Project Natal: MS Full Body 3D Motion Detection

There's already a video on here of Sugar Ray Leonard and other people with darker skin using Natal just like anyone else.

Yep saw that too. An interesting theory though as black skin does actually absorb light rather than reflecting it and id assume infrared light would behave the same as other light. Then again black cloths absorb light and we have seen it work with people in black suits.

Dont think there is anything in this at all but would be a serious issue if it was.

Could be that it doesnt work too well with overweight body types yet or something which would be an easy software fix rather than a inherent hardware issue.
 
They still have to figure out why it didn't work for Ngai. :) Best is to understand the differences in the environments (The Sugar Ray Leonard video may be taken under different conditions).

That video was taken in a extremely wel lit white coloured room if i remember correctly. It certainly is possible i that any inherent problems with certain skin types could be amplified in a suboptimal environment.

A black guy in a dark room using a projector would be the ultimate test lol
 
There are 2 levels of issues. The PR one and the tech one. The linked article responds to the former (because it uses the words "will obviously work" instead of a simple "works"). I am refering to the tech issue. They still have to work out why it didn't work for Ngai when it worked for the other MS staff just moments ago. The answer may or may not be skin color related, but it's an obvious one to start with.
 
I'm sure Microsoft is noting issues during demos and working on them, but I'd say there's a 0% chance we'd ever hear about it.
 
The discussion started because Arwin made a comment about seeing in the dark; a point I took for granted. If the problem is indeed caused by near-IR light and object color, specifically skin color, then I am simply wondering whether they need to use the RGB camera to help out; or if they can use another solution (e.g., use a "better" IR light ?). If the former is used, then it's back to Arwin's question.... it would be sensitive to some lighting conditions too (but should be much better than a pure RGB camera solution like PS Eye).
 
Here another aritcle with impressions and some more info:

http://games.kikizo.com/features/microsoft-natal-xbox-360-why-its-a-big-deal-p1.asp

The reason there is a thing going around about Natal not being able to detect black players is that a black journalist had difficulty controlling the Burnout demo. Guess what? So did a white female that was in my closed-doors session. I could make a 'women drivers' gag at this point, but I won't because this is a serious point. The Burnout demo was probably not the most finely tuned thing Microsoft could have used to demonstrate Natal, and could have done with some further optimisation. The fact that many people of all colours who've tried this particular demo have had difficulty trying to steer about, accelerating by putting right foot forwards, only shows that this is probably a bit of a last minute demo to have modified - it's another company's game code, essentially hacked into a makeshift Natal demo. Kudo stated in the session that it was not designed specifically for Natal and that it was just to make the point that Natal has no latency issues when it comes to fast paced games like racing games. It certainly does not mean that this is something specifically to do with dark-skinned players. I hope we're quite clear on this.

Read the whole article though. He does hands on impressions with Natal + Milo.
 
Interesting point and Iim sure during development the tech hasn't been purely the domain of White Supremacists! On a side issue though, I don't see how the choice of Burnout or the last-minutedness of the demo can lead to detection issues if the tech works as I'm led to believe. As I understand it, Natal is supplying joint info to the game. The developer only need map the right leg/foot joint to acceleration, and a vector between both hand joints to determine steering. This isn't at all hard and should be absolutely reliable if the joint positions are accurate. Which leads me to think that the hardware isn't stable at this point, and the issues with Burnout are being presented here as something of a scape-goat. I think MS have shown their hand a little early, and now they have to make sure they can iron out the bugs. They don't want to repeat the mistakes they've made with their other peripherals!
 
If I were Microsoft I would be planning aother demo for Ngai, but it looks like they've already decided to respond to the rumor...

http://www.joystiq.com/2009/06/11/microsoft-project-natal-obviously-works-with-people-of-all-et/

Tommy McClain

While im doubtful it is a problem with skin colour the quote in that article: "it will obviously work with people of all shapes and ethnicities at launch" does little to disprove the rumours, in fact it makes it sound more suspicious...
 
I'm sure Microsoft is noting issues during demos and working on them, but I'd say there's a 0% chance we'd ever hear about it.

I think this should be pretty obvious by the amount of monitoring hardware they are using in addition to the NATAL mockup. Allowing them to track in various ways how Natal is performing.

I'm sure MS is using this whole experience not only for PR but to gain more data points on the performance of Natal. Where it works well and where it doesn't work well.

Likewise, it gives them a chance to gain feedback from professionals that are unfamiliar with the system. Or at least more knowledgeable than pulling random people through an AD in the paper for a focus group test.

And apparently allowing them to test in diverse locations each of which could provide it's own challenges to the system. The Late Night with Jimmy Fallon studio for example.

Having watched the Fallon show and other video's it's quite obviously not ready for release. However, presumably now that hardware and software specs have been set, the next 1.3 years (until holiday 2010) will probably mostly be devoted to working out kinks in the software and improving recognition and response. As well as taking input from devs that now have the NATAL dev kit on ways to improve the system.

We'll probably never find out why it doesn't work for a few people while working well for everyone else. Which is a shame as I'm quite curious to know why it fails with a very small subset of people that don't appear to share anything in common.

It is a bit curious why the PR department decided to show it at this years E3 rather than next years E3 if the system is over a year away still.

Regards,
SB
 
While im doubtful it is a problem with skin colour the quote in that article: "it will obviously work with people of all shapes and ethnicities at launch" does little to disprove the rumours, in fact it makes it sound more suspicious...

Love conspiracies don't you? If you want to ignore all the other people that have used it with no problems go right ahead.

Tommy McClain
 
If it worked with 100% of people, yes. However, it has failed some people. Ergo it doesn't work with 'all shapes and ethnicities' now. MS are prediciting it'll work when launched, but that doesn't change the reality that at this point, is does not work for everyone. If it did work for everyone now, there wouldn't be any reports of it not working for MS to issue a statement 'it will work' ;)
 
I think this should be pretty obvious by the amount of monitoring hardware they are using in addition to the NATAL mockup. Allowing them to track in various ways how Natal is performing.

I'm sure MS is using this whole experience not only for PR but to gain more data points on the performance of Natal. Where it works well and where it doesn't work well.

Likewise, it gives them a chance to gain feedback from professionals that are unfamiliar with the system. Or at least more knowledgeable than pulling random people through an AD in the paper for a focus group test.

And apparently allowing them to test in diverse locations each of which could provide it's own challenges to the system. The Late Night with Jimmy Fallon studio for example.

Having watched the Fallon show and other video's it's quite obviously not ready for release. However, presumably now that hardware and software specs have been set, the next 1.3 years (until holiday 2010) will probably mostly be devoted to working out kinks in the software and improving recognition and response. As well as taking input from devs that now have the NATAL dev kit on ways to improve the system.

We'll probably never find out why it doesn't work for a few people while working well for everyone else. Which is a shame as I'm quite curious to know why it fails with a very small subset of people that don't appear to share anything in common.

It is a bit curious why the PR department decided to show it at this years E3 rather than next years E3 if the system is over a year away still.

Regards,
SB

While a side effect of this PR move may be gaining aditional experience they would be incredibly stupid if one of the reasons for showing it off at this stage was for testing purposes.
 
While a side effect of the PR may be gaining aditional experience they would be incredibly stupid if a significant reason for showing it off at this stage was for testing purposes.

Eh? I never stated that additional data gained through presenting NATAL was the main reason for showing it.

In fact I think I was pretty clear at the end questioning why MS PR decided to demo it this year instead of next year when the system itself is presumably more than a year away from launch.

Regards,
SB
 
Love conspiracies don't you? If you want to ignore all the other people that have used it with no problems go right ahead.

Tommy McClain

Read post 301 i havent ignored anything. I think you may be being a little touchy.
 
Last edited by a moderator:
If it worked with 100% of people, yes. However, it has failed some people. Ergo it doesn't work with 'all shapes and ethnicities' now. MS are prediciting it'll work when launched, but that doesn't change the reality that at this point, is does not work for everyone. If it did work for everyone now, there wouldn't be any reports of it not working for MS to issue a statement 'it will work' ;)

I'd be more worried if this was launching in the new few months or even this year. As a demo, it's worked well and the issue it has will surely be addressed as time passes on. MS has been getting a chance to test this tech out with various people in various locations. It's one thing to tell your paid testers to follow your commands. It another for media personalities to flail their limbs about wildly and just to see how the tech responds. So far, it'd seem that the positive results seem to greatly outnumber the negatives.

You'd think being on a tech site, someone wouldn't have to hold you down to hammer the concept of "improvements during the development cycle" but for you and quite a hanful of others that seems to be the case. Perhaps, it's just reahearsel for the dance if the tech does fail at launch. Who knows....In any case, if you expect Natal to be perfect, then I'd start the dance now.
 
Eh? I never stated that additional data gained through presenting NATAL was the main reason for showing it.

In fact I think I was pretty clear at the end questioning why MS PR decided to demo it this year instead of next year when the system itself is presumably more than a year away from launch.

Regards,
SB

Yep sorry bout that i did edit out the significant out of my post shortly after posting.

The way i read your post it seemed like you was suggesting one of the reasons for MS to show natal publicly was to get feedback from press and test it out in diffeent environments.
 
I don't think there's anything to conclusively say it has detection issues with particular ethnicities or shapes. The "conspiracy" talk is annoying. The product is still a year or more out, so bugs are basically given at this point. The only question is whether it'll work well on release date, and we won't be able to find that out until people start reviewing the retail product.
 
I'd be more worried if this was launching in the new few months or even this year. As a demo, it's worked well and the issue it has will surely be addressed as time passes on. MS has been getting a chance to test this tech out with various people in various locations. It's one thing to tell your paid testers to follow your commands. It another for media personalities to flail their limbs about wildly and just to see how the tech responds. So far, it'd seem that the positive results seem to greatly outnumber the negatives.

You'd think being on a tech site, someone wouldn't have to hold you down to hammer the concept of "improvements during the development cycle" but for you and quite a hanful of others that seems to be the case. Perhaps, it's just reahearsel for the dance if the tech does fail at launch. Who knows....In any case, if you expect Natal to be perfect, then I'd start the dance now.

I dont see anyone saying it wont work as advertised come launch day, just that for sum reason it has not worked for some people right now. As you stated this is a tech site and i think its perfectly natural that people would want to discuss the technical reasons for it not working in those rare cases.
 
I don't think there's anything to conclusively say it has detection issues with particular ethnicities or shapes. The "conspiracy" talk is annoying. The product is still a year or more out, so bugs are basically given at this point. The only question is whether it'll work well on release date, and we won't be able to find that out until people start reviewing the retail product.

There is no conspiracy man. Nobody has said there is a problem with it to do with ethnicities just that its a possibility in its current state. AzBat posted an article that seemed to be a response from microsoft that denied any issues, i then followed up by saying it didnt deny it but said it would obviously not be a problem upon release. Where is the issue here?
 
Back
Top