First Look: Kinect for Xbox One in action
About Video Comments (0 ) Share (0) Transcript
First Look: Kinect for Xbox One in action7:22 /
Kareem Choudhry, development manager for Microsoft's Kinect, shows off new ways the Xbox One will track the user's body, including infrared imaging, face recognition, and heart-rate monitoring.
-All right. So, we're looking at the actual depth view. So what you can see is, you know, the 3D rendering of where he is in through your space. If you come a little bit closer, you can see like some incredible fine detail. You can see his thigh, I think that's your belt couple that we're looking at. [unk] you're good to go. If you bend down a little bit, we'll be able to see the detail in your face. You can open your mouth with your tongue out, three times in fidelity of what we have in the first sensor. It's really kind of the ground-- one of the groundbreaking areas that's active and [unk] in it really light things up, your tongue light, it's gotta be correct. -Yeah, I know. -Yeah. -I know. -Yeah, it can actually show that. -Yeah. So, the other thing I'd love to point out is incredible feel of you have, and you all notice how close he actually is to the sensor, which is a dramatic set form from Kinect D1 and you can actually see the sofa all the way our back. So, you now, living rooms of all sides, no matter where the furniture is, we're gonna be able to, you know, bring this experience to use. Usually forward over where we were in the first generation. The other thing I point out is the incredible horizontal field of you. If you wouldn't mind out walking from side to side, put your arms out, show us your wingspan. -Like this or-- -Actually facing the sensor. -Oh, so you want me do side to side? -Yes, side to side. Just to give people like, you know, an overall view of how big of an area you can see. So, everyone in the living room, no matter how, where you are, how your living room is oriented, where yourself is, you're gonna be in perfect space, pretty phenomenal. So now, this is a great view for all the technology that we run but it's not quite so human consumable. So, for that, we can do our 1080p HD color. That's, you know, 6 times of resolution of what we have in Kinect D1. You can see the incredible crystal clear high quality color key that we have. This will be the premier experience for spending in a living room. Everybody is involved. You can see everyone no matter where they are. The sofa, next to sofa, far back on the wall, we're really gonna be able to light up and task experiences both for e-mail and Skype, living room to living room communication and gameplay scenarios. So, this is a new feed available in the next generation of Kinect that we call Active IR. And what it's doing is it's using Infrared technology to see everything that's happening in the room. Actually, you can see the gentleman in the front. We can see all of you in the back. So, no longer that we have any kind of light independence whatsoever. Pretty fantastic stuff. I'm gonna back up once more for the color key and I'm gonna, you know, use my flashlight into kinda show another aspect of this technology. Make one, close your eyes for this, so I don't blind you. Typical scenario, you're gaming and you got a lamp right next to you. So, you can see how he's all bloomed out, we can't really see his face, it doesn't show up that bright. I'm gonna switch back to active IR, completely subtracted out. So, not only can see in the dark, we have complete independence from any of the light sources in the room. So, this is what creates a phenomenal consistent platform we're doing all of our computer vision and all our detection. You're all familiar with Kinect Skeletal Tracking, welcome to the next generation. So, as you can see, we've got our next generation skeleton happening right here. I'd like to point out just the incredible stability that we have, the more you move around, the more it's gonna represent what you're actually doing. So, you can shrug your shoulders, you can kinda lean forward, you can move your head back and forth, and we've got, you know, really good anatomical recognition. We've got our adrenalin over on the side there as you were taking pictures shown up. So, point out, you know, the accurate spine, the shoulders, all the expressiveness, all running in real time on our hardware platform. Other key joints that I'll point out is the hands. So, if you hold up your hands, we actually have joints that represent the tips of your hands and the thumbs as you rotate your hands around as you open up and as you close that, you can actually see that happening in the skeleton itself. Some pretty advanced stuff. So, we've got, you know, kind of great coverage of how old the bones are connected. This is how it all oriented next to each other. So, this is Walkman, you saw every glimpse of him in the video and as you move around, as you rotate your arms and your shoulder, you can pick up your leg and swing that all around. We can actually see you orientation being represented in real time and enough fidelity with the depth map, could be able to detect what you're accurately doing. So, pretty phenomenal stuff I might say. Have the sign, giving the way, kinda do whatever you want. Good stuff. So, this is Walkman. You know, but we didn't just stop there, you know, I'm 6'4". My scale only weights 8 pounds. If you wanna know what's going on with me and the body level, you gotta be able to understand all my muscle, all the forces that are going on, what torque am I experiencing as we will get the Muscle Man. So, Muscle Man is a human-based physics model applied directly on top of our skeleton and our high fidelity depth map. And what you're seeing here is a representation of the actual forces of you scaling in its [unk]. So, green representing no force and then red being the most intense force. So, as you shift weight between your toes and feet, you can see that we're detecting which foot he's got all his weight on the floor. You know, probably you do a little bit of a squat, you can see the force. He's feeling it in his thighs but we're seeing it on the screen. So, you don't have real time representation of what's happening on his body level. And the last thing we can use the technology for is you know, keeping track of momentum. So, if you can give me a nice tight little jam, nothing too big. You see that white dock, you know, so we're actually able to track what's going on. Now, if you really put your back into it and shift your weight, you know, once you make a circle, there we go, and you can see he's shifting weight from one foot to the next, have a really accurate representation to what's happening. All right. So, we have, you know, really great telemetry on what's happening with the body but as any good physical trainer will tell you, like this isn't enough, you will know its heart rate as well. So, what we're doing here is we're using both our color key and our active IR key to actually see this pulse in his face. There's very subtle fluctuations in both the intensity and in his skin color is imperceptible to the human eye that we're actually able to detect in real time to detect what its heart rate actually is. So, let's talk about the body but, you know, it's not just about the body. We've got these fantastic controllers as well. So, I'm gonna come join in, I'm gonna come in to give you, all right, I will hand you controller. All right. So, you may have missed it but it actually just identified me. He did say yes additionally and now it actually says green, I'm pre-enrolled on this box. So, back quickly, it actually identified who I was. I walked into the room, my Xbox tells me, we're both holding controllers. [unk] I give you the wrong one, you might switch anything. All right. Don't push any buttons. Give it a second. All right. Well, we can detect the real time, let's try walking again. All right. There we go. So, as we're stopping, controller we're able to detect in real time who's holding what controller, you know, we can kinda do in more [unk] on java and join us. Anyone else? Oh, come on. All right. Thanks. So, we're on here, we got the three of us. You know, we're having controller. They can hand you controller if you want. -Cool. -All right. So, controller one should go over there a little bit and it's gonna find you. The other thing that we can use, you can actually swap places. So, let's go ahead and swap. And we all kind of move around. So, in real time, it's actually able to detect who's on what side of the screen. And so, no longer do we have to swap controllers or swap on itself as it actually gets the screen to line up to how we are. There's nothing more annoying than being on the right time to solve and mind on [unk] screen.