The non-standard game interfaces discussion thread (move, voice, vitality, etc.)

[BadTB25]

erm, I don't want to add fuel, but Kinect can't even do finger recognition (unless maybe you have to stand there making a sign for 10 secs while it figures out what sign you're making in 'sign mode') so IMHO this is all about is stopping others from doing it.

You sure know with certainty a lot of what Kinect can and can't do for a product that hasn't officially launched.

Don't get me wrong, I'm not thoroughly convinced myself, but I'm not ready to make sweeping statements about what it is and is not capable of doing...yet. Also, if some of our posters are correct, Kinect could just be a Beta for what is included in the Next Box that would be able to include more skeletal points including each digit.

 

[Shifty Geezer]

You sure know with certainty a lot of what Kinect can and can't do for a product that hasn't officially launched.

Don't get me wrong, I'm not thoroughly convinced myself, but I'm not ready to make sweeping statements about what it is and is not capable of doing...yet.

We've heard finger tracking isn't available, and we have the tech specs that show the camera is low resolution, making digit separation difficult. Finally we have depth-buffer images to examine, and we can see there's no way to identify the fingers. There's seemingly too much noise in the depth info that is being filtered out, losing detail and leaving only broad silhouettes.

A developer could create a finger-tracking interface if the user moves closer to the camera, just as they could any webcam or the like, but the default skeleton tracking can't be able to track fingers with that depth view.

Also, if some of our posters are correct, Kinect could just be a Beta for what is included in the Next Box that would be able to include more skeletal points including each digit.

The next gen is bound to be better, so I wouldn't call this a Beta any more than any first gen product is. That also doesn't affect what Kinect can do now.

As for the patent, it sounds like a broad patent to cover any input strategies, and it's a smart move. The idea of lip-reading would help with voice recognition. I can't see how sign-language could be patented when gesture input is nothing new, and sign language is just a vocabulary of gestures that MS doesn't own. The EDGE reporting said these were actual features of Kinect, rather than a patent. Whatever happened to journalistic integrity? What really makes this patent shine though is this technical explanation :

In this situation, for example, when the user kills another user's character, that victorious, though speechless, user would be able to tell the other user that he had been 'PWNED'.

Smack-talk in a patent?!

 

[BadTB25]

Are those depth-buffer images actually from Kinect? My impression was that it was from some other tech that maybe shared with Kinect.

Also, is there no other way that you could do finger recognition that doesn't utilize the depth-buffer information. From everything I've read, it seems that developers have a wide variety of tangents that they could pursue from using the MS libs to creating their own.

My point was that for a product that is really in its infancy, we shouldn't be too quick to state its limitations nor praise its potential with absolute certainty.

 

[scently]

Are those depth-buffer images actually from Kinect? My impression was that it was from some other tech that maybe shared with Kinect.

Also, is there no other way that you could do finger recognition that doesn't utilize the depth-buffer information. From everything I've read, it seems that developers have a wide variety of tangents that they could pursue from using the MS libs to creating their own.

My point was that for a product that is really in its infancy, we shouldn't be too quick to state its limitations nor praise its potential with absolute certainty.

Please guys read this sentence twice because for some reason we are treating the current ( very early) state of the software as the final and absolute state the system is always going to be.

Seriously, from the whole ps3 cannot do msaa to the brilliant mlaa solution, and in general the natural evolution of software development of fixed hardware, we should stop judging this system with such absolute conviction.

Really, what is preventing them from adding more skeletal points eventually.

 

[Shifty Geezer]

Please guys read this sentence twice because for some reason we are treating the current ( very early) state of the software as the final and absolute state the system is always going to be.

So no-one should evaluate the current state of the technology nor endeavour to make educated predicitons?

These calls to abandon discussion bemuse me! Progress requires new ideas, exploring them, finding what works, finding what doesn't. Nothing said at this point is concrete, and to claim now that Kinect can't do finger tracking does not mean it will never happen; it's a probability consideration, factoring in known info and likely outcomes. Does finger tracking exist now? No, probability 100%. Should a person buy Kinect in the expectation that it'll have finger tracking? No, probability 95%. Will Kinect gain finger tracking? Unsure, but 65% chance of it not happening. Or maybe 80% certainty that it won't happen. Or 80% it's bound to happen. the information we consider gives us a chance to determine how things will develop. But regardless, we're only exploring in theory the practicalities of a new tech. What possible reason is there to be against this?! Why should we not question and speculate?!

...we should stop judging this system with such absolute conviction.

Perhaps this is the issue here. Who is judging? How can a list of features be considered a judgement?

Really, what is preventing them from adding more skeletal points eventually.

Presumably the same thing that prevented them from suporting 48 point skeletons or whatever the original spec was, and the same thing that prevents 4 players on screen at once; finite processing resources.

 

[BadTB25]

I'm all for discussion and speculation. My problem it when we use words like "can't" as if it is a foregone conclusion.

I don't want to sound like a Kinect Defense force as my feelings on Kinect is lukewarm to optimistic, but as scently mentioned software changes as devs find new workarounds for tasks that were thought to be not possible or probable with fixed hardware.

Is the 48 point skeleton limited by hardware or software?

 

[scently]

So no-one should evaluate the current state of the technology nor endeavour to make educated predicitons?

These calls to abandon discussion bemuse me! Progress requires new ideas, exploring them, finding what works, finding what doesn't. Nothing said at this point is concrete, and to claim now that Kinect can't do finger tracking does not mean it will never happen; it's a probability consideration, factoring in known info and likely outcomes. Does finger tracking exist now? No, probability 100%. Should a person buy Kinect in the expectation that it'll have finger tracking? No, probability 95%. Will Kinect gain finger tracking? Unsure, but 65% chance of it not happening. Or maybe 80% certainty that it won't happen. Or 80% it's bound to happen. the information we consider gives us a chance to determine how things will develop. But regardless, we're only exploring in theory the practicalities of a new tech. What possible reason is there to be against this?! Why should we not question and speculate?!
Perhaps this is the issue here. Who is judging? How can a list of features be considered a judgement?
Presumably the same thing that prevented them from suporting 48 point skeletons or whatever the original spec was, and the same thing that prevents 4 players on screen at once; finite processing resources.

So what you are saying is that they cannot add more points with future software SDK and that it can never support 4 players? Seriously for someone used to software evolution in consoles, I am disappointed and perplexed by your answer.

It could simply be a case of optimization. The xbox360 supported just 1080i at launch and now it can do 1080p. Seriously which hardware feature explicitly prevent the addition of more points or adding 4 player?

 

[Shifty Geezer]

Are those depth-buffer images actually from Kinect? My impression was that it was from some other tech that maybe shared with Kinect.

I don't know, but the depth info must be at least similar if MS are using it. The Engadget video shows how much noise there is in the source, which is too much to get finger tracking at normal playing distance. At 50 minutes, the engadget video feed is almost 1:1 with the 3D camera resolution. I cropped out an image of Kudo Tsunoda's red data and skeleton and it was 360x266 pixels (not neatly cropped). It's a very raggedy silhouette, but the skeleton tracking's predictive placement overcomes the missing data.

Also, is there no other way that you could do finger recognition that doesn't utilize the depth-buffer information. From everything I've read, it seems that developers have a wide variety of tangents that they could pursue from using the MS libs to creating their own.

There's always optical and Kinect has a higher resolution optical camera, but that'll suffer from EyeToy like contrast problems. Recognising hand gestures is possible, but actually tracking finger movements is going to be extremely difficulty visually unless you get up close to the fingers

My point was that for a product that is really in its infancy, we shouldn't be too quick to state its limitations nor praise its potential with absolute certainty.

That's true if someone doesn't want egg on their face when proven wrong. Goonergaz probably should ahve said, "at this time." however, his sentiments seem substantiated by known info at this point, such that the patent suggesting finger tracking probably only relates to securing IP rights rather than describing what the first-gen tech is capable of. Putting it another way, if Kinect wasn't capable of finger tracking, would MS leave that out of their patent?

 

[Shifty Geezer]

I'm all for discussion and speculation. My problem it when we use words like "can't" as if it is a foregone conclusion.

Okay, it's worth pointing out I've entered this patent debate from being introduced to it via the EDGE article, which headlines, "Kinect Will Recognise Sign Language." I read the article and got annoyed at people making claims that something can be done when perhaps it can't, so in essence we're both in this debate from two sides of central.

So what you are saying is that they cannot add more points with future software SDK and that it can never support 4 players? Seriously for someone used to software evolution in consoles, I am disappointed and perplexed by your answer.

Okay, I'm a huge proponent of human ingenuity and a firm believer that amazing solutions will always be found, hence my enthusiams for flexible hardware and software renderers. Developers have access to Kinect's source data, so they can do whatever they want with it. We know there's enough processing potential for 2x20 points, so a single 48 person skeleton could be done, and if the limit is processing power, there's nothing to stop a developer devoting 50% of their CPU and GPU cycles to image recognition should they want. One never knows what other extremes developers will go to. 4 players using a simplified tracking method is also possible.

This argument is about the patent and claims that Kinect has features like lip-reading and ASL input. Goonergaz was saying that (at the moment) Kinect doesn't even have finger tracking (within the default libraries) and as such, the patent can't be listing features but instead securing IP rights. That some developer some time may find a way to track fingers doesn't validate the patent as a list of Kinect's abilities. Hell, Kinect hasn't got any abilities beyond two sources of data, as everything that's achieved through that is done in software!

Perhaps a distinction needs to be made between the Kinect hardware and Kinect software libraries?

 

[obonicus]

Smack-talk in a patent?!

Oh, Internet. Is there anywhere where you're not funny?

 

[Cyan]

I'd gladly follow the discussion, it's so open to debate but I have no time right now so I will post some amusing pics of an Ubisoft game for the Wii. The game is Just Dance Kids. It might sound wicked but I laughed, cool stuff. I don't know if it will come out for PlayStation Move or Kinect, but it seems to recognize people's faces....

Mod edit : Removed gallery.

 

[Shifty Geezer]

Does this game include a camera peripheral? Or are the characters pre-recorded game characters that respond to Wii waggle?

 

[BadTB25]

Judging by the perfect synchronization, I'd guess that it is prerecorded.

 

[Silent_Buddha]

Presumably the same thing that prevented them from suporting 48 point skeletons or whatever the original spec was, and the same thing that prevents 4 players on screen at once; finite processing resources.

Actually if you watch any of the 2 player Kinect experiences on Youtube you'll see this is far more likely to be a space limitation than a processing limitation. Just watched an adult and kid play the Konami dancing game in Japan on Youtube. And there is just no way to physically fit another player into the Camera FOV, as it was they were already crowding each other to some extent when their arms are outstretched.

As an example just grab another person and stand fingertip to fingertip with arms outstretched. Doesn't leave much room for a 3rd player.

It's entirely possible that once "couch potato" games start being developed that we may see 3 player games, but as pretty much all of the launch titles feature vigorous physical activity, there's just no room for more than 2 players.

Regards,
SB

 

[Arwin]

The resolution suggests this will be an HD game. The four player support rules out Kinect. The orange puppet in the bottom-right corner having a glowing ball in one hand seems to suggest this is the PS3 + Move version? There is some face mapping in another PS3 + Move game that may be similar to what happens here.

 

[Silent_Buddha]

The resolution suggests this will be an HD game. The four player support rules out Kinect. The orange puppet in the bottom-right corner having a glowing ball in one hand seems to suggest this is the PS3 + Move version? There is some face mapping in another PS3 + Move game that may be similar to what happens here.

If it's based on the Just Dance title (implied due to the Just Dance Kids title ) coming out for Kinect, then it's just one player being actively tracked while the others are there just for fun and to play along.

Regards,
SB

 

[goonergaz]

I'm all for discussion and speculation. My problem it when we use words like "can't" as if it is a foregone conclusion.

Sorry to upset you but it can't do finger tracking (at the moment), we know this (as already stated by shifty) and because of the fact that in the bowling game it can't even track wrist movements (OK this is something I have decided myself from goonerlogic).

I suggested 'sign mode' as I dare say it can do finger tracking if it's only tracking hands and you stand closer to the camera (as pointed out due to lack of resolution).

I don't mind egg on my face, I'm an Arsenal fan and used to it! LOL

 

[Arwin]

If it's based on the Just Dance title (implied due to the Just Dance Kids title ) coming out for Kinect, then it's just one player being actively tracked while the others are there just for fun and to play along.

Regards,
SB

Look closer. There are four players being scored, one in each corner.

 

[Shifty Geezer]

Actually if you watch any of the 2 player Kinect experiences on Youtube you'll see this is far more likely to be a space limitation than a processing limitation...
It's entirely possible that once "couch potato" games start being developed that we may see 3 player games, but as pretty much all of the launch titles feature vigorous physical activity, there's just no room for more than 2 players.

That's a reasonable argument, but my counter would be that if MS had any form of four-player tracking, wouldn't that be listed in the features even if no games use it due to space limitations? Similar to PS3 supporting 7 controllers, but no games targetting that because it'd require over £200 on controllers. I expect features to represent the best possible representation. eg. I've just ordered a new USB audio card that list Mic, line, and Hi-Z guitar inputs, but what the features don't tell you is you can only use one at a time. If MS had four skeleton tracking working in the lab, I'm sure it'd have made it to the features as 'up to four players tracked'.

 

[goonergaz]

And there is just no way to physically fit another player into the Camera FOV, as it was they were already crowding each other to some extent when their arms are outstretched.

Just dance supports 4 players dancing around (and will support 8) and singstar dance 2, I wonder if this is the only advantage of the way Wii works vs PSMove/Kinect?