Qualcomm kick-starts standalone VR with its VR820 headset

While Oculus' Rift and HTC's Vive have dominated the conversation about VR in the enthusiast space, there's a burgeoning range of companies thinking about ways to deliver VR experiences that don't require a tether to a powerful gaming PC. Samsung's Gear VR is just one example, but Intel's Project Alloy and the AMD-powered Sulon Q also promise a cordless, entry-level VR experience. Qualcomm is getting in on this game today with an ambitious VR reference platform of its own called the VR820. Produced in partnership with Chinese electronics developer Goertek, the VR820 will serve as both a standalone product and as a reference platform for other companies to develop VR headsets of their own.

As standalone VR platforms go, the VR820 appears to pull out all the stops. The headset uses a Snapdragon 820 SoC to power a 2880×1440 AMOLED display refreshing at 70Hz. That figure translates to a 1440×1440 pixel array per eye. The VR820 includes a pair of eye-tracking cameras that allow the headset to perform foveated rendering, a technique that makes more efficient use of graphics-processing resources for VR. It also has a pair of front-facing cameras that Qualcomm suggests can be used for hand-tracking or augmented-reality applications. The VR820 also purports to perform "six degrees of freedom" head-tracking without any external trackers, unlike the Rift and Vive.

The VR820 is doubtless Android-powered, since Qualcomm touts its own Snapdragon VR SDK as the development toolkit for the platform. Qualcomm expects to start delivering VR820s in the fourth quarter of this year, and it anticipates that partner devices will become available "shortly thereafter." While there's no official word on pricing yet, Qualcomm told The Verge that the VR820 could have a sticker "similar to higher-performance tablets."

Comments closed
    • moose17145
    • 3 years ago

    I may be alone in this… but a couple of things are going to have to happen first for me to adopt VR.

    1. Laser Eye surgery (seriously… the idea of wearing glasses with these things on top of them is enough to make me just go “Nooooooope!”
    2. Wider game adoption than what we are currently seeing
    3. Price needs to come down a little along with the tech improving a little bit further (IE I would hold out for at least the 2nd Generation of the Vive).

    • Chrispy_
    • 3 years ago

    So very, very Asian.

    The “reference platform” has a stupid dragon decal on the front.

    Whilst it may appeal to a few non-Asians, I’d assume the vast majority of people think it looks childish, stupid, or chintzy.

    • ronch
    • 3 years ago

    Really don’t get why VR is so exciting. So I strap two miniature computer monitors to my face. And then?

      • GrimDanfango
      • 3 years ago

      Well, in the case of proper, non-mobile VR – and then the accurate positional tracking system calculates the exact position of your head, 90 times per second, and displays the optically correct viewpoint in whatever game for each of your eyes. It doesn’t look like two screens hanging in front of your eyes, it looks like you are inside a game – with the Vive, you can literally walk around something that doesn’t really exist and look at it from any angle. You can also reach out and pick up imaginary objects using the tracked controllers.
      I find that quite an exciting notion 🙂

    • ronch
    • 3 years ago

    Different companies seem to claim different resolutions as adequate for VR. As someone with low vision these companies don’t need to work hard to push a quadrillion pixels for each of my eyes. Heck, 320 x 240 should be fine. Heh.

    Edit – DAMMIT autocorrect!!!
    Edit 2 – DAMMIT brain!!

    • Laykun
    • 3 years ago

    This looked pretty “me too”/pedestrian until I read this

    [quote<]eye-tracking cameras that allow the headset to perform foveated rendering[/quote<] Now that is legitimately interesting, and if done well could mean this will be one little beast of a headset. EDIT: Too bad it's so damn ugly.

      • GrimDanfango
      • 3 years ago

      Huh, hadn’t actually noticed the bit about 6dof tracking. That could actually make for a reasonable VR solution.

      I’m not sure how well that’ll work if they’re going to fall back on accelerometers instead of a precise external tracking solution.

      Unless they’re managing to use the dual front-facing cameras to track the environment… I rather doubt that though – real-time image-based matchmoving is surely a way off yet… even for a powerful desktop workstation.

      Maybe you can place tracking markers around your room… I guess that could work.

        • Laykun
        • 3 years ago

        I used to work with some AR researchers before they shipped off to work with Qualcomm, I have no doubt they have expertise to pull off solid tracking with the external trackers.

    • Srsly_Bro
    • 3 years ago

    How much resolution per eye is “enough?”

      • Jeff Kampman
      • 3 years ago

      AMD has said that photorealistic VR will demand 4K per eye at a minimum and 16K per eye for “true” immersion.

        • GrimDanfango
        • 3 years ago

        Isn’t this one of those pointlessly subjective notions to attempt to quantify? I mean, about the only answer that really means anything is – whatever resolution corresponds approximately to the same dot-pitch as the size of the individual rods-and-cones in your retina.

        Short of that… “enough” is surely just a matter of taste.

        4K would maybe be “enough” to reduce pixel aliasing to bearable levels.

        “Photorealism” could be achieved at a lower resolution than even current headsets provide – being entirely an issue of complex physically-based-lighting, realistic ambient occlusion, radiance bounces, etc… you could surely have something low-resolution that was also so convincingly rendered at to be indistinguishable from video footage?

        In fact, resolution in computing terms works directly against photorealism… the higher the definition, the more computing power and artist time it takes to fill it.

        4K would be nice… but would have drawbacks…
        16K would be even nicer… but would have massive drawbacks…

        (higher resolutions will also mean that the goalposts for untethered, wireless VR will keep moving further away – 2x4K at 90+hz will be extremely tough to send wirelessly without compression – compression causing unacceptable latency in VR)

          • Joerdgs
          • 3 years ago

          The 16K per eye number is the resolution required (with a FOV of today’s HMDs) to simulate the amount of detail that any normal person with 20/20 vision could see in the focus of their vision. The thing is, the center of your vision is just a tiny area so rendering the full screen at that resolution would just be a waste of resources. What you’d want is eye tracking tech that tells you where exactly the player is looking on-screen, and only render at full resolution in that area, while skimping the rest.

            • GrimDanfango
            • 3 years ago

            But again, where is that 16k-20/20 vision statistic actually coming from? Has someone precisely measured the effective limit of resolution of a 20/20 human eye? It sounds somewhat more like one of the usual “## fps because the human eye can only perceive that many frames per second” idea…

            I strongly suspect that there are details that typical 20/20 vision could detect finer than the dot-pitch of a 16k display covering a typical VR FOV… especially considering that due to the optics used in VR, the resolution is effectly *less* in the dead-center than at the edges, where the pixels become compressed. If that’s the case, then the 16k figure is just another subjective figure plucked out of the air.

            Again, not denying that it’d be great to have, or that it’s a likely enough figure beyond which diminishing returns would render larger resolutions pointless… it just niggles me, when the most they’ve done is approximated a ballpark “probably-good-enough” value, that they state it in a way that sounds more authoritative. It’s akin to that “experts claim…” nonsense from newspaper headlines. Nothing wrong with a ballpark figure, I just wish they wouldn’t present it as some authoritative fact when it’s an inherently subjective matter.

            I agree that foveated rendering can’t come soon enough… it will enable massively higher VR performance and thus open the doors to at least decently performing 4k screens.

    • Shinare
    • 3 years ago

    My Daughter brought with her the last time she visited her Samsung Gear VR + Note 5. It was my first experience ever with VR and I must say that I was both impressed and disappointed at the EXACT same time. Sounds hard to do, but I really was. But the lasting impression I got was this: VR is going to be great, really great. Someday.

    It just needed to be way better resolution. Not sure how to quantify it, but “way better” sounds about right. Every moment was like waiting for a video stream to catch up so it would pop into HD video quality rather than the crappy internet connection resolution. But that moment never came.

    I had the chance to watch some netflix (like sitting in a movie theater), play some VR games she had downloaded, and even watched some of the 2016 Rio Olympics that were recorded for the VR. I was really excited for it all, except that it just left me wanting (begging) for something better. 5 years?

    Edit: And one other thing: I’m an older gentleman, and the strain on my neck was palpable after playing the games. I even found myself supporting the headset with my hand when watching netflix.

      • GrimDanfango
      • 3 years ago

      I think this is indicative of a big problem with all these cheapy mobile solutions – a LOT of people’s first (and probably last) experience with VR is going to be soured by thinking that this is what VR actually is.

      Gear VR and the various others coming to market aren’t VR… they’re toy-VR. To me, the most fundamental element of a truely immersive VR experience is precise positional head tracking, and these devices omit it entirely.
      Without it, manufacturers are reducing VR to be just another meaningless gimmick like 3D television.

      If you’re really interested in VR… wait until you can try out someone’s Rift or Vive. Anything less will only disappoint.

      Also, drop the expectation that VR gives any real benefit to video content. I know anyone and everyone are going to be pushing 360-degree video as the next big thing, but until they develop a working light-field video camera that actually lets you move your head around inside a video stream, it’s still just going to be another gimmick… and that light-field tech certainly is years off yet… if it’s even possible.

      The only really distinct, non-gimmicky application of VR for now is for PC(/console when it arrives) gaming, with an accurately tracked headset. Anything else is just muddying the water and giving people a bad first impression.

      That said, even the Rift and Vive are certainly 1st-gen products, and have lacklustre resolution… but play the right game on it, and it’s very easy to lose yourself in it regardless.

      (additionally – the neck strain thing may very well be less with the Rift/Vive… the Gear VR literally attaches a whole phone to the point furthest from your head, so I can’t imagine the balance is at all comfortable. I hear the Rift is especially comfortable in that regard.)

      • psuedonymous
      • 3 years ago

      In terms of resolution, you’re limited by the smartphone inside GearVR: in terms of rendered images, the framebuffer size is somewhat below (and often a lot below) the effective display resolution. In terms of prerecorded 360 video, the maximum resolution that can be played back (limited by the FFC in the SoC) is FAR below the display resolution. About the only way to view the display at it’s maximum possible resolution is when viewing static 360 images. And even then, the phone has a limited capability to perform antialiasing, and antialiasing could almost be considered a mandatory requirement for VR on desktop.

      Also hobbling things is the panel itself: panels used in the Rift CV1 and Vive are designed to have a high ‘fill factor’, which reduces the ‘screen door effect’ by trying as much as possible to eliminate the dead space between individual subpixels (the Rift CV1 has an additional diffractive layer in it’s composit lens which further reduces this). The phones used in GearVR lack this optimised panel (which is instead optimised for battery life and brightness) so even at an equivalent angular resolution, GearVR will have a [i<]perceptual[/i<] lower resolution due to the pixel structure being much more visible. GearVR is the best mobile HMD available, but it;s only competition is Google Cardboard. Along with the lack of positional tracking (one of the cornerstones of VR) GearVR doesn't hold a candle to proper desktop VR devices.

    • CuttinHobo
    • 3 years ago

    Sir, the possibility of successfully navigating the stand-alone VR market is approximately 3,720 to 1.

    • wingless
    • 3 years ago

    I wouldnt’ mind a headset with a small backpack or hip mounted pack for the compute, graphics, and battery.

    We can have more power and heat dissipation without having a ton of weight on our face!

      • drfish
      • 3 years ago

      [url=https://techreport.com/forums/viewtopic.php?t=81535<]Ahead of it's time.[/url<] RIP. Seriously though, totally agree, one little cable running down from the glasses/headset and the form-factor that people are likely to prefer gets a lot easier.

Pin It on Pinterest

Share This