Varjo Promises Human-Eye Resolutions For VR This Year

by Kevin Ohannessian • June 19th, 2017

I am standing in a virtual living room rendered via Unity. Except I am seeing things I never have before in VR. There is a virtual television on the wall to my left displaying a 4K video of a city. The floor, the couch pillows, the clothes hanging on a rack to my left, are shown in such extreme detail that I am actually seeing a life-like world. It’s a glimpse of VR’s future.

“We believe that by chasing for the human-eye resolution as fast as possible, it jump starts everyone so they can work toward the end game of VR and AR,” said Urho Konttori, founder and CEO of Varjo.

Varjo is a startup of 19 employees that is only 10 months old based in Helsinki, Finland. The talent comes from the likes of Microsoft and Nokia. This work they have done comes from their original $2M funding, with a second round of funding being pursued currently.

My experience in that Unity-based living room and other demos, using what Varjo is calling the “20/20” prototype, shows that they are working in the right direction.

They retrofitted an Oculus Rift with an extra layer of lenses inside the unit. An OLED microdisplay projects an image onto a glass plate over the Oculus lenses. That center piece is said to have 70 times the resolution of the Rift’s image.

Below is a comparison.

“The comparison is apples to apples, taken through the same sense with the same SLR. It’s a world of difference,” said Konttori. “From our point of view, looking at what professionals need from VR, you need human-eye resolution.”

This prototype is a proof-of-concept demo to be sure: the size of the ultra resolution section is on the small side, about a small fraction of the real estate of the entire VR picture (see my approximate illustration of what it felt like below). The edges are a bit jittery in the transition from one image to the other, because the microdisplay has persistence versus the Rift’s low-persistence display.

But, even with that limitation, the demos give you a sense of what VR and AR at such crazy resolutions would be like. A virtual Windows 7 desktop had file explorer open, a 4K trailer to Guardians of the Galaxy Volume 2 and AutoCAD. Whatever was centered in my view was astonishingly detailed.

A video of a prototype back in Finland showed how this tech can work with AR. Using video see-through, a person could have multiple “displays” while they work in the office, at home, or on an airplane. Another example demo from video shows someone manipulating what looks like a model of a home floating in front of them, moving furniture around, for example. Then they click a spot and the virtual home surrounds them at human scale. It looks like you are standing in a VR house that is as detailed as any that could be in a 4K video game.

Varjo wants to address the limitations of current tech.

“The problem with the VR and AR of today is, firstly, that they are all inherently low resolution. We’re talking about one one-hundredth of human-eye resolution. It’s not quite legally blind,” said Konttori. “Optical see-through systems…have narrow fields of view. It’s crippling all the time. And everything you create in AR is very hazy and artificial, ghostly. It’s not life-like. If you’re  a professional who wants to showcase a product in mixed reality, it needs to be accurately represented and not as a holographic image.”

The company is aiming for a professional market, aiming to use its tech to create a SteamVR headset that companies will use. They think they could use software and hardware gaze-tracking to move the 70X resolution section around to where you are actually looking, and not just locked in the center of the screen. The microdisplay could physically move or maybe the glass plate that receives the projection could tilt — Urho Konttori is cagey about what the final mechanism will be be like. But he wants to be clear that this isn’t foveated rendering.

“Foveated rendering is used just to decrease the quality and thus decrease the GPU consumption. Where as we are going to enhance the quality. Maybe it’s ‘foveated display’ or something like that,” said Konttori.

He stresses that the total number of pixels being pushed is less than a 4K computer monitor, and so the software can be run today on current desktops and gaming laptops.

There current schedule is to have a prototype of their own headset by the end of this year and giving away hundreds of units to partners. After some feedback and further development, the final headset is supposed to come to market by the end of 2018.

And the cost of a VR/AR headset with human-eye resolution aimed for business use?

“This will be multiple of thousands in price, but not tens of thousands,” Konttori said. “We can not say the price now because that depends on so many things between now and then. This is achievable for prosumers, but you need to be quite committed to shell out multiple thousands. But technology trickles down over time. It’s not going to be that many years until this will be available at consumer price points.”

Before they get to that consumer unit, which he says should be a standalone, untethered headset to get wide adoption, they need to have the enterprise headset finished. And that requires the aforementioned prototype coming later this year.

“It is in many ways a dev kit, but we don’t want only software developers to use it,” Konttori said of this year’s prototype. “It will also have certain limitations. We haven’t decided which compromises we make. And sometimes dev kits are also perceived to be very close to the quality of the expected product, but there will be some things that we will be improving during the next year.”

“One of the most important things is that we want to keep the schedule. We want to make it that we bring this to the attention of many interested companies, to start the next transformation of computing together.”

What's your reaction?
  • Kacey Sherrard

    your screen door effect youve applied to the picture is a bit of an over exaggeration isnt it? 😛

    • Adderstone VR

      In this article they claim that it’s photos taken with the same camera through the Oculus’ lens. (Thus not an applied effect)
      Considering the graphic showing the tiny area of the Oculus’ field of view that is enhanced to this clear image…I think these images are zoomed in quite a bit and thus the screen door effect looks so over exaggerated.

      • Kacey Sherrard

        ah k that would explain it, but as someone else mentioned, it implies to people (who are unaware) that that is the resolution of the vive/rift, which we know of course isn’t the case, it could just give non-VR users the wrong idea. It would ahve been better to show an example with a bigger picture and then another showing a zoomed in portion of that bigger picture. purely for people who havnt tried VR yet so they dont think the resolution on the current headsets sucks 😛

        • Hugh Bitzer

          I wouldn’t say it’s exaggerated at all. Pixellation is still extremely visible inside the CV1 (assuming you have decent/normal eyesight). It’s easy to stop noticing after awhile (which may be why the image seems enlarged), but there’s no denying the first impression is still there.

          • Kacey Sherrard

            well on the latest vive, the only time i see the screen door effect is on cheap indie games that didnt optimise the 3d effect or are lacking tessellation, or on white screens. I have 20/20 vision too 🙂 And its not ‘seems enlarged’, the imagine is enlarged and is only a small section of the visible area, the guy commented back to me to explain this.

  • Matti Pouhakka

    Way to go! Varjo ROCKS. And we make the content

  • Rigelleo

    With mirrors MEMs it may be possible and it’s a very clever idea!

  • It’s going to be utterly amazing when consumer headsets are capable of this kind of resolution across the view area, as well as getting field of views closer to the human equivalent and all that kind of thing.

    • 1droidfan

      By the time that happens, as well as rendering tech that can render photorealistic images at 90fps+, they probably will have ditched headsets and will simply transmit to your optic nerve via a wireless signal.

      • Tyler Soward

        I’d be afraid the system would crash at some point and scramble my brain LOL

  • Buddydudeguy

    The images are unfairly exaggerated in it’s low res and screen door effect-ness. But ya, I can’t wait for higher res and foveat rendering.

    • daveinpublic

      It’s not over exaggerated, it’s just zoomed in on a portion of the view. They did the same to their own example, it mentions it right in the article.

      • Buddydudeguy

        Exacttly. It’s zoomed in. Which is exaggerating it.

        • Robert Cole

          I had a great experience yesterday with a complimentary 1 hour session of ‘Star Trek Bridge Crew’ on the Vive.

          The most noticeable visual artifacts were god reys from the fresnel lenses and noticeable jaggies on the 3 NPC (AI) crew and lesser jaggies around some of the environment.

          Some AA / super sampling would have cleaned up the latter, but I don’t believe their PC had a GPU with enough grunt for that.

          Still a great experience (Kobayashi Maru!) and resolution was pretty good overall

          Even the multitude of control panels and overlays were clean and relatively crisp, I did however start to feel some eye fatigue by the end of my session

          this is not something I’ve experienced before in 100+ hours in steamVR and openVR apps, perhaps from having to look up and down at the menus constantly from seated position

          • Buddydudeguy

            It’s a OK looking game. Pretty sure SS doesn’t work with it though. I notice no difference.

  • MowTin

    Is this a form of floveation? Wouldn’t the high res area have to track your eye movement?

    • Jack H

      Yes, it is! It’s optical or hardware foveation, where the foveal display should be shifted around to follow the eye.

  • daveinpublic

    Looks like all those predictions are not as far out as they thought. 10 or more years till we get full resolution for our eyes. Ha! They underestimated the exponential speed of today’s progress. They aren’t filling your entire view with these pixels, but still, putting them where it counts, in the center of your gaze, is all that matters. You can bet that Oculus and Vive are paying attention to this.

  • Cool project! The problem, as lots of people has told on reddit, is that they haven’t showcased the actual product, with the “moving display” nor they have explained how to achieve this. So we have to see if they’ll be able to make it work.

    • Rigelleo

      My Bet is for mems mirrors, it may not be so difficult.

      • Sean Lumly

        This follows my intuition exactly. One projector projects an image onto a flat ‘screen’ that will be the VR display (similar to most HMDs) and masks out the foveated region. The second projectors output is positioned with a MEMs mirros and renders into the foveated region at a much higher resolution (albeit a much lower area) at the center of view.

        In the end, the total pixel count shouldn’t be that much different than what current headsets render, but without the need to have an ultra-high 70MPix resolution display capable of rendering a sharp foveated region wherever the center of vision may be.

        In the end you *see* an ultra-sharp image, but without the need to have an ultra high resolution display/projector.

        This is my guess.

  • MowTin

    Can you imagine crystal clear VR?

  • Jonathan Jones

    Super high resolution displays already exist. The reason they aren’t in the market is the price, and that you can’t push enough pixels without foveated rendering technology, which isn’t quite there yet. This company is saying that they have created an ultra high resolution display… but… it’s too expensive and they need to implement foveated rendering. So… literally nothing new. Perhaps a new technique. But nothing with a future. Ultra high resolution displays are already cheaper than this I believe.

  • wowgivemeabreak

    I hope this works out and if not or in addition to, other companies are able to do similar stuff. I love vr/the rift but damn, the idea of having more real life clarity gets me all warm and fuzzy inside whenever I think about the future of vr along with an untethered headset. It better not fail before then.

  • James Friedman

    ironically I was just thinking about how long it might take for crystal clear visuals to hit VR headsets. My guess was a good 5-10 years till the screen and the hardware to drive those displays are perfect. Hopefully sooner

  • Guilherme Sampaio

    I WAJNT rwdtaewfgsaerfg gimme gimme gimme