What Facebook Gets With Oculus Rift Purchase

How the latest generation of the Oculus Rift virtual reality headset works and the new SDK that makes the magic possible

At CES in January, we got some dedicated eyes-on time with the latest prototype of the Oculus Rift virtual reality headset, code-named Crystal Cove. Just last week at the Game Developer’s Conference, Oculus released a second-generation Rift developer kit that builds Crystal Cove into something even closer to a consumer-ready device. It’s an entirely new experience that (almost) brings virtual reality into your living room, and with the news that Oculus was acquired by Facebook on Tuesday for a staggering US $2 billion, the future of consumer VR is even more of an inevitability.

The original Oculus Rift developer kit was released about a year ago. It featured a small cellphone-type screen and optics that feed a slightly different view to each eye to generate a stereo picture, with a field of view large enough to completely immerse the user. This visual system is combined with a very fast, very sensitive inertial measurement unit (IMU) that tracks head orientation. Combining immersive vision with a picture that updates itself based on even the smallest head movement results in a virtual environment in which you’re free to look around.

What’s exciting about the Rift isn’t simply that it has these capabilities but that the combination and implementation of them is done so well that it’s possible to suspend your disbelief and experience a virtual environment without having to suffer through compromises that are continually reminding your brain that it’s being tricked.

As impressive as the first Rift developer kit was, it was obvious that there was a substantial amount of improvement that could be made to both the software and the hardware. The Crystal Cove prototype we tried on in January featured a higher-resolution, lower-persistence OLED screen (a massive upgrade in quality), lower-latency head-orientation tracking, faster updating, and most notably, a head-position tracking system that enables users to move their entire heads from side to side and forward and backward, as well as look around.

Last week, Oculus announced the availability of a new version of their developer kit (DK2). Very similar to Crystal Cove, DK2’s HD (960 x 1080 pixels per eye) OLED display updates so quickly (75 hertz) that it virtually eliminates motion blur, and it incorporates the same infrared head tracking as in the Crystal Cove prototype, relying on an external camera and infrared LEDs embedded in the front of the headset to figure out where you are. The optics in between your eyes and the screen have been completely reworked, and there’s a USB port on top of the headset to encourage people to add their own accessories and hacks.

Having tried several iterations of the Rift from the first prototype through the current dev kit, I’m continually impressed by how much better the hardware is getting in such a short time. Part of this is that the Rift is piggybacking on cellphone technology, and as we start seeing phones with 4K screens (and eventually 8K screens), that’ll make the Rift that much more immersive. Additionally, developers are getting better at working with systems like the Rift to make virtual environments that aren’t terrible.

Virtual reality is very, very difficult, which is why it’s so hard to do well: We have a lot of practice in looking at the real world, and even the smallest discrepancy between our head movements and what our eyes see can be enough to cause instant nausea. You can’t settle for just “close enough” with VR; it really has to be perfect, without compromises.

Usually, saying that something has to be without compromises means that it’s going to be very, very expensive, but Oculus has been deliberately targeting an affordable consumer experience from day one. The first Rift dev kit was $300, and DK2 can be preordered for $350. However, as impressive as DK2 is, the consumer version (expected sometime in 2015) is going to be even better, Oculus says—“a leap beyond” DK2 with substantial improvements to the display, tracking systems, and overall design. We’ve been assured that there won’t be a third developer kit. The next major hardware release from Oculus should be an affordable peripheral designed to give everyone an immersive virtual reality experience, with content to match.

We can’t wait.

[ Oculus DK2 ]


Laird Malamed: We are creating the Oculus Rift, which is a virtual reality device where you can put it on and be in any kind of virtual world that you can imagine.

Laird Malamed: So you’re looking down on those trolls. And you can, you can try to—it’s like a tower defense system. So if they get through, they’re building an enemy base against you.

This is how it’s done.

It’s a hundred degrees freedom of view plus inside the Rift, and in your space here, you have about a 3-foot volume.

So the Rift is a combination of a sensor package, which sits on the device, and a camera prototype that sits outside the device to track all of your movements. So as you move around—I look, I bend in, all those sort of things—all that information is sent back to a PC. The PC processes it and sends over HDMI to a screen inside here—a cellphone screen actually sends the new image. So if I look down this way, that’s what I see. I look over here, I see that.

What people weren’t realizing or noticing is that when you breathe, when you talk, your head moves slightly. With the positional system in the Crystal Cove prototype, all of that is tracked. Inside the SDK is a body model. So we know how a body can move because there are certain limitations to the way our neck and our body moves. All the sensor data is then married to that virtualization of a body. That information then says, “Okay, well, if the eyes on that body are here, and they’ve just done that, tell the cameras of the game engine to move to that position.” And the cameras do that, they of course render the frame, that frame is sent back over HDMI to the device itself, and that’s what the user sees. Repeat and rinse.

This now new device takes us to a whole level further. So we’ve added Amoled 1080P screens. So it’s superhigh resolution, very fast switching times. We’ve added positional tracking. So you see some markers here on the device. These are being tracked by a camera that knows the shape of the device, and so as it turns, the device can understand the orientation that your head is going through. Adding that to accelerometers, gravitometers, and a gyro that are on the device itself and we have a really good idea of where your head is at any time. And we’re doing this 1000 times a second and then displaying the screen at over 60 frames a second, so that you get a superresponsive, low-latency experience. One of the things that we’ve added with the Crystal Cove prototype is latency was about 60 frames per second—60 milliseconds before. We’ve cut that in half.

So a question we get asked a lot is, “When are you freakin’ shipping it?”

Short answer is…when we’re coming out is soon-ish. It’s not years from now, but it’s not, you know, in the next few weeks or few months, either.