Live AIV encoding is new enough that most teams are doing it for the first time. This is the checklist we would hand a production team before their first live Apple Immersive Video (AIV) event.
It is deliberately practical. For the background, read live AIV, from camera to Apple Vision Pro first.
1. Camera
- An AIV camera. Apple Immersive Video needs per-lens calibration, so the camera has to be built for the format. The Blackmagic URSA Cine Immersive is the one we see on almost every production.
- The calibration data for each camera body. It is specific to the lenses on that unit. Mixing it up between cameras produces a picture that is wrong in a way viewers feel before they can describe it.
- A decision on camera count. Every extra position is another full live encode. Spectrum Front Row runs five.
2. Signal path
- ProRes over SMPTE ST 2110 from camera to encoder.
- A network path sized for it. A 16K, 90 fps feed needs a 100 Gb/s class link. Check switches, optics, and cabling, not just the endpoints.
- The encoder in the venue. This signal is not something you send to a distant cloud region to be encoded.
3. Encoder
- Real-time MV-HEVC at your ingest resolution and 90 fps. Ask for proof at the real resolution, not a scaled-down demo. More on why in live MV-HEVC.
- Zero dropped frames over the full event length. Test for the length of the event, not for ten minutes.
- A plan for failure. Know what happens if an encoder, a link, or a power feed goes down mid-event.
4. Bitrate ladder
- More than one rung. Home connections vary, and a single high bitrate will stall for a lot of viewers.
- A sensible range. Apple recommends 4320 x 4320 per eye at 90 fps for delivery, with average bitrates from about 25 to 150 Mbps. For reference, Spectrum Front Row’s top output is 180 Mbps.
- Aligned rungs. The player has to be able to switch between them without a visible jump.
5. Metadata
- Venue description (AIME) referenced from the playlist.
- Stereo layout and AIV projection signaled in
REQ-VIDEO-LAYOUT. - HDR signaled correctly. AIV is P3-D65 with PQ.
- Apple Spatial Audio in the audio group.
If any of these are missing, the headset will not treat the stream as Apple Immersive Video. Our post on how to stream Apple Immersive Video shows a full sample playlist.
6. Delivery
- An origin and CDN that are comfortable with high-bitrate HLS. At these bitrates, egress adds up quickly. Know the cost before the event.
- A flat simulcast, if you want one. An HEVC or H.264 feed for YouTube or Twitch can come from the same source.
- A latency target. We delivered under 20 seconds glass-to-glass on Spectrum Front Row. Decide what your event needs.
7. Playback
- An Apple Vision Pro app or player that supports AIV over HLS.
- A real headset in the loop during the event. Someone should be watching the actual stream on the actual device the whole time. Scopes do not catch everything.
8. Rehearsal
- A full-length rehearsal on the real network.
- A lighting check in the headset. Venue lighting that looks fine on a monitor can look very different in HDR at 180 degrees.
- A named person to call. When something changes an hour before the event, you want an engineer, not a ticket number.
How we cover this list
Zeus handles sections two through six in one 3U system: ProRes 2110 ingest at up to 16K and 90 FPS, real-time MV-HEVC, packaging, and delivery to SpatialGen, AWS, GCP, Azure, Akamai, and others. Every Zeus comes with a year of direct access to our engineers.
The longer explanation of the encoding stage is in live Apple Immersive Video encoding.
Planning a live AIV event? Tell us about it.
SpatialGen
SpatialGen
contact@spatialgen.com