Details On Natal's Motion Capture Technology 121
An anonymous reader writes "Following yesterday's announcement of a late 2010 launch date for Natal, more details are emerging on exactly how Natal works. Alex Kipman, the project's lead developer, explains that Natal uses only 10-15% of the Xbox's resources to calibrate to a new player inside 160 milliseconds, track one or two players simultaneously, and use rudimentary knowledge of body anatomy to estimate where hands or other body parts are even when they can't be seen by Natal — for instance when they are held behind the back."
Accuracy (Score:2, Insightful)
In my mind what will affect user experience the most is its accuracy, and latency. It uses a infrared camera, so I'm wondering if anyone knows the camera's specs, especially regarding resolution and latency. Without knowing these, it is difficult to evaluate what it is good for. The article quotes a several cm. accuracy, and 10 ms recognition time, but it not clear what the sensor's limits are, and how much latency there is in the processing.
Re:Debate? I gots yer debate! (Score:3, Insightful)
Re:awesome (Score:3, Insightful)
I agree that motion sensing won't kill controllers; but it will become an essential feature. Would you like to go back to using your computer with no mouse?
Re:Natal Demo (Score:2, Insightful)
I call shenanigans on this demo. once the footage of milo actually starts, you may notice that the avatar is looking out from the screen directly towards the girl at points.
Granted, she is rather easy on the eyes, but this tends to suggest the thing has been staged. Either that or the tv used has a 3d capability more advanced than anything in existence.
Re:Natal Demo (Score:3, Insightful)
Microsoft never made any such insinuation whatsoever. I don't blame you for having this mistaken understanding of the situation as there's so much FUD going around about the Milo demo created by fanboys of other consoles but it's still completely and utterly wrong so rather silly to keep repeating.
The Milo demo does one thing and does it well, it demonstrates how Natal can be used as a CONTROLLER to get PRE-SCRIPTED RESPONSES (whilst I hate caps, these points are fundamental to what Natal is and how it can be used). This is no different that when you walk upto a character in an RPG with a classic console controller and the NPC speaks, the NPC isn't recognising you as a person has turned up and speaking specifically to you, it's detected you are in range through basic math and played back a specific response. The Milo demo is no different other than the fact the controller is different- they have demonstrated that Natal as a controller allows for different inputs just as a classic controller usually has two directional sticks and a multitude of buttons. It can scan images, it can respond to gestures and it can respond to voice recognition (something that has worked for years now- see Endwar for a previous successful example of it's use in games).
It's far from a fabrication and all the conspiracy theories that fail to actually show anything worthwhile are utterly stupid. One example was some dumb Youtube video pointing out Milo is not looking at the woman- well duh, it's pre-scripted, it's designed to look good for the camera. The real issue with the Milo demo is this, fanboys saw it as a serious threat to their favoured console because they made the false assumption in their fanboy rage that it was demonstrating strong or near strong AI which is clearly absurd to anyone thinking rationally.
Natal is what it is, it's a controller, it will allow more immersive interactions, but no one's really pretending that alongside Natal Microsoft have managed something that might not even be possible, and certainly isn't possible with today's technology- strong AI. Microsoft made it quite clear the Milo demo was to show off the potential of Natal, they never once tried to pretend that alongside Natal they'd made a massive leap in AI. If you made that assumption either mistakenly or maliciously because of fanboyism then that's your insinuation not Microsoft's. Interestingly what this also means is that because Milo was pre-scripted, the whole thing is also perfectly possible to have in your living room but the caveat is that it wont be as exciting as your insinuation suggests, because like the actor, you'll also have to perform the pre-set actions that get pre-scripted reactions. Watch the presentation and tech demo again- do they even once suggest that Milo is dynamically figuring out how to respond to the users actions? Not in the slightest.
What Natal will be able to do however is allow players to use more than just their thumbs in games- there's no reason you can't play Call of Duty like you always have with the controller but also now be able to physically lean left or right to peak round corners of buildings and so on. In games like Fable there's no reason you wont be able to make gestures like pretending to throw a ball to make your character throw a ball for your dog. There's no reason you wont be able to make gestures to draw things which can be turned into objects ala Scribblenauts and Crayon Physics. It does one thing and does it well- opens up the console for much more immersive controls just as the Wii did and just as Sony's new motion control system also will, the only difference being that Microsoft have done it in a different way to Sony and Nintendo.
All that said, I've no idea why you would think they haven't perfected motion and facial recognition for use in games, these have been solved problems for a fair while now to a level where they're useful in many games and applications and as pointed above, the same goes for voice recognition. Natural language processing and natural speech