Slashdot Mirror
Why Natal Is a Big Deal
Kikizo has an editorial piece evaluating the Xbox 360's upcoming motion-control scheme, Project Natal, and discussing why it's a bigger step forward for interactive gaming than many people think. Quoting: "[Natal] accurately perceives players in 3D space, simultaneously tracking over 48 joints on your body, enabling it to accurately redraw your skeleton in real time as you move about. On a separate 'debug screen' in the closed-doors session, we could witness for ourselves the 'mind's eye' of Natal, visually showing how it completely understands where we are, how we're moving, where we are in 3D space, how far in front of my face my hand is, whatever. It can supposedly even track individual hand and finger movement when it switches into this more finely-tuned mode. ... There is a surprising feeling of tactility and iPhone-like fluidity and precision to the way Natal works." Another interesting bit of news about Natal is that Wii-hacker Johnny Chung Lee is part of the development team. We've discussed some of his creations in the past.
Xbox Project Natal : Felicia Day
"I can't wait to throw a fireball."
http://www.youtube.com/watch?v=DYuJivFFa-c
The project Natal is indeed different for an Eye tool of a wii mote. This is a depth sensing camera (you have a RGB image and an image with the distance to the camera). This camera is made by Primesense (you can check theirs patents), it works by projecting a grid (infrared so you can't see it). By analyzing the deformed pattern with a camera, the depth is computed.
Having the deep information is really useful because you get the 3d cloud of the user and with some math, you can guess where is the user and what he is doing. For example you can find the biggest cluster of point (the user). The mass center of these points will give you the position of the user (this is already enough to to a lot of thing).
But the technology is not perfect. It will never be a full replacement for motion capture because it's subject to occlusions and there are a lot of ambiguous cases so the system will always need to cheat.
This technology seems kinda fake to me. If you refer back to E3 where they first showcased the device, there were several odd and unexplained things going on. 1) The twitchy avatar character shown (especially when the guy tried to show the bottom of his shoe and couldn't.) seemed to show that the technology wasn't really complete. 2) The other presenters wore dark clothing that seemed to contrast better with their surroundings. Yet, the people in the promotional video wore more colorful clothing. 3) The device only seemed to be able to detect only wide movement and not subtle movement like the promotional video suggested. 4) Most core gamers would like to take advantage of this technology. However, most gamers like to play in dark rooms. It seems to me that it'll be more difficult for this camera to adjust to harsh lighting conditions (dark rooms, lens flare, moving background lights, etc.). 5) It didn't feel as though the camera could decipher between more than one person, because no one (presenters or journalist in the closed room) tried to test it with that in mind. Sure it could detect more than one person; but could it tell the difference between the two? 6) Where are the games that utilize this technology? It seemed that all that was available were tech demos. I'd very much like to see this technology put to use in actual real time environments instead of controlled environments. It gives me the sense that this project is a little premature and may not see the light of day for a long time.