even turn around and reach behind you (as in Rez or Panzer Dragoon), because then the camera stops being able to see your hands and arms. So now we're talking about a VR experience where you don't even have 360 degree vision (no analogue stick for rotation). There's just too many limitations in my opinion to make it a good gaming implementation rather than simply a technical showpiece. Kinect is often criticized because the limitations of what you can control with it prevents it having enough software to justify its purchase. From what you're saying here, it sounds like a VR + Kinect would be useful for so few games you could count them on your virtual fingers.
Similar threads
- Poll
- Poll