Lead developer exposes Natal inner-workings
Natal’s lead developer Alex Kipman has revealed that the upcoming peripheral will demand 10-15 per cent of the 360’s CPU resources.
He added that the device can recognise any pose in just 10 milliseconds – quick enough to suggest that the final product won’t suffer from lag issues.
And in under two-tenths of a second, says Kipman, Natal can recognise the body shape of a new user stepping in front of it.
"Natal has to work on the existing hardware without taking too much hardware processing away from the games developers,” added Kipman in an interview with trade publication New Scientist.
The lead designer added that he and his team had collected "terabytes" of data of people making poses likely to appear during game play.
Key body parts were meticulously identified in each image and frame, with all the data feeding into a computer rig.
According to Kipman, all the collated data resulted in a software package that can recognise 31 body parts in any video frame.
The software package itself weighs a mere 50MB, and is so sophisticated, says Kipman, that Natal can calculate where the player’s hand is even if its hidden behind their back.
“When we train this 'brain' we are telling it: this is the head, this is the shoulder. And we're doing that over millions of frames," says Kipman.
"When it sees a new image it can tell you the probability it's seeing a certain body part based on that historical information."
Microsoft confirmed yesterday that Natal would be released in 2010.
Essay Pro