Elegant, you say.
I'd say elegant².
Fantastic. And to think that facebook spent $2b buying Oculus Rift.
Interesting, I wonder how well it works?
From following the Oculus Saga with moderate interest, they appeared to be dealing with the fact that VR (like CGI humans that aren't terrifying plasticky horror dolls) is apparently what project management types (in an all-too-accurate bit of gallows humor) call "90% done, 90% remaining"; and their prototypes have been gradually sprouting extra bits and pieces(like the IR head tracking stuff) in order to keep users from running into nausea too quickly.
Are accelerometers and other sensors in phones good enough that, with a bit of software witchcraft, this actually works OK, or will you discover, should you try to push this beyond the 'neat tech demo' time limit, why VR headsets, even the slick new ones, not the wildly expensive legacy ones sold to customers with deep pockets, cost more, are bigger, and have more sensor systems?
Here's a snag: Google don't sell a kit, and tell you to buy replacement lenses for the Durovis OpenDive, a moulded plastic version of this thing. And they're like $25...
Here's a cheaper option, which sadly renders the cardboard craft session moot. I really like how Google encouraged folks to DIY, but with the lenses they should've lined up a supplier and had them for sale in the Play store.
BTW, speaking of stereoscopic vision, here's an interesting article I read today: http://www.bbc.com/future/story/20120719-awoken-from-a-2d-world
From that Ebay ad:
Item location: Yerevan, default, Armenia
The original Rift was just pretty much just this, only with ski goggles instead of cardboard.
This seems like a neat toy, but I would hardly call it "elegant".
That looks like a nice Crackerjack Box solution, but really VR is all about the details and there is no way this thing has gotten all of the details right. Putting a screen in front of your eyes is the easy part, building a very low latency and high accuracy head tracking tracking system is difficult. Integrating that into a low latency renderer is even more difficult.
If you're curious what kinds of things are explored (and often discarded) in this quest, check out John Carmack's twitter feed:
His twitter handle is an object lesson in not putting your company name in your personal feed. He doesn't work for iD anymore and Armadillo Aerospace is defunct.
My guess is the point of this is to get the mainstream news to report "Hurrr durrr. Facebook paid $2 billion for a company and Google has duplicated their technology. How? With some cardboard. Now I'll read the teleprompter repeating Google's press release, mentioning the things Google actually wants reported today."
Next up, Google Glass: Styrofoam Edition.
And wait until you see the Galaxy Tab in paper mache!!
Plastic and metal have had their time in the sun, it's time for cheaper, lighter, way less durable materials to step up.
Not even close.
This seems a lot more like Google trying set up their marketplace as a hub for VR content before Oculus/Samsung cut them out of the loop.
I'm also annoyed that they fail to acknowledge the FOV2GO project which has been around for two and a half years and was one of the things Palmer Luckey was working on before founding Oculus. Cite your sources, Google.
Barring something vaguely more put together, at least as a demo, I'm inclined to go with @Jardine's publicity-snipe hypothesis. I don't doubt that they can achieve View-Master quality without difficulty(since, after all View-Master did so sometime before 1940 with nothing but toy-grade plastic), quite possibly better because a backlight screen is nice and bright and bold, and you aren't confined to static images.
What is less clear is whether there is much of a market for '3d' that is neither unobtrusive-invisible-magic nor serious-techhead-nerd-stuff-but-damn-does-it-work. Given how...delightfully... '3d' worked as TV and movie ticket price increase gimmick, I'm not optimistic. The slightly nutty hardcore set are going to want VR that is indistinguishable from magic, which is something of an asymptotic goal. The dabblers are going to lose interest, fast. They might move some units on the 'worth what you pay' grounds; but that's only because you don't pay much.
I suspect the big challenge will be figuring out a way for people to try it out before buying (for something like the Oculus Rift I mean). I might be willing to buy a VR headset so I can fly around in 3D in Elite: Dangerous and Star Citizen going pew pew pew, but it would really suck if it makes me nauseous. Or if it doesn't fit comfortably over my giant head and glasses. Or if the resolution isn't good enough.
Basically, you gotta make friends with an early adopter right now. The Rift isn't ready for primetime yet, but it's getting close. The model that is out there now is a great research platform, but it's not polished for home use. It's a hacker toy. In a year or two they should have a mass consumption model that is shaping up to be a great product.
I would disagree that it's "not even close." I'm willing to bet that the most accurate and important sensor on that device isn't the accelerometer, it's that big window on the back for the video camera.
Video processing is very powerful these days, and motion-tracking is a pretty fundamental part of it. By itself I expect that would be enough to get very accurate orientation information, and with the error-correction of the accelerometer it's even more accurate.
Add to that the double-lens 3d cameras that are coming out on some phones these days, and you have the promise of very fast, very very accurate motion tracking.
One big downside of using a phone is that VR requires a lot of GPU grunt, and that's something phones just don't have. It takes a lot more GPU resources to render two 960x1080 scenes than it does to render a single 1920x1080 scene for various hardware reasons. A current gen Oculus requires a pretty beefy computer to run it, and it only has effectively a 720p screen. And this is after years of optimization and tweaks (if there's anyone who can tweak the software to make it run faster, it's John Carmack). There are even some doubts that a PS4 will be strong enough to drive the thing.
This cardboard headset is a toy. If they find a cheap supplier for the lenses so it can be given away or sold for like $5 to give people a taste of VR, then that's great, but anybody who thinks this makes the Oculus obsolete is deluded.
I don't doubt that this will be inferior to the occulus for many reasons, not least because it's literally made of cardboard.
I guess I shouldn't have replied to the overly-narrow question of whether a phone's sensors can provide accurate information about a phone's orientation. I think they can, as I said, but it's a bit of a side note.
I haven't been able to try any of Google's software as I have no phone (can I make one out of cardboard?) so you may well be right. It seemed to cause enough of a stir to get many developers to start working on things though.
I don't think it's going to grow as rapidly as some people think, but I do believe it's here to stay (for realsies this time!) and will slowly but surely become an important part of many fields outside of entertainment. I'm a slightly nutty hardcore type, but I don't think it needs to be The Matrix. It just needs to avoid being nauseating, and that point is getting close. Unless you really hate Facebook.
I've been saying for some time they're going to need to set up showrooms so people can take them for a test drive, not only for the reasons you outline, but also because it's nearly impossible to accurately convey what it's actually like.
The VR community is doing a pretty good job filling that role at the moment, and you could probably find someone local who'd be willing to demo a Rift if you wanted to try it. Give it another month or so and then check out the second development kit though. Elite is stunning, but far better on the newer hardware.
It's accurate for orientation, but too slow. That's not a huge thing to fix though, there's nothing stopping anyone from producing an IMU which matches the performance of the one in the Rift.
Translation is harder. Assuming the lighting's good and there are enough reliably identifiable points, it would be possible to use the camera to correct errors in the double integration of acceleration, but latency is still an issue.
Changes need to be reflected on screen within about 20ms to prevent nausea in most people. When rendering is likely to take a large portion of that time your options are pretty limited. John Carmack has come up with some good stuff for that though, so it will happen at some point in the future.
There are still plenty of cool things you could do with this, I'm just slightly concerned people will try it and assume it represents VR as a whole rather than just the thrifty tip of the iceberg.
I'd say the most useful thing is built into the Cardboard app.
YouTube: Watch popular YouTube videos on a massive screen.
There are 3D videos on Youtube, but I bet this thing might actually work fairly well for watching regular video while laying down without holding the phone up.
Won't somebody think of the Giant Heads!? I feel your pain fellow melon-topped one.
next page →