Sony's display renders two views for one tracked viewer

↗

What the developer documentation specifies

Sony's current developer portal for the Spatial Reality Display, now hosted at xyn.sony.net after the manufacturer's own site restructured its addresses, states that "you can develop applications using various game engines and graphics libraries," naming Unity, Unreal Engine, OpenXR, OpenGL and both DirectX 11 and DirectX 12. The documentation also records a hardware-generation compatibility rule: content built with an older version of the Unreal Engine plugin, in Sony's words, "can only play on ELF-SR1," the first-generation display, and must be rebuilt with the current plugin to run on the newer ELF-SR2. Beyond a standard controller, the documentation notes that a "content creator can incorporate gesture control into their design such as with a gaming controller, or Leap Motion Controller for hand gestures," meaning interaction design is left to the developer rather than fixed by the hardware.

The mechanism: one face, tracked, driving a stereo pair

The technical documentation hub links to a white paper describing the display as "glasses-free 3D technology that revolutionizes design, visualization, and gaming with real-time head motion tracking." Stripped of the promotional framing, that sentence describes a well-understood mechanism: a camera on the display tracks one viewer's eyes or face, and the render engine generates a stereo image pair, continuously repositioned so each eye receives the correct perspective for wherever that viewer's head currently is. This produces stereoscopic depth and head-coupled motion parallax for that one tracked person, without glasses, a fundamentally different mechanism from a lenticular panel showing dozens of fixed angular views to many people at once: here there is exactly one correct view, continuously recalculated, belonging to whichever face the camera is tracking.

What the documentation does not settle

The retrieved developer pages describe engines, plugin compatibility and interaction options in specific, checkable terms, but the pages found do not state the tracking camera's angular range, latency, or how many faces it can register before choosing one to follow, details a venue would need before relying on the display for a queue of visitors rather than a single seated user. The single-viewer mechanism also implies a plain, undocumented-but-inevitable constraint: anyone standing beside the tracked viewer is looking at the same stereo pair calculated for someone else's eyes, and will see a distorted or doubled image rather than a second correct view. Nothing in Sony's own materials claims otherwise; the display is consistently described, in the sources retrieved, as a single-user tool for design and visualisation work, not a shared-audience display.

Questions to bring to a demo

• Is the audience meant to be one tracked viewer at a time, and what happens to the image for anyone else nearby?
• Which hardware generation, ELF-SR1 or ELF-SR2, is running, and was the content built against a matching plugin version?
• What interaction method is wired up, controller, hand tracking, or something custom, and is that documented as supported?

Sony's documentation is precise about the software path into the display and candid, if indirectly, about its single-viewer design; the eye-tracking hardware's own operating limits are the part a prospective venue would still need to test rather than read.