spatialcomputing.info

Varjo, makers of high-end XR headsets, today announced it is introducing an XR cloud streaming service for enterprise customers. The streaming solution leverages the headset’s foveated rendering to deliver high quality XR experiences on less powerful machines. The company plans to eventually allow other headsets to make use of the service as well.

Varjo today revealed that it has been developing an XR cloud streaming service for its enterprise customers. The feature comes as an expansion to the recently introduced Varjo Reality Cloud, and the company says its goal is the make it easier for companies to scale the use of XR internally.

Varjo’s high-end headsets offer some of the highest resolution and sharpest visuals available in any headset on the market today. Rendering virtual content for those headsets understandably requires some pretty hefty hardware… a low-powered laptop simply won’t cut it. The need for computers with beefy GPUs to power the company’s headsets has constrained the ability of customers to scale the use of Varjo headsets more widely.

Varjo sees cloud rendering as a solution to this issue and has set out to build its own XR cloud streaming service. The idea is that the rendering is done on powerful servers in the cloud and then streamed to the headset. This approach can significantly reduce the power required of the local computer that the headset is plugged into.

Of course, Varjo is far from the first company to have experimented with this. And their situation is particularly challenging because of the high resolution of their headsets compared to many others on the market.

And yet the company seems to be rising to the challenge with what it says is its own solution built atop AWS’s low latency ‘Wavelength’ platform. I got an exclusive early preview of the feature running on the Varjo XR-3 headset.

When I first put the headset on I saw a high fidelity virtual car rendered against the real backdrop of the room as seen through the headset’s cameras—a demo I’ve seen in some form or another plenty of times in Varjo’s headsets. But this time the headset was plugged into a fairly thin laptop instead of a desktop-class rig. The car in front of me, explained Varjo’s Chief Technology Officer, Urho Konttori, was being rendered entirely in the cloud on an AWS server in Oregon… a few hundred miles North from where I was standing in Silicon Valley.

The scene in front of me appeared entirely sharp on the headset’s ‘retina resolution’ display, and aside from the occasional hiccup that showed a bit of artifacting around the edges of the car, it ran just as smoothly as I’d expect from a locally rendered XR experience.

Impressively, this wasn’t a highly controlled demo on some kind of dedicated, staged connection. According to Konttori, the laptop running the stream was plugged into the general internet connection that happened to be available at the space the company had reserved for its demo. The upstream and downstream connection was running over the regular old internet to the server in Oregon and back, which Konttori said demonstrates  the system is robust enough for real-world deployments.

That’s partly thanks to the company’s “foveated transport algorithm,” which it says makes use of the headset’s eye-tracking to help compress the downstream bandwidth requirement to just 35 megabits per second.

Konttori told me that Varjo is going the extra mile by focusing on the feature’s usability. Once a company has configured their software to run in the cloud, users can start a cloud session with a single click right from the Varjo software. Further, the company says it plans to support link-based invitations to cloud sessions, so that anyone with a headset and the Varjo software installed can instantly view the right content without installing any additional software.

Varjo sees this ease-of-use as being particularly important for making it easier for customers to scale the use of XR headsets internally. While an auto manufacturer’s design team, for instance, might work with headsets regularly and have the high-powered machines to run them, other stakeholders are less likely to have an XR capable machine with all the requisite software ready to go. With cloud streamin, those stakeholders could more easily join design review sessions to sign off on the latest changes.

Konttori said that electric car maker Rivian is doing precisely this using an early version of Varjo’s cloud streaming feature integrated with the industry standard Autodesk VRED software.

“With Varjo Reality Cloud, we are able to make high-fidelity immersion a key part of our design development and scale it effectively across locations,” said Trevor Greene, Lead of Visualization Design at Rivian. “This is a turn-key solution that allows users with very different skill levels to be brought into an immersive environment to collaborate—something that hasn’t been possible before.”

But even if stakeholders could jump into an immersive design review with the click of a link… how many of them are likely to have a dedicated XR space—with a headset, tracking beacons, and all? To make XR truly scalable within organizations, you’d really want to support a wide range of headsets from Varjo’s ultra high-end all the way down to standalone devices.

Indeed, Varjo says it plans to do just this. Though the XR cloud streaming feature will only be available on its own headsets at launch, Konttori told me in the future the company plans to open the system up to other headsets, whether they’re PC or Android based. This could of course mean the possibility of using more affordable standalone headsets like Vive Focus 3 or Quest 2 as end-points for high-quality cloud rendered visuals, making it that much easier to scale XR more broadly inside of organizations.

– – — – –

Varjo’s XR cloud streaming feature is expected to be made widely available to enterprise customers in the first half of the year. It will be sold as an additional feature, on top of the existing annual fee for the company’s enterprise headsets, though the exact pricing model has yet to be revealed.