Project Natal vs Kinect: Fantasy vs Reality November 9, 2010

Update:  I have been corrected (by many commenters) about the facial recognition capabilities in Kinect.  I’m altering the article based upon those comments.  I had been misinformed.  I don’t have one and I tremulously apologize for the incorrect statement.

A little under a year before I started this website Microsoft at E3 2009 released this video during their Pre-E3 show.

The media and the internet shit a brick.  I remember.

A year and a half and one name change later the product is now on store shelves, and scene to scene you can clearly see that a lot of what was shown in that video could not make it to the final version.

Was the concept far too grand for Microsoft to deliver on?  Were the people who made the video completely clueless as to the technical capabilities of what they were promoting?  I’ll leave you to decide on that, but I will tell you specifically what parts turned out to be fantasy and which parts got it right.

First off, the opening, where the teenager is playing a fighting game with his whole body.  One player interacting with an AI controlled opponent is quite likely to find a place in Kinect’s library, and who knows, with Xbox Live you could have two players controlling avatars fighting each other from their respective homes.  Sure makes fighting in the parking lot after school obsolete.

Next came the driving game.  While you can see a whole family is watching the tv only two people are actually interacting with the game.  The girl is clearly seen driving and the father is doing the motions of the pit crew.  While this does fit within the rules of Kinect (two players only) the Xbox 360 does have a racing wheel which would work a heck of a lot better.  Hold out your hands for a few minutes and you’ll notice how quickly your arms tire without holding them on something.  The pit crew motion control is much more interesting here, and I think it would work a lot better matching one player with a control device while a separate one does a strictly Kinect motion.  Recent news has said controller plus Kinect interactivity is possible, so it is technically feasible.

Then there was the little boy playing as Godzilla and destroying a city in his wrath.  I think a game like this already existed requiring the vision camera, but I could be wrong.  I do remember XBLA games where you’d have to stand in front of the vision camera and swat balls headed your way and move things around, and the motions the boy is using is very similar.  Please note the vision camera had problems capturing the imput correctly sometimes and Kinect shouldn’t have that problem.  Regardless, one player being captured, perfectly feasible.  It gets a pass, but it really doesn’t look like much fun.

The next video, a girl and her mom playing what looks at first to be Dance Central, but then turns out to be a generic split screen soccer game.  I think that’s already been made, and it fits with the two player limit.  If that game’s not already out, it should be out by Christmas.

Then comes the infamous skateboarder video.  The boy clearly is seen scanning his skateboard into the Xbox 360 to play a Tony Hawk Ride-esque skateboarding game without a board.  One would think if the boy knew already how to use a skateboard and already had a skateboard he would just…oh…GO SKATEBOARDING!  That aside, the Kinect has a 640×480 camera in there, not exactly high-definition by anyone’s standards.  Capturing an image of the board’s appearance from that distance would not give a very sharp texture to apply to the virtual board, and it would be very apparent on the family’s HDTV that the board logos would’ve been blurry when upsampled like that.  You also see some voice command work here when the skateboarder orders the Kinect to “SCAN” and it responds on command.  The final version of the Kinect does in fact have a mic…but voice commands are….well I’ll talk more about it later.

Ahh yes, then we get to the infamous teenage girl scene.  The Xbox in the video clearly recognises the girl’s face and loads her profile.  This is one of the famous bits that made a lot of nerds go crazy, and it is pretty cool, and I’ve been told that the Kinect can in fact do this, so it gets a pass.  As she talks to her friend who also happens to be conveniently using her Xbox at the same time (and she also conveniently has a Kinect installed on her own 360) they rattle on about dresses, and wave their hands around a conveniently placed dress shop which suddenly is selling their wares on Xbox Live, and superimposes a dress on a 3D image of herself before deciding the first one she virtually tries on is the one she likes.  This scene is a complete lie.  I don’t know one woman who likes the first dress she tries on, ever!  Heck I even know a girl that will spend hours dressing her Rock Band avatar with virtual apparel after playing one song, at the expense of the time of an entire group who’d rather just be playing songs.  The Xbox will be able to have full voice and video chat from multiple Kinects allowing for two people to have a full-duplex video chat from their living rooms with voice.  They also claim to have limited voice command support, so telling it to contact a friend online is certainly possible, but voice accuracy technology is still iffy.  Try telling an automated phone service where you’re trying to reach and you’ll likely get an operator every time since the service is useless.  Then there’s the whole dress shopping experience… a total pipe dream.  The framework to have an entire dress catalog does not currently exist in Xbox Live (unless you’re dressing virtual avatar clothing), and there are no real world stores which have deals to sell physical goods through the service.  You can also clearly see that the dress is being stored on a stored 3D image of the girl and not her current physical state on the real time video chat.  While Kinect might be able to take a 3D image of you, there’s just no reason for a user to do it.  You would have to stand in front of the camera and do a very specific series of turns that the system would prompt you to do, and then it would layer it onto a 3D image based upon your 3D frame taken from your current motion capture state.  Plus, since the camera’s resolution is so low, it would not stand up to being displayed in full 1080p and would have to be kept windowed or it would blur.

Next is the game show.  You clearly can see four people on one family’s side playing through Xbox Live with three people of another family in a generic trivia contest.  Aside from the fact the teams are imbalanced (which is kind of unfair) you can clearly see more than two people on each side playing.  Kinect can only support two people, so having a whole family doing a trivia contest is kind of a fantasy.  It’s also kind of unnecessary since trivia games on Xbox Live Arcade already exist using a controller, and would be better suited to just pushing the A button, even my mother would be able to figure that out.  The fact the system would accept voice responses for the trivia questions is kind of neat but time will tell how good that voice recognition software is.  Also we know that Chuck Norris’s push up count is still climbing as I write this, so that’s a trick question.

The movie selection system comes next, and you can clearly see the arm control system in place, another scene that made nerds go crazy.  This is not a fantasy, this is now a part of the Xbox 360 system, as is the full selection of movies to watch in full 1080p through Zune Marketplace and Netflix.  Technicians and programmers have been swapping back and forth about if Kinect can support arm motions while sitting or standing, originally it couldn’t, but after E3 it could…can we get a ruling on this?  Please post a comment if you own a Kinect and know if it can or not.  In the last bit of the video you see the woman telling the Xbox to shut off.  People interpreted this that you’d be able to turn the Xbox 360 on and off just by voice command.  You can’t do either.  Update:  The Fall 2011 update has added the ability to turn off the Xbox 360 by voice command.

So thank you for reading through this meaty article comparing what was originally promised with Kinect back when it was called Project Natal and what is actually coming with Kinect.  If you don’t get anything out of rewatching the video a year and a half later, at least you get to hear some of Wolfmother’s “Joker and the Thief”.