Gracia’s transferring absolutely volumetric captures can now be streamed, together with to blended actuality on Quest 3, with no app or content material obtain required — and an Apple Imaginative and prescient Professional app is coming quickly.
Gracia first launched in 2024 as a Quest app for viewing static photorealistic objects, and, way more notably, a PC VR app for transferring photorealistic scenes, in any other case describable as volumetric video.
And to be clear, this isn’t simply easy parallax such as you’d get from stereoscopic 3D video, nor limited-perspective artificial depth just like the spatial scenes of Apple’s visionOS 26. Gracia goes a lot additional, providing absolutely volumetric content material you can bodily stroll round in VR or in your room through blended actuality.
Gracia Transferring Volumetric Scenes Now On Quest 3 & Pico 4
Gracia’s transferring volumetric scenes, dynamic gaussian splats, at the moment are accessible on Quest 3 and Pico 4 headsets standalone.
As with virtually the entire exceptional developments in 3D reconstruction over the previous few years, the expertise behind Gracia is Gaussian splatting – becoming thousands and thousands of semitransparent coloured blobs (Gaussians) in 3D house in order that arbitrary viewpoints may be rendered realistically in real-time. Transferring splats are sometimes referred to as 4DGS, four-dimensional Gaussian splats, the fourth dimension being time.
By the top of 2024, Gracia had already launched standalone apps for Quest 3 and Pico 4, eradicating the requirement for a PC, with the tradeoff of some high quality. However the largest downside with this standalone stream was the necessity to obtain every scene, requiring you to face there with the headset on watching the multi-gigabyte file arrive in your gadget. With the corporate’s expertise on the time, streaming would have required a 2.4 gigabit web connection, one thing solely a tiny fraction of individuals have and that almost all standalone headsets cannot even reliably hit wirelessly anyway.
This crippled Gracia’s founding ambitions of being the “YouTube of really volumetric content material” (although the corporate has now pivoted from this), as a result of if YouTube required a multi-gigabyte obtain earlier than every video performed, it simply would not be a viable service. And so the standalone headset app targeted on quick 10-second clips, quite than longer content material that might help extra use circumstances.
On the time of the standalone app launch, Gracia informed UploadVR that it was engaged on bettering its compression by an order of magnitude. Now, a 12 months and a half later, the startup has achieved that lofty objective.
Gracia streaming on Apple Imaginative and prescient Professional.
In March of this 12 months, Gracia launched a world-first: the power to stream choose 4DGS scenes, with no app set up or full scene obtain required. And you do not even must obtain an app; Gracia’s new tech stack works fully in WebXR, which means you’ll be able to simply navigate to gracia.ai/retailer in your Quest 3’s internet browser and have a performer volumetrically rendered in your lounge in seconds.
On Apple Imaginative and prescient Professional, since Apple nonetheless would not let WebXR apps use passthrough as a background, you will as a substitute see the captures in a featureless black void. To get round this and let Imaginative and prescient Professional house owners show the captures of their lounge, Gracia is engaged on an Apple Imaginative and prescient Professional app, presently in testing, which must be launching comparatively quickly.
China’s 4DV AI Releases 6DoF Volumetric Video WebXR Demo
China’s 4DV AI launched a WebXR demo of volumetric 6DoF movies created from 20 digital camera views, and you may attempt it in your VR headset proper now.
There’s, so far as we’re conscious, nothing else like Gracia’s streaming on this planet right now. China’s 4DV confirmed transferring scenes in WebXR final 12 months, however these have been solely 10 seconds lengthy and predownloaded. Gracia is as soon as once more main the pack with regards to rendering transferring volumetric captures in VR.
The startup says it achieved this by growing a way for sending keyframes and movement change deltas, encoding solely what components of the seize change over time quite than sending all the scene for every body. Basically, Gracia is making use of a video codec type method to Gaussian splatting. It additionally makes use of the brand new WebGPU graphics API in your browser to realize the utmost attainable efficiency, rivaling native apps, to get rid of probably the most frictionful step of all for any client app, that preliminary obtain requirement within the first place. The compression makes the stream possible; WebGPU makes rendering the streamed splats quick sufficient within the browser.
In addition to permitting for near-instant streaming of volumetric captures, this implies captures haven’t any exhausting file-size cap, in concept supporting use circumstances like total concert events.
To stream Gracia’s transferring volumetric scenes at comparatively top quality, 120K splats per body, you will want a continuing 75 Mbps connection. A decade in the past that might be a non-starter, however as of 2026 round 90% of houses in developed international locations now have sufficiently quick web for this.
Nonetheless, some do not, and typically the headset will probably be too distant from the router or on a community the place different units are hogging the bandwidth. For that, Gracia has demonstrated a 17 Mbps mode, providing 15K splats per body, although on this mode the standard takes a big hit.
Gracia streaming compression ranges.
Three scenes are presently accessible to stream from Gracia’s web site: a 36-second clip of an individual engaged on a bicycle wheel, a 39-second clip of a physician analyzing a affected person’s shoulder, and most spectacular of all, a 4-minute efficiency by a musician singing with a guitar, full with spatial audio.
Gracia can also be eager to level out that whereas its captures are well-suited to be positioned in your lounge through blended actuality, they will also be positioned into current digital Unity or Unreal environments, together with for VR video games, if a developer needed to do that.
Gracia captures may be positioned into digital scenes.
To be clear, by the best way, by “stream” on this article I am referring to watching content material with out having to totally obtain it. Gracia’s expertise doesn’t help livestreaming. And the seize and processing necessities are the elephant within the room right here.
Capturing these scenes requires round 60 shutter-synced and genlocked cameras, which studios sometimes cost a minimum of $15K for, and Gracia fees $800 per minute for processing plus a “minimal cost per phase”. iPhone 17 Professional’s genlock help could enable for a decrease marginal seize value, although for finest outcomes creators will need to use skilled cameras like these from Blackmagic and Sony.
Nonetheless, the seize and processing necessities apart, what Gracia has pulled off right here is exceptional. I’d a lot quite watch a musician or comic carry out as a 6DoF quantity in my lounge, or most popular VR surroundings, than a 3DoF 180° 3D video, and I think for those who’re studying this, you agree.
When you personal a VR headset, head over to gracia.ai/retailer in your headset’s internet browser, or your desktop browser for PC VR, and take a look at it out for your self. And even for those who do not, you’ll be able to attempt a flatscreen render in your browser to get the fundamental thought, although with out the sense of depth and scale.
The subsequent frontier for the volumetric scenes house will probably be growing the expertise to permit anybody to seize volumetric scenes with a few iPhones and to considerably cut back the processing time. As unattainable as this may sound right now, so did the thought of even streaming the sort of content material just a few years in the past, and the tempo of progress within the 3D reconstruction scene, leveraging the speedy developments in AI, stays staggering. The “YouTube of really volumetric content material” just isn’t attainable right now, however I think it very a lot so will probably be by the top of this decade – and it might grow to be a core use case of XR expertise, propelling its mainstream use by the compounding benefits of a social content material feed platform that actually takes benefit of it.
I am actively writing on UploadVR once more, and this text is one in a sequence of “catch up” items the place I report on among the attention-grabbing issues which have been taking place within the business in current months. And sure, VR Obtain is coming again very quickly!
