Anyone have a link to the Natal demo where the person is talking to the AI controlled person? I can't seem to find it anywhere.

Quote: Original post by Moe
Anyone have a link to the Natal demo where the person is talking to the AI controlled person? I can't seem to find it anywhere.

I assume

">this is it.

I've seen a sensor system like this in person... it might have actually been the same thing. I was impressed. The distance sensor basically gives you the equivalent of a depth buffer to work with (which is a lot nicer for rough shape tracking than a color image). The depth resolution is pretty good and the capture rate is very high. At the time, it was hooked up to a mid-range laptop which was doing some very impressive 3D feature detection on the input.

The thing could recognize hand gestures at 5 feet easily.

The main thing I was concerned with at the time was the processor time required to handle the image / depth information. I could totally see it working on one of the 360's cores though. You'd never have enough processing power on a Wii (and still have time to run a game). In that case the sensor would need its own embedded processor, which would put costs too high.

So... if it's the same thing I saw, it'll probably work better than the Wiimote (position is better than acceleration as far as I'm concerned).


(edit) I think that this is the device I saw in person.

why on earth interview Peter Molyneux? with his reputation
as wikipeida states
'Molyneux has nevertheless acquired a reputation for issuing over-enthusiastic descriptions of games under development, which are found to be somewhat less ambitious when released.'

Quote: Milo was such a smoke and mirrors bullshit demo it was disgusting. Lots of people getting really hyped up over a demo that's 90% fake and 10% misleading.

I do find it odd though the number of ppl that apparently have brought into it hook line + sinker

overpromising and underdelivering helped wolfram alpha, didnt it?

Quote: Original post by Nypyren
I've seen a sensor system like this in person... it might have actually been the same thing. I was impressed. The distance sensor basically gives you the equivalent of a depth buffer to work with (which is a lot nicer for rough shape tracking than a color image). The depth resolution is pretty good and the capture rate is very high. At the time, it was hooked up to a mid-range laptop which was doing some very impressive 3D feature detection on the input.

The thing could recognize hand gestures at 5 feet easily.

The main thing I was concerned with at the time was the processor time required to handle the image / depth information. I could totally see it working on one of the 360's cores though. You'd never have enough processing power on a Wii (and still have time to run a game). In that case the sensor would need its own embedded processor, which would put costs too high.

You can't use the Xbox 360's CPU to do this stuff, though. Adding that much processing load to the CPU cores would probably break a lot of existing games. The sensor bar itself would have to contain a processor of some sort to process the raw feed in real-time and send the data back to the Xbox. I think Microsoft's going to have a hard time getting the cost of this thing down - I've heard that the cost would be in the range of $200.

I think Natal is great for everyone. Their demos showed it being the only input, which could be great for some things, but it could also serve simply to enhance other games.

For instance, the microphone allows you to give voice commands without sitting around with a headset on (something that I always hated).

The IR stuff gives better scanning and aligning of textures, like the skateboard in the demo video.

Face tracking could extend to knowing if the player's eyes or mouth are open or shut, or even advance to trying to detect expressions in the face. That could lead to different dialog trees or other things as you play the game.

A first person shooter could allow users to make swiping motions with their hands or controllers to do melee attacks or other actions.

Depending on the resolution and abilities of the device, perhaps going so far as to infer where the player is looking by tracking the eyes. Could be used to adjust depth of field to literally match where players are looking.

Not to mention all the nifty system-level stuff it can do. Having it sign you on just from seeing you. Maybe voice commands for playing music. I'm not real keen on the motion controls for navigating around, but it'll probably be nice for some people.

Anyone else a lazy gamer that doesn't want to get up and do a workout to play games? Back when I was working min. wage jobs that required standing all day. No way would I come home to play a game like this or the wii.

There really isn't much to say about the product at this point. I watched it on gamespot and they said the dev kits come out tommorow for it.

Quote: Original post by NickGravelyn
Depending on the resolution and abilities of the device, perhaps going so far as to infer where the player is looking by tracking the eyes. Could be used to adjust depth of field to literally match where players are looking.

Have there been tech demos where Natal does that? There's technology for gaze tracking already in use, but typically they need a very high-res camera image of the face (usually by sticking a camera right up close).

Quote: Original post by Trapper Zoid

Quote: Original post by NickGravelyn
Depending on the resolution and abilities of the device, perhaps going so far as to infer where the player is looking by tracking the eyes. Could be used to adjust depth of field to literally match where players are looking.

Have there been tech demos where Natal does that? There's technology for gaze tracking already in use, but typically they need a very high-res camera image of the face (usually by sticking a camera right up close).

No they didn't show anything like that. I'm not sure if their tech is up to it; it's just something I thought could be really cool if they pulled it off.

Quote: Original post by LockePick
Milo was such a smoke and mirrors bullshit demo it was disgusting. Lots of people getting really hyped up over a demo that's 90% fake and 10% misleading.

My exact thought.

In that Milo video, you can see the boy jump from a rock to another near the lake. It was far too natural to not have been motion captured. Even in games today, such a sequence would have been scripted.

Yet the video is made in a way to make the viewer think that the boy has complete freedom over the "area" and a smart AI to control it. Calling it "misleading" isn't strong enough.

However I'm pretty sure their tracking/face recogn/voice technology works relatively well.

Y.