To answer the OP's question in more detail, here is my guess:
The heavy lifting is performed on the iOS to TVOS device, where a Dolby Digital+ decoder handles the encoded stream and decodes the base channels plus objects, rendering them into a 3D binaural output using a generic HRTF (hopefully modified by actual users inter-aural distance). The rendering takes into account the positional feedback from the Airpods relative to the renderers position to provide the 'anchoring' of the audio.
This is pretty much what the Realiser A16 does (except it handles full-rez TruHD Atmos as well) if using the supplied default HRTF.
I have a music blu-ray with Atmos (I have several), how would I go about playing that on Apple (or anyone else’s) headphones?
For now, you rip, transcode to DDP in an .MP4 container, and copy that to your iPad. See
Music for Airpods: How to Convert Existing Surround Music to Spatial Audio Music
Longer-term, I expect Apple to release a box with the following:
- Multiple (6 to 8) HDMI inputs,
- Two HDMI outs, one of which supports eARC so TV apps can provide streams and one with video and audio to feed receivers
- Support for 4 concurrent Spatial Audio capable clients
- Ability to drive 4 or 5 HomePods as a surround setup (they have a patent to this effect).
A feature I hope they add but doubt it, is to have a 'sub out' that is engaged during spatial audio to drive a tactile transducer or something like a
SubPac (I have one and it's great). And have it dynamically adjust the delay based on the distance of the Airpods from the ATV.
The above would be the perfect solution for apartments and places where deploying a full-blown stereo is not possible, as this plus a couple of HomePods would allow for a decent music system, and a killer immersive audio setup when used with AirPods Pro or Max. They could call it the Pod-Caster