Older Release Video
Newer Demo Video
Now it all looks very cool and it's been announced three of the games shown at the E3 press conference will become launch titles along with the promise of further demos involving adaptations of Beautiful Katamari and Space Invaders Extreme being released.
Oh Katamari, you will rock so hard if this comes to fruition.... epically the new version in HD :)
Anyway, the basic idea behind Natal is that the camera can process your bodies movements in 4 cm chunks and recognition of 31 different body parts in any video frame. No controllers, no wiimotes nothing just you're body movements (AKA Super Eye-toy). It's also been programmed to recognise the location of your hands even when they may be obscured from view understanding that most humans have two hands, feet legs and so on. If you're a meat packer with only one limb you may be in for some trouble but I doubt that Xbox's new peripheral is the biggest of you're concerns.
Microsoft collected "terabytes" of data from people in poses likely to crop up during game play, both in motion capture studios and their own homes. Frames from the home videos were manually labeled to identify key body parts, and the data was then fed into "expert system" software running on a powerful cluster of computers. The result was a 50-megabyte software package that only leans on the 360's processing power by around 10 - 15%. Not bad considering the number crunching involved. Also it can recognize any pose in just 10 milliseconds, not quite as good as Hollywood's Motion Capture tech millimetre precision but decent enough for us regular gaming folk.
No comments:
Post a Comment