Distinguish looking around from moving around.
Three degrees of freedom, often abbreviated 3DoF, describe rotational movement: turning left or right, looking up or down and tilting the head. Six degrees of freedom, or 6DoF, add positional movement along three axes. This distinction concerns the viewer’s supported relationship to the scene, not simply the label on a headset.
A position-tracked headset can play a conventional 360 video, but that does not make the recording position-tracked content. Leaning sideways will not reveal the previously hidden side of a chair unless the media representation contains information that supports that change of viewpoint. Stereoscopic depth is also different from free positional movement: two eye views do not describe every possible viewing position.
Set the audience’s expectations accurately. A spherical documentary may offer a strong sense of place from one location. A rendered simulation may allow exploration and interaction. Both can be effective, but promising one while delivering the other creates confusion and sometimes discomfort.
Choose a delivery route your audience can reach.
A browser-based spherical player allows exploration with a mouse, touch or device orientation when available and permitted. It can reach people without a headset, although small screens and input friction affect attention. A separate flat edit may still be the clearest route for viewers who cannot use the spherical controls or who need a concise version.
Headsets such as Meta Quest, Apple Vision Pro and HTC VIVE represent different hardware and software ecosystems. Do not assume identical codec support, file limits, browser features, controller behaviour or accessibility settings. Check current official documentation and test the application and device combination you intend to support. A platform name is not a complete delivery specification.
The Khronos Group’s OpenXR standard addresses an application interface for access to XR devices and runtimes. W3C WebXR specifies browser interfaces for immersive experiences on supported systems. Neither is a video codec, a guarantee of universal compatibility or a replacement for testing. Related formats such as glTF and specifications such as OpenUSD address other parts of the asset and scene pipeline.
Understand what emerging representations add.
Volumetric capture records a subject or scene in a form that can support viewing from multiple positions within limits. Depending on the system, the result may involve geometry, textures, point clouds or other representations. Capturing fine hair, transparent materials and fast movement remains difficult, and a studio rig may demand substantial lighting, calibration and data management.
Photogrammetry estimates geometry from overlapping photographs. It is established for many static subjects, but reflective surfaces and moving objects can cause problems. LiDAR measures distance using light and can complement imagery, although sensor characteristics and processing choices affect the result. Neither technique removes the need to check scale, missing surfaces and alignment.
Neural radiance fields, commonly called NeRFs, and 3D Gaussian splatting are families of scene-representation techniques associated with novel-view synthesis. They can produce compelling reconstructions from suitable captures. Their behaviour depends on training or optimisation, scene coverage, renderer support and the distance from captured viewpoints. A convincing demonstration does not establish robust editing, interaction, distribution or accessibility for a finished production.
Distinguish reconstruction from invention. Missing areas, view-dependent effects and generated detail may not correspond to recorded evidence. For journalism, heritage documentation or scientific communication, describe how the scene was produced and where uncertainty remains. Visual plausibility is not the same as measurement accuracy.
Define a release that can be tested.
Write a short delivery contract: supported devices, expected posture, input method, media format, loading conditions and accessible alternatives. Test on the least capable device you claim to support, not only the development workstation. High pixel counts and complex scenes can exceed decoding, rendering or memory limits even when file transfer is fast enough.
Check the complete experience: permission prompts, loading feedback, captions, volume, pause, recentering and exit. In immersive rendering, inconsistent frame delivery and high latency can undermine comfort. In video playback, buffering or dropped frames can interrupt orientation. Measure actual behaviour instead of inferring readiness from a specification sheet.
Protect privacy when adding spatial interaction. Gaze, hand, room and movement data can be sensitive. Collect only what is necessary, explain the purpose and avoid assuming that device permission is a substitute for informed consent. This informational website does not request those sensor inputs or run an immersive viewer.
Finally, provide an ordinary route to the essential information. A transcript, image sequence or flat video can make the project useful beyond one generation of hardware. The future of spatial media will include new capabilities; good delivery still begins with a clear promise that the current experience can keep.