Monday, 1 June 2009

E3 Update: Project Natal

Sadly, Treehouse budgets preclude us from jetting to E3 to cover press conferences in person, so we have to make do with live video feeds and reading smart alecy girls' typed updates. Microsoft were first up this year, and have dropped a couple of bombshells. In amid the Beatles, more Halo, game play footage of Modern Warfare 2 and Metal Gear Solid coming to Xbox, was a fairly extraordinary piece on the motion capture camera everyone's been poo pooing. This was clearly the crown jewels in the display, as Steven Spielberg and Peter Molyneux were wheeled out to wax lyrical about it, and some bloke named Kudo (Tsunoda- lead designer) who couldnt resist taking a few pops at Nintendo's lame waggle approach to inclusivity.

Project Natal (working title, for 'tis but a prototype) consists of what must be a fairly spiffy camera, capable of facial recognition, body motion capture, voice recognition, object recognition - blah blah blah, big whoop, whats new? - the presentation kicked off with a Nintendo style video of people playing various games, martial arts, racing, quiz games, all by just moving their bodies. Right, bit suspicious, although the motion capture looks impressive. Still cant be real. Oh look, they're navigating menus Minority Report style, and chatting on the video screen. Also, voice controlling movies - now that is handy, but hardly revolutionary.

Next up, after a potentially diastrous start in which Kudo's avatar spazzed around like a double jointed gymnast (lotta lights onstage huh Kudo?), an onstage demonstration of some tech demos; a block smashing game where a girl bounces around kicking punching and heading balls down a corridor. Hmm interesting; plausibility, rising.. Next a painting game where the guy paints a landscape with elephant. Not too interesting in itself (but clearly fun) but it's here I noticed how closely the movements were being replicated; as the guy just stood talking to the audience, the avatar onscreen mimicked precisely. Full body motion capture, with real subtlety of movement. Ooh.

Next up, Molyneux, and the part of the demo that puts this in the almost creepy folder. Lionhead have had Natal for a couple of months, and have come up with an AI boy, Milo, that the player can interact with. Using the facial and voice recognition, the creepy kid looks at the demonstrator's face, reacts with emotion to what she says while having a seemingly proper conversation, "throws" objects to her, is passed a drawing. Er, really, its a bit creepy. She then plays with the water in the river Milo lives by, her reflection shimmerring realistically in the surface as she sloshes the water about. That sir, is pretty freaking amazing.

Whether or not this will turn out to be what it appears to be, we'll know in the next few days as fortunate journos with their deep deep pockets get to try it out, and lord knows, we can't trust Peter "ZOMG" Molyneux. If it turns out to not be a fancy pants eye toy, and does do what it suggests though, well, I'm not one for hyperbole, but it could literally change the entire world forever. Or just be a ton of fun. One of those.

Image courtesy of (pinched from) Kotaku

Check after the break for some early videos.

The cheesy "product vision". Hmmm, really?

And Milo. Imagine Fallout 3 with NPCs with this tech.

1 comment:

S1nnerman said...

I agree with you. The potential for this is huge. Almost beyond imagination. Will it work as well as it was demonstrated? I don't know. I am having a hard believing that it will. Very expensive voice recognition software doesn't work well at the best of times even now and neither does image recognition in anything other than in very high end cameras. Are we to believe that a humble home console and an accesory can manage both these things so seamlessly? Maybe it can. Only time will tell. If it works, I'll buy a 360 without a second thought. From me, that is high praise indeed.