FOVE Eye-tracking Headset Gets Final Specs and Pre-order Date

42

FOVE has announced that their first eye-tracking VR headset, the FOVE 0, will open for pre-orders on November 2nd, and has also released the final specifications of the device.

Much like Oculus, FOVE began as a successful Kickstarter which raised $480,000, nearly twice its goal, back in mid-2015. The company went on to raise an $11 million Series A investment in 2016, and now plans to open pre-orders for their first headset, the FOVE 0, starting on November 2nd.

fove-0-vr-headset-5

The headset’s unique differentiator is built in eye-tracking, which the company says is accurate to within 1 degree, and fast enough for foveated rendering (a technique which improves rendering performance by only drawing the scene sharply where the eye can detect high detail).

FOVeated-rendering-SMI-2
See Also: NVIDIA Says New Foveated Rendering Technique is More Efficient, Virtually Unnoticeable

The benefits of eye-tracking are potentially many; FOVE promises that the headset will be able to detect the player’s eye movement and represent it within VR to make their avatar more realistic. They also say that the user’s gaze can be used to make UI selections, target enemies, create depth of field, or trigger context-sensitive events that only happen when the user is looking in a certain location within the virtual world.

FOVE hopes to be the first VR headset on the market with eye-tracking, and we expect other major headsets like the Oculus Rift, HTC Vive, and PlayStation VR to add eye-tracking in future product iterations.

SEE ALSO
Eye-tracking Glasses Give Glimpse into Professional Pianist's Perception

fove-0-vr-headset-4

You’ll of course note the zero in the headset’s name. This denotes the FOVE 0 as a sort of precursor headset, which the company tells us is “primarily designed to be for VR hobbiest and developers.” Launching a developer-focused version of the headset first is similar in approach to Oculus who launched two development kit headsets (DK1 and DK2) prior to their consumer VR headset.

Along with the November 2nd release date, the company has announced the final specifications of the FOVE 0 headset.

FOVE 0 Specifications

  • Display: WQHD OLED (2560×1440)
  • Display Refresh Rate: 70Hz
  • Field of View: 90 ~ 100 degrees
  • Tracking:
    • Head Tracking
      • Rotational
      • Positional
    • Eye Tracking
      • 120FPS infrared x2 (accuracy <1 degree)
  • Weight: 520g
  • Audio: 3.5mm audio jack
  • Connections: HDMI 1.4, USB 3.0, USB 2.0
  • Accessories: Positional tracking camera, face cushion

Compared to the 2160×1200 displays of the HTC Vive and Oculus Rift, the 2560×1440 display of the FOVE 0 has 42% more pixels, on par with Samsung’s Gear VR headset. Unfortunately the display only has a refresh rate of 70Hz, which comes in significantly below the 90Hz of the Rift and Vive (not to mention the 120Hz of PlayStation VR), and slightly lower than the 75Hz of Oculus’ Rift DK2 development kit. A lower refresh rate means more latency and less smooth motion inside the headset. However, with the success of Gear VR, even with its 60Hz refresh rate, 70Hz ought to work fine as a starting point for FOVE 0, though it’s an area we expect to see improve in future versions of the headset.

fove-0-vr-headset-2

Robust positional tracking has historically been very difficult for companies other than leaders in the VR space to achieve, but FOVE says they’ve created their own positional tracking solution based on an infrared camera and (we presume) infrared markers inside the shell of the headset, similar to Oculus’ ‘Constellation’ tracking system. While in previous hands-ons with prototypes of the headset we found the rotational tracking to be on-par with leading headsets, we’ve yet to have a chance to try the company’s positional tracking solution.

SEE ALSO
Vive to Get Eye-tracking Add-on with Optional Corrective Lenses

For those with glasses, the company says that FOVE 0 will work with clear contacts, but they don’t recommend colored contracts or wearing glasses with the headset as the frames can block the infrared eye-tracking. It’s unknown if the headset has IPD or lens-to-eye distance adjustments, but we’re in touch with FOVE to find out.

fove-0-vr-headset-3

In addition to showing their latest prototypes at the Tokyo Game Show this week in Japan, the company has announced that the Xenko game engine has committed to official support of the FOVE 0 headset. At this time, it sounds like the company will ask developers to build for the headset using their own proprietary SDK rather than tapping into Valve’s OpenVR platform to enable compatibility with SteamVR games. This makes sense of course, because the company needs to give developers eye-tracking data to work with, which isn’t something encountered on other consumer headsets to date.

FOVE encourages those interested to register for the email list on their website to be notified the moment that pre-orders launch on November 2nd. So far the company hasn’t announced a price or final release date for FOVE 0.

This article may contain affiliate links. If you click an affiliate link and buy a product we may receive a small commission which helps support the publication. See here for more information.


  • VRgameDevGirl

    I would love to play Around with eye tracking, so many ideas.

    • I’d love to see a game where you aim with your eyes, or maybe even something like virtual desktop using your eyes as a mouse.

      • Bryan Ischo

        Why would aiming with your eyes be significantly better than head aiming, which most people thinks sucks? In current headsets you tend to want to keep your eyes fairly straight forward anyway, because of poor quality lenses and limited field of view, which makes head and eye aiming almost identical.

        • I didn’t say it would be better. I just think it would be cool. Also I don’t think it would be as bad as you think, even the dk2 had pretty decent lenses.

          • Bryan Ischo

            The DK2 actually had *better* lenses in most respects. But all current VR headset lenses suck, some more, some less, in my opinion.

          • Nevertheless, my original suggestion still stands. I’ve always thought it would be cool to aim with your gaze.

  • Ainar

    Can’t wait for this one, but they need to do something about the glasses issue and of course the refresh rate. I guess I’m going to have to wait until the consumer version.

  • I love this tech, it’s wonderful and I’m certain it will be part of the future of VR…. assuming there is a future for VR.

    This is one of those “Cart before the horse” things. VR sales have stagnated. Once they sold all the units they could to developers and enthusiasts, demand just sort of dried up. Much as I predicted, when the $600 price tag for the Oculus was announced, high prices have all but killed mass market interest. Unless Playstation VR lauches with overwhelming demand, we’ll see a VR crash again mid-way through next year. Samsung’s GearVR is considered an “overwhelming success”, but only a fraction of Galaxy owners own one, and fewer few use it regularly. And if you own a Galaxy phone, it’s only $100! Does anyone REALLY see a future on the PC side with prices well north of $500?? Unless those PC VR units reach $300 or less, you’ll never seen any mass adoption. It’ll fade into a niche again.

    I sincerely hope this all works out and there is a bright future for VR. I hope that tech like eye tracking becomes common place. But if we, as a community, want to see that happen, we should “encourage” the hardware makers to pursue MUCH lower prices over fancier tech.

    • Andrew Jakobs

      uhh.. here you get a GearVR with every new Samsung Galaxy S6 and S7 (at least when you have a subscription)..
      And I do see a future even if the headset (with controllers) is short north of the $500 border, but at the moment it’s just for early birds and developers.. Next year you’ll see cheaper headsets.. It is just like with the smartphones, at first the good ones were very expensive and not for the masses, and not much content was available for them.. Years later they were ‘cheap’ and content is booming…
      At the moment content is also the bigger problem, there are some games, but not enough YET.. Let’s not forget ‘WE’ aren’t the masses, and currently the major headsets are targeting ‘US’ not the masses, that’s for the next incarnation..

    • Dave

      Same thing happened in the 90’s. FOVE is so Full Of Vapor even their specs are vague and their video clip just goes to show how unhealthy it would be for their models.

    • One huge benefit of foveated rendering via eye tracking tech is it requires much less GPU power to deliver a full frame of video which (hopefully) looks exactly the same as if it was a full resolution frame to the viewer. That is the top benefit of eye tracking.

      This means the recommended PC hardware required to run VR on this headset will be lower than the current minimum GTX 970’ish. Also, a higher resolution but lower Hz could work and open the door to many more people.

      It would be interesting to test if consumers can really feel the difference between locked 90fps vs locked 70fps though.

      • MosBen

        Well, more likely is that the next generation of VR HMD will see a significant jump in graphical fidelity, or FOV, etc, without needing a comparable jump in hardware.

  • Joe Bazaar

    Foveated Rendering WITH Depth of Field is a very big deal! That’s what lightfield displays are promising.

    That would be more like real world vision as current displays have a fixed depth.

    • Bryan Ischo

      I’m not sure how much depth of field matters. The fact is you still won’t have accurate accomodation. I personally believe that inaccurate accomodation is part of what contributes to VR not looking or feeling quite realistic, something does always seem a little off for me. Maybe lack of depth of field blurring is part of it, but because most of that blurring would occur away from the focal point, I don’t think it really matters much.

    • Simon Therrien

      Adding dof in real time with post-processses will help but one the the biggest leap will be to be able to simulate the parralax translation of the focal plane due to the rotation of both eye from their central rotation nodes. This will greatly help with depth perception.. amongs other stuff..

    • Charles

      Depth-of-field shouldn’t be something generated graphically in VR – that’s not how it works. Depth-of-field is an imitation of what happens automatically when your eyes focus on a nearby object, designed for being viewed on a flat screen. With a VR headset, any depth-of-field effect should only occur passively as a neurological result of the user focusing on a nearby object. In real life, focusing on a nearby object obviously does not change the light being projected towards a person’s face.

      • Yes and reality shouldn’t be virtual, but when trying to recreate it, the fact that we have limited processing power means that’s approximations must be made. Right now it’s hard enough to render 1440p@70hz for all but the more powerful PCs. To do real light-field as you’re suggesting means doing that * there number of depth field to be supported which would decrease fps to the eye by an order of magnitude. Even if that wasn’t an issue, there isn’t a technology right now to get all that signal to a display nor is there a display to bring it to your eye at the resolution VR requires (AR can get away with less and present solutions lack FOV in the best cases right now due to this limitation.) By tracking the eyes and approximating this way, Fove had made it possible to go the other way and improve fps to the eye (sure, limited to 70Hz for now), and with substantially higher fidelity thanks to foveated rendering, and an added level of realism seen in no other HMD at this point.

        To those scoffing about the SDK, how many of you have tried to implement a VR SDK in an app, let alone tackle Foveated rendering or even simulated DOF. When you’ve done that and can explain how it should be done, THEN you are in a position to have an opinion and maybe you should be talking to the team to seek employment instead of trolling the internet.

        To those reading comments regarding the “just 70hz” who don’t know better already; ignore them. There is so much more to VR than FPS and each factor is largely content dependant. There is no magic value where immersion breaks or where it’s suddenly overwhelmingly real. The faster; the better, but that comes with very real tradeoffs thanks to persistence, tracking fidelity, and visual fidelity. Look me up on LinkedIn (or come find me at OC3) if you want to know why I’m qualified to say, but 70Hz is plenty fast, ESPECIALLY given the added benefits of the eye tracking, resolution and optics being used.

        Keep in mind also that short of real lightfields which are >5 years away (I’ve seen all the present candidates), this is the best solution available and it’s not being seriously tackled by anyone else in VR unless it’s in extreme secrecy, and I say this having been to the Oculus HQ and knowing a few people people there worth knowing, working in the bleeding edge of VR with Samsung, and having met with the Morpheus team at SCEA early last year (to give you an idea of the involvement I have in hardware). That said, I was working next to these guys very early on before their Kickstarter and know first-hand a fraction of how dedicated they are and it infuriates me (though it shouldn’t since it’s so typical) that so many people would try to demotivate them for their 5 seconds of trying to look important on the internet. Hopefully their comments will be read by few that actually take them seriously and don’t know better, but I post this for the benefit of those few and in case their team is taking any of those comments even a little too heart.

        Even with (let me count….) 5 high-quality HMDs sitting here, I’m very much looking forward to the Fove 0 and I’m confident they will continue to make progress on the implementation to make the hardware more useful with other platforms and the SDK easier and easier to implement. Even the prototype i tried last year which was glued together; was very enjoyable and they’ve come a long way since.

      • Changing focus doesn’t change the light available to the eye, but it does change the image the brain registers from that light. Non-lightfield HMDs (everything we have access to today) present all light from a single plane and our eyes focus on that plane. That means there is a VERY shallow depth of field (in photography terms) and everything in it is focused which is not realistic. Our eyes have a very shallow depth of field thanks to a HUGE aperture; more-so in low-light (which is why our eyes do so much better than cameras). Combine that with parallax effects; everything in front of and behind the object in focus is doubled in addition to being blurred; and what you get is VERY different in real life (or light-fields) than in present HMDs. The same way monoscopic displays are “flat”, 2D displays (not having a light-field is what I mean, not to be confused with stereoscopy) are also flat, even when stereoscopy is present.

        There’s an argument to be made about a monoscopic display with proper DOF simulation or light-fields (which actually provide stereoscopy naturally anyway) over a 2D display with stereoscopy. That’s the thing about Fove; they’re going to provide both to a very good degree and improvements from there will be marginal until light-field displays are available and those will come in low resolutions initially just as we saw with the consumer Lytro camera.

        • Charles

          I wasn’t aware of this “light field” stuff, but I just read up on it. Apparently the eyes get some extra depth cues from “retinal blur” and “accommodation”. But I’m still having trouble understanding exactly what these are – I can’t find a good explanation. But I thought of one possible explanation (which I didn’t see written anywhere): when the eyes rotate, the position of the pupil changes slightly, and so the position from which they are receiving light changes slightly, which would noticeably change the angle of a very-close object and would thereby be a different image than if the person were focusing at infinity.

          So it looks like you were right that focus-related depth is not completely accurate with current screens, but this still isn’t the same thing as DOF, because DOF means the blurring of background objects when one focuses on near objects. DOF blurring does happen with current screens, but there are inaccuracies with depth cues that can cause eye strain and slight misjudgement of distance.

          Here’s my previous reply again, because it got detected as spam and deleted for some reason.
          ————————————-
          In regards to depth-of-field, I don’t understand what you mean. All VR headsets have depth-of-field, right now. All 3D displays have depth-of-field, right now, and have always had it. Because when it comes to stereo images, depth-of-field is not a technological feature – it’s a biological reaction to the images being presented to the eye and the viewer’s chosen object-of-focus. Prove this to yourself by putting on your HMD and bringing an object close to your face, then moving your eyes to look past it at something else, then moving your eyes back again. The eyes automatically rotate inwards to focus on a close object, which puts everything else out of focus. In 2D, it’s a completely different story, because the two eyes are seeing the same image, so you can’t change focus. The 2D depth-of-field effect simulates the biological reaction to focusing in 3D on a nearby object by intentionally putting background objects out of focus. This is only necessary for 2D, because in 3D it just happens.

          In regards to only 70Hz, I agree it’s probably fine. Having owned the 75 Hz Oculus DK2 and the 60 Hz Gear VR, I felt like 75 Hz was fine, while 60 Hz was not. 70 Hz is pretty close to 75 Hz.

          Definitely looking forward to the standardization of the foveated rendering being implemented in this headset.
          ————————————-

  • Ashley Pinnick

    Fove had a small booth at E3 2015 and it was really interesting to give a try. I’m glad to see they’ve got a release date on the horizon! The game I played was a shooter where I used my eyes to focus on targets and shoot lasers from them. Lots of fun, a little eye strain though.

  • Display Refresh Rate: 70Hz ???

    • VR Development is still in a learning phase. When you think of what the human eye can see in terms of FPS then it isn’t as straight cut as higher Hz = win

      Here is an interesting article on eye FPS http://www.100fps.com/how_many_frames_can_humans_see.htm

      As you can hopefully see from that article, real-time rendering techniques need to evolve differently for VR so that FPS/Hz are not the be-all answer. A faster display rate is only part of the solution, Oculus and Valve said the sweet spot is 90 Hz but that was with current graphic frameworks. Sony have gone with 120Hz.

      So, should 70Hz work? That depends on advancements in real-time VR rendering. theoretically it could be ok, they seem to think so but more research needs to be taken into how we see and how you display your content to the viewer. Just like 24fps film and how TV devices display it smoothly using frame blending and other techniques, this needs to be researched for VR too. Even 300 Hz is not the answer if your content is just delivered as is to the viewer, you can still make them puke due to a lack of visual understanding.

      There is a lot of research happening, foveated rendering at 70Hz may be the answer, I would love to test it out. Defocusing, DoF of the peripheral view may also make the Hz issue less of a factor. All these new methods of rendering are appearing as we try to simulate what the human eye can see.

      Just my 2p
      Geoff

      • Erik Malkavian

        @D3Pixel:disqus Your crazy when you say Oculus and Vive have found a sweet spot with 90 fps. Frame Rate is KING! Period. Watching anything on a screen that is inch from your face requires VERY high frame rates. 120 mhz is really the bare minimum and one of the reasons I am waiting until 240 fps is available because I refuse to subject my eyes to that kind of pain.

        • Erik, I see you own the GearVR and have tried 60fps. Thats bad, I have tried it. FPS it is not linear in terms of visual smoothness, it is exponential with diminishing returns. Freesync and G-Sync which do not increase native Hz can make things appear very smooth too, It is not all about fps which is a brute force solution.

  • MosBen

    Now that there are actual consumer HMDs on the market, I’m much less interested in these startups. How much room is there in the market place for HMDs? It seems like all of these indie VR hardware startups are just going to be bought up by the bigger players.

    • Ian Shook

      I was just going to post that foveated rendering and eye tracking is more of a feature that all headsets will eventually have. Seems they should’ve licensed the tech instead of making another HMD. It’s like a head-strap company making an HMD to show off their head-strap.

      • Jerald Doerr

        Well said…. But dude…. As long as the headstrap works I’ll use

    • Jerald Doerr

      Well said…. But dude…. As long as tha headstrap works I’ll use it!

  • yag

    How can they track saccadic eye movements (to do foveated rendering) with only 120hz eye trackers ? I was thinking the 250hz trackers from SMI were already too slow…

    • Jerald Doerr

      Here’s the thing… With a PC VR setup most games can be played with any headset…. So hardware/software is about the only deference… And that’s good becouse the more people that designed different VR head sets the better designed VR headsets will fallow .

    • Scott C

      The assumption I’ve been making regarding the foveated rendering research is that the detail-rendered area is enough bigger than the actual retinal target that the small saccades aren’t crucial to catch individually. So long as you get the more gross eye movements tracked, the saccades will fall within the detail-rendered area.

      • yag

        Makes sense. So faster eye-trackers will bring more and more performance gains.

  • Jason Héll

    This is the only significant high-end HMD that’s on the near horizon, and not in some Never-never Land, like StarVR, and all the rest.

    The resolution swings it for me, not eye-tracking. As I have DK2 I never considered purchasing CV1 because of the disappointing resolution, the necessity of splashing the cash on a new computer to support it, and the awful ‘godrays’ (and all the other things wrong with it). I’ll certainly get Fove – it will be the perfect stop-gap until the release of CV2, which I’ll certainly get as well.

    I find myself watching 2D/3D video in VR more than playing games, so I won’t be too concerned if an avalanche of games doesn’t come my way.

    • Jerald Doerr

      “Not eye-tracking. ” That’s only because you don’t understand what the benefit is for developers and and the end user… Not just that, it is like adding nitrous to your car, it will add an extreme performance increase in rendering if used right….

      • Jason Héll

        Don’t assume superiority when ignorant of the facts. I understand the benefits of foveated rendering and what it can do for dogs, pigs, and wild fowl perfectly well. It’s simply not my number one priority.

  • PrymeFactor

    No compatibility with Glasses wearers?

    LOL

    • Scott C

      Yeah, this is a serious problem. If it’s only a problem with some frames, that might be one thing. But there’s a demographic out there that doesn’t wear contacts for some pretty compelling (comfort, prescription, etc.) reasons that will seriously damage the marketability of a consumer product if your solution for glasses-wearers is to switch to contacts.

      The best case scenario is that this is a limitation they’re working to alleviate by continuing to experiment with sensor placement as they iterate. If this is an inability of the sensors themselves to detect through common lens types…

    • Nigerian Wizard

      What if you have a lazy eye?

  • Pobrecito hablador

    If foveated rendering is only made available to developers through a proprietary
    SDK, they will fail. Developers would support the headset, but it’s unlikely
    that they will spend time developing for specific features of the hmd. A
    chicken egg problem for them to solve, since its most appealing feature won’t
    be supported by almost no one because the low market share they are expected to get. I know that it’s a tough decision as they don’t want to “commoditize”, but the must acknowledge that they don’t have the power of Oculus. You can’t get away with a walled garden if you’re not one of the top dogs in such a reduced market that high-end VR is now. Maybe offering support through OpenVR could help.

    • Scott C

      This highlights the problem with OpenVR, though — it’s only “open” in the sense that Valve lets you use it. Valve has no motivation to develop support for things it’s not already got a vested interest in. There’s no way for the community or industry to submit new features for inclusion. Valve is the gatekeeper, and gets to pick winners and losers by virtue of selecting what to incorporate, and when.

  • Nigerian Wizard

    I really feel FOVE will have a hard time penetrating the market. Maybe if they pulled the stick out of their ass and made it so it wasn’t proprietary, they might have games in the future besides some literally-who startup japanese engine.

  • Erik Malkavian

    Only 70 frame rate is a real deal breaker. I own the GearVR and 60 frame rate from my Note 5 is not nearly enough, in fact the same for 90 fps. The minimum for a screen that is inch from your face is 120 and preferably 240. I was supported of Fove but I kept seeing those frame rates drop in spec updates and knew it would be SERIOUS trouble.