The guide to professional VR and AR tools

What’s Facebook’s next VR headset after Quest?

ac1d8 oculusfirstgeneration
Published 2019-11-20

The success of Oculus Quest and the hints of future things to come makes us wonder what the next steps are towards VR’s mainstream adoption. Here’s our analysis and speculative conclusions about a few years ahead.

Rift, Go, Quest – and the winner is…

When Mark Zuckerberg presented the lineup of these three products and said “this is our first generation lineup” it sounded like an implication that next generation would simply mean a Rift 2, Go 2 and Quest 2.

After the release and subsequent success of Quest, that scenario doesn’t seem very likely. Especially so if we take into account what Abrash hinted about next generation – a “quantum leap” rather than just iterative development.

Now that there is a true all-in-one device which, in Facebook’s own words they’re “selling them as fast as we can make them”, in hindsight it looks like pure PCVR and feature-limited cheap mobile VR were just steppingstones. And the Oculus Link was the last piece of the puzzle.

What we’re likely to see from future Oculus releases is probably just Quest all the way. Let’s look at why (in retrospect), and how (as hinted from what we know about Facebook’s r&d efforts) we should see a more focused go-to-market VR strategy from Facebook henceforth.

Carmack believed from the start in full-featured, wireless VR based on mobile architecture… and said “there are problems when a company isn’t laser-focused on one single product” in reference to Oculus efforts on both mobile and high-end VR at the same time.

Rift S – irrelevant?

In our view, Rift S was a reactive product release, made to please the enthusiasts and tech media who were expecting an answer to iterative innovations on the PC-VR side from competitors. It fixed the biggest pain of the original Rift – the messy setup and “variable tracking results” (it was originally released with only one Constellation camera, one more was shipped with Touch almost a year later, and yet another additional camera was recommended for a reliable roomscale setup) by delivering the best inside-out tracking on the market.

At the same time Rift S merely improved the visual specs enough not to have an effect on the strict minimal PC spec requirements – by also lowering the screen refresh frequency from 90 to 80 hz. With the shift from Rift to Rift S, Oculus went from “(comparable to) best in class” to “affordable and accessible PCVR”.

“Rift S by Lenovo”

Most notably, Facebook didn’t even manufacture the Rift S themselves but partnered with Lenovo. Audio and optical comfort took a hit compared to the original Rift – Rift S even lacks a physical adjustment for IPD. Furthermore, it didn’t offer a wireless accessory.

Rift S was a “play not to lose” move, but Oculus isn’t really competing with other VR makers (neither are the others competing amongst each other). Their real competition is “all the reasons why the mainstream isn’t using VR today”.

Oculus Go – who needs 3dof VR?

Oculus Go is really the Gear VR, but in a nicer package. The Gear VR had its eulogy from John Carmack at OC6, but we know that Carmack saw full-featured 6DoF mobile VR as the “end goal”. During the development of Gear VR and Go, the mobile VR team at Oculus managed amazing feats of engineering to cram out impressive processing from limited hardware, but it all really needed Quest for those optimizations to be fully realized as a component in fully interactive virtual worlds.

Go is marketed as “a media consumption device”. That, along with the social utility in Oculus TV, Oculus Venues and ability to run AltSpace, Rumii and other 3rd party social VR apps, seems to be the only argument for its staying power. But it’s clearly not getting any more love from developers or press, and it hasn’t reached the number of consumers that warrants continued interest from media content providers.

And seeing how the actual manufacturing of Go wasn’t done by Oculus, (they used Xiaomi as a partner just like they did with Lenovo for Rift S) it doesn’t look like it’s getting much love from its maker either.

Xiaomi, the manufacturing partner for Oculus Go, also made a version of the product for sale in China, where Facebook is banned. That project doesn’t seem to be doing well.

Our verdict: The use case of media consumption in VR isn’t enough of a value proposition to sell a $200 device). It will be nice as a side feature in future VR and AR headsets, and that’s where the legacy of mobile VR can live on.

The path ahead: Quest as a hybrid platform – standalone + PCVR in one

Here’s the thing: Wired VR will never be a thing.

The average consumer knows this, Facebook knows this, and any Quest user, whether new to VR or seasoned enthusiast, knows it too.

“But what about PC-level graphics/performance??” Yes. We will be able to opt in to that too, and not just by going back to cables with Link.

Link’s Endgame

Link isn’t about the cable. It’s a significant ecosystem evolution for Oculus because it starts the roll-out of distributed/remote processing.

Because Quest only has a USB connection and lacks HDMI, for Link to work with the limited USB bandwidth there’s some engineering magic going on to encode/compress the video signal on the PC side… AND then decode/decompress it on the Quest side. This is big. The logic of Link (along with earlier image optimization innovations like ASW) is a crucial piece of the puzzle of “accessible and powerful mass-consumer VR”.

The Quest needs onboard processing to work as a PCVR headset, unlike all other PCVR headsets like Rift, Vive, Index, Pimax etc which are all pretty much “dumb accessories” that only need to capture/send sensor data and receive an audio/video feed.

This, incidentally or not, is the path to “true” wireless VR. We NEED be able to wear sleek headsets and run high-end VR – from our PCs, through cellular networks (5G/edge) as well as from an external but disconnected mobile processing puck – all without being tethered. And we’re getting there.

The Directional Beam project

For near-field wireless streaming, Facebook has another innovation up their sleeve: Directional Beam. This will be used for primarily Wireless PC VR, BUT it will also allow distributed processing for the Quest as a standalone unit.

Read the patent application for Directional Beam concept here.

VR headsets need to become smaller – more like goggles. Facebook knows this, and Abrash has presented this imaginary future concept:

Apple has a design in the works that’s said to be about a third of the size of Quest. Huawei is about to release a sleek and nice-looking headset, even if its only 3DoF. Skylights Aero is another example. And that can only be done by offloading processing to something like the Magic Leap “puck” – but wirelessly. (Regular consumers have zero tolerance for wires.)

This will be accomplished by Directional Beam… built into yet another innovation that hasn’t been talked about much: A “Portable Compute Case“.

Portable Compute Case: A multi-functional accessory for your wireless VR headset

Patent application for Portable Compute Case

Imagine ripping out the actual processing unit from Quest and putting it in a case that you carry and charge the HMD in (similar to Snap Spectacles or Airpods).

And when you take out the headset, the two units communicate wirelessly – using less power and with high bandwidth thanks to Directional Beam based on positional tracking. The only processing that needs to happen on the headset is the decoding of the compressed data signal, which should be manageable with a tiny, power-efficient chip. (But the bulk of the headset’s volume and weight will still need to consist of battery.)

What about foveated rendering and varifocal displays?

Eyetracking and foveated rendering will be yet another necessary step in reducing the amount of data to be wirelessly transmitted to the headset. Said to potentially save up to 80% of processing, foveated rendering sounds like a no-brainer as a component in the “standalone headset + portable compute case OR PC-connected directional beam unit” pipeline.

The next substantial generation of VR from Oculus will HAVE to include foveated rendering. The question is whether varifocal (for visual comfort and realistic representation of depths of focus – necessary for long-term use of VR) will be consumer-ready at about the same time as foveated rendering. And what, if anything, will Oculus release while on the path there?

Speculation: The new lineup for next generation Oculus VR hardware

The Half-Dome 3 gives us a good hint on the look of the next primary Oculus product.

Our guess is that Oculus will launch a direct follow-up to Quest (“Quest S”?) sometime during 2020, followed by some silence and then have 3 versions of Quest ready by 2022:

$150: “Quest Lite”:

Like Go but with 2 cameras for positional tracking and pass-through (so basically like a Vive Focus). Gameplay can be 6DoF or 3DoF. Controller only 3DoF. Aimed at media consumption, social VR, virtual productivity and light gaming. This is really a hit or miss – will they instead choose to simply abandon the thought of catering to the minimum VR threshold?

Hacky enthusiasts could use it with Virtual Desktop and maybe Vive/Index Controllers along with Lighthouse tracking (just like some people are combining WMR with Vive controllers today).

$300: “Quest S”:

Like Quest but with stronger processor. Can be used with the Link cable for PCVR.

$500: “Quest Half-Dome” including Portable Compute Case (same mobile processor as “Quest S”) :

True hybrid “thin client”-style VR headset with very sleek design and full optical features – 140 degree field of view, eyetracking, foveated rendering and varifocal displays. The ultimate in comfort and mobile performance… and with a Directional Beam PC unit, the ultimate wireless PC headset – and not just for enthusiasts. (Probably still only for Windows. Mac support is unlikely when looking at Apple’s inevitable XR entrance.)


via Gfycat

Bodytracking is a bit of an unknown factor. We believe this will be a key part in making VR extra immersive and adding a new dimension of expressiveness. Our educated guess is that Oculus will NOT go the route of placing trackers on the body but instead simply rely on one external camera or Kinect-like 3D scanner. (Just like they’ve demonstrated as being done in research.)

Occlusion of hands (when the player has their back turned to the camera) would be mitigated by the hand controller tracking, and conversely the bodytracking could provide redundancy tracking data for hands when outside the headset camera field of view.

(Furthermore, if it’s a 3D scanner, it could double as an body/face avatar capture device for use at home, to enable an approximation of the work being done with Facebook Codec Avatars.)

There’s a few possible ways to tie this ability into the user experience described above:

  1. Include the camera in the “Compute Case” / “Directional Beam unit” – simplest to the user, but would add cost to the core product.
  2. Build it into the existing “Portal” product line.
  3. Make it possible to use your own phone via the Oculus companion app (not very elegant, would need to be propped up on something like a tripod accessory.)

Rift S will be rendered obsolete. And as for what Oculus should do about Go, it would probably be wise to “kill their baby” (more like “Carmack’s baby”)!

If not outright killed, maybe they could license the whole product to Xiaomi, to let them focus on the Asian market, where “personal cinema” has more value because of small housing. Again, that’s not a very likely outcome either…

The new Quest product line, along with these components being developed in secret at Facebook Reality Labs, would logically be the more focused play about to be revealed by Facebook in order to win the mass-consumer VR throne.

(End note: There are interesting initiatives from other players that also point to the inevitability of wireless VR / remote processing as a crucial enabler to mass-consumer VR, from players like Nvidia, HTC, Sony, Microsoft, Valve. This time we chose to focus on Facebook/Oculus, partly because they have the clearest explicit aim of making VR mainstream and partly because of the exciting potential of their efforts in the full hardware stack.