Nintendo's 'Natal' ?

[WereKitten]

dcIKeeL said:

Wow ok. The Natal camera is not just one camera but two 3D cameras. They can register depth as well as the standard stuff cameras like the eyetoy register. The fact that it's 2 cameras gives it more versatility and allows it to differentiate between different people and objects. Voice command and voice recognition are not the same thing. Sometimes these to techs can be called the samething (voice recognition). Voice command is simply a device that's able to register voice and coupled with the right software gives actions to the words and phrases it registers.

Voice recognition in the way Natal implements the term is being able to not only register voice, but doing it more intelligently. Instead of relying almost entirely on software, Natals' hardware can do most of the registering AND assigning on it's own. More importantly, Natal can differentiate once again between different people. Natal essentially can distinguish your voice from that of a friends or family member. The hardware is also advanced enough that when coupled with appropriate software allows for recognition of tones and emotions in both voice and facial recognition.

It's really annoying when people claim that Natal is the same as other cameras in video gaming today. People that claim this nonsense should be banned for excessively absurd trolling and flamebaiting! Next time anyone gets stuck on this nonsensical nonsense, ask yourself: If the existing camera tech you are considering was truly as innovative and advanced as the Natal camera, why aren't it's manufacturers using it?

Actually for what I know it's a "normal" RGB camera plus an infrared one that uses an illumination grid of marker lines to reconstruct a depth map. If you want you can call the second one a "3d camera", though basically it elaborates displacement/resizing of the markers in a way not too dissimilar by what the Sony system does with the image of the glowing ball placed on the wand controllers.

The vantages of using this infrared+grid setup is that they want to map whole bodies, not a discrete set of controllers, and match it to a skeletal model. So it's more in the whole architecture of the modelling software (depth map-> skeletal model) than in the tech of the camera itself.

The voice technology as well as the face recognition is basically software, though maybe the actual final piece of hardware might have specific silicon to accelerate some processing. Nintendo could indeed implement some of it, with the CPU and memory limitations of the hardware they work with.