Metaverse Research InstituteMetaverse Research Institute

Take 500 Photos, and Sony Turns the Cup on Your Desk into a Photorealistic 3D Asset

Can a photo of a glass cup preserve the transparency of glass? Can a stair handrail keep its slender metallic sheen in a 3D world?

The industry has struggled with these questions for years.

In October 2026, Sony offered its answer: the spatial content creation suite "XYN" has been updated in Japan and the US, with a core algorithm that generates photorealistic 3D assets directly from photos.

What It Is

Grab a mirrorless camera or an iPhone, circle an object and shoot — up to 500 photos. Drop them into the cloud app "XYN Spatial Scan," and Sony's proprietary "XYN Venus algorithm" hands you back a photorealistic 3D model.

Here's the key point: this algorithm is neither photogrammetry nor 3D Gaussian Splatting.

Those two mainstream 3D reconstruction techniques are familiar to insiders, but they share a common weakness — the transparency of glass, the sheen of metal, and slender structures like railings tend to come out murky or simply break apart.

The Venus algorithm attacks precisely this hardest problem. Reflections, gloss, and the textures of complex materials are all faithfully reproduced.

Why It Matters

Let me ask: spatial computing, the metaverse, virtual production — how many years have these concepts been around, and why is real-world adoption still slow?

One bottleneck is content. 3D asset creation is notoriously time-consuming and expensive; a single detailed model can cost a modeler several all-nighters. Without a massive supply of cheap, realistic 3D assets, even the best hardware is just an empty room.

Sony's logic is easy to read: the physical world is the largest asset library. Snap a few photos, and objects move from the physical world into the digital one in bulk. It's building a content highway for the spatial internet.

Pricing has loosened this time, too. With online payments now supported, individual professional creators can subscribe — no longer enterprise-only. Beyond the free tier, the cheapest plan costs 3,300 yen per month, roughly 150 RMB. Sony is lowering the barrier and waiting for a breakout moment.

What It's Like to Use

The workflow is quite smooth:

  • Shoot: The mobile app "XYN Spatial Scan Navi" adds an Object Assist mode — frame the scan range by the subject's size, then follow the constant-distance guide around the object
  • Select: Use a bounding box to specify the generation range and extract only the object you want
  • Compute: Everything runs in the cloud, independent of your local machine, with browser-based preview
  • Export: Download PLY-format models in four quality tiers — Basic, Enhanced, Advanced, Extreme

The "XYN Spatial Renderer Plugin" for virtual production was also extended: alongside Unreal Engine and Disguise RenderStream, it now supports Aximmetry, plus a popping-reduction feature that suppresses flickering during camera movement.

The Detective's Take

Photo-to-3D isn't new technology. What's formidable about Sony lies elsewhere: it has cameras, sensors, rendering engines, and the entire film-industry ecosystem. XYN essentially puts the whole "physical world to digital world" pipeline in one company's hands.

Once the cost of scanning an object approaches zero, digital twins, virtual showrooms, and 3D e-commerce displays will go from luxury exclusives to commodity standards.

Someday, that limited-edition espresso machine in your shop could be scanned once and appear simultaneously on an e-commerce page, in a virtual showroom, and inside a customer's AR glasses. That's when the boundary between physical and digital truly starts to dissolve.

One last question: when everything in the real world can be copied into virtual space with one click, will your most valuable assets still be the physical ones?