Immersive broadcast covers live production in which the audience receives something beyond a single fixed camera feed: a view they can steer, a scene rendered around them, or a real-time graphics layer registered to what the camera is showing. The distinguishing constraint is that all of it must happen live, within the delay budget of a broadcast.
01Approaches
Several distinct techniques sit under the same heading, and they have very different production requirements.
Panoramic and 360-degree capture records the full sphere around a camera position, letting each viewer choose their own view direction. It preserves the live feel with a manageable pipeline, but the viewer stays at the camera's location and cannot move within the scene.
Volumetric capture reconstructs subjects as three-dimensional data using arrays of cameras or depth sensors, so a viewer can move around the subject rather than only look around from one point. It is the most capable approach and by a wide margin the most demanding in capture infrastructure, processing and bandwidth.
Real-time graphics augmentation composites rendered elements into a conventional camera feed, tracked so they hold their position in the scene. This is the most widely deployed family, covering everything from tracked on-field graphics to virtual sets, and it works with an otherwise standard production chain.
Virtual production stages invert the idea by rendering the environment on large LED walls behind the talent, so the scene is captured in camera rather than composited afterwards, with the rendered perspective driven by camera tracking.
02What makes it difficult
Latency budget. Every stage — capture, stitching or reconstruction, rendering, encoding, packaging, delivery, decode — adds delay. Interactive elements, where audience input must affect what is broadcast, require a round trip that fits inside that budget, and conventional streaming delivery is not designed for it.
Synchronisation. Multi-camera capture, tracking data and rendered elements must share a common timing reference. Rendered content that is a frame or two out of step with the camera reads immediately as wrong, even to viewers who cannot say why.
Bandwidth and encoding. Delivering a full sphere means delivering resolution the viewer is not currently looking at. View-dependent delivery, which sends higher quality only for the region in view, reduces the waste and adds complexity and a dependency on fast reaction to head movement.
Playback diversity. The same production is consumed on headsets, phones, desktop browsers and conventional screens, and a design that only works well in a headset reaches a small fraction of the audience. Planning the flat-screen version as a first-class output rather than a fallback is usually the difference between a project that ships and one that demos.
Production discipline. Immersive broadcast merges live television practice with real-time engine development. Both disciplines have to be in the room, rehearsal requirements are higher than for either alone, and the failure modes are additive.