Has This Stealth Company Solved Vision-Quality VR?

A new technology for VR headsets puts the so-called screen-door effect in the past.
Has The Stealth Company Varjo Solved VisionQuality VR
Varjo

When Urho Konttori handed me the VR headset, I almost laughed. The founder and CEO of some Finnish company I'd never heard of had just told me he and his team of 19 people had managed to leapfrog virtual reality 20 years into the future—and he gives me an Oculus Rift? "It's just the housing," he said. "We added some things inside." Fine, I thought. You've seen plenty of demos where the reality didn't match the hype. Just do it, then you can go back to the office. So I put the headset on.

The demo itself was quick, maybe 10 minutes, and consisted of a series of static VR environments that I could examine at will. There was a simple room with a TV in the corner streaming video; a shapeless environment with some floating computer monitors; a plane cockpit. Because this was an Oculus Rift, the image quality was exactly what I expected it to be: fine. However, a small clear rectangle was there as well, sitting in the middle of my field of vision. If I looked at something through that small rectangle—the text on the virtual computer monitors, the tiny numbers in the plane's instrument panels—it stopped looking like VR. It just looked like...well, like real life. And it's the first step in Varjo's plan to create ultra-high-end headsets—for corporate use at first, but someday soon, for civilians like you and me.

If virtual reality's recent evolution has seemed like a blur to you, that's probably because it is. A lot of companies have engineered a lot of smart solutions to VR's many bugaboos—latency, tracking, input—but they've all run into the same limitation: the display just isn't that sharp. Sure, it's good, and OLED screens keep getting better, but the resolution is still a far cry from anything even resembling human vision.

Think about: the closer a screen is to your eye, the more obvious its pixels are—and in a VR headset, the screen is less than two inches away from your face. The clearest thing about it is your realization that you're staring at a screen. That's why people still complain about the "screen door effect," the annoying visibility of the space between the pixels that can make things in VR look like they're hiding behind gauze. It's also why VR isn't a super-likely candidate for desktop replacement anytime soon: you might be able to hang monitors all over a virtual environment, but if you can't read the text clearly, then it's all for naught.

The ultimate goal, of course, is to be able to match the resolution of human eyesight. In order to give you an image that’s indistinguishable from real-life vision, VR needs to display more than 2,000 pixels per inch. That’s about 10 times more densely packed than a Macbook’s “retinal display”—and according to Jason Paul, who oversees VR strategy at NVIDIA, it’s also going to be long time until we get there. "Based off how many pixels we would need to be able to push and mapping that out alongside our upcoming new GPU releases," he told Upload VR in May, "it would take us about 20 years to achieve resolutions that can match the human eye.”

The folks at Varjo didn't want to wait that long. As ex-Nokia employees (Finland represent!), many of them were familiar with small-screen optics, and the way they saw it, there was already a display out there that could simulate human vision. Sony's broadcast cameras use an OLED microdisplay that's only 0.7 inches diagonally, but manages to cram in a full HD resolution:1920x1080. That works out to more than 3,000 pixels per inch—enough to let you see everything there is to see, without eyestrain or effort.

Varjo

Varjo's early prototype—the one that I saw—mounts the microdisplay on the top of the headset, and uses a mirror inside the headset to reflect it onto the pre-existing display. The result is clarity of a kind I've never experienced in VR; I could read virtual monitors as though they were sitting on my desk, and examine the cockpit readouts like I was in the plane. In the prototype version, that view is fixed in place, but Varjo is hard at work marrying the microdisplay with gaze-tracking technology, so that you get that ultra-clear view regardless of where look in your field of vision. (Other VR companies are implementing what's called "foveated rendering" in order to lessen the graphical load: by only delivering peak resolution where a user is looking, a headset can deliver lower-resolution graphics anywhere else, where your peripheral vision won't be able to tell the difference. Varjo isn't downgrading the rest of the graphics—it's vastly upgrading the area you're focusing on.)

There's one catch, though. That Sony microdisplay? It's $900 a pop, considerably more expensive by itself than any other VR headset on the market. That's why Varjo is targeting enterprise rather than consumers: The first version of this thing is going to cost in the thousands of dollars. Later this year, the company will be releasing a technological preview unit—no strap, no bells and whistles, purely to be used as a development tool—then giving it away to would-be partners. Architecture, real estate, auto makers, hospitals. Next year comes the headset itself. And no, you're not going to use it to play games or hang out with other VR users; you're going to use it to train as a pilot, or design buildings. Real stuff.

Varjo sees the headset as more than strictly VR, though. I also saw a video demo they filmed through the headset, combining video passthrough with the microdisplay to create lifelike holograms not unlike Hololens' or Magic Leap's. The augmented- and mixed-reality applications are likely farther away, but with an optical workaround like Varjo seems to have figured out—and the fact that they build all this in a mere 10 months—it's not inconceivable that their tech trickles down to the consumer level within two or three years. One step closer to the Holodeck, no screen door required.