Preserving Immersive Audio Metadata
Preserving immersive audio requires carrying the spatial information needed by a renderer through the supported playback stages.
A moving sound needs both its audio and instructions about its position through time. If the goal is letting the receiver render the original immersive scene, the required spatial information must remain available until that receiver's renderer uses it.
Preservation depends on where the scene is decoded, transformed, and finally rendered.
Make a boundary ledger
For an illustrative Atmos source:
| Boundary or operation | Information carried onward |
|---|---|
| Copy compatible encoded source data | Audio and relevant spatial information in that bitstream |
| Object-aware decode to a suitable metadata-bearing output | Reconstructed audio plus required spatial information |
| Render to a chosen speaker layout | The generated speaker-channel presentation |
| Render to binaural headphones | The generated ear signals |
| Ordinary channel transcode | A new channel presentation, without automatically retaining source objects |
The first two routes can leave a downstream renderer the scene description. The next two have already completed a render. The final route creates another coded channel result from the signal it was supplied.
Rendering can be the intended end of preservation
A headphone player can use the original object directions to generate convincing ear cues. Once those cues are delivered to the headphones, there is no need for the headphones to receive and render the original objects again.
Likewise, a renderer can generate the supported speaker layout's PCM channels. That presentation can be correct for its target even though the metadata is no longer traveling onward.
The distinction is therefore the intended renderer boundary, not whether the word PCM appeared anywhere in the chain.
Use an object-aware decoded route when required
A conventional PCM decoder/output can supply the compatible channel presentation of an immersive track. Its array of 5.1 or 7.1 values does not by itself encode the original object's position updates.
An object-aware decoder can retain the appropriate reconstructed components and metadata. Supported Dolby MAT output can carry such information onward. The downstream system then applies the scene information to its own supported arrangement.
The compressed source form still matters: Dolby Digital Plus object carriage and TrueHD delivery require different parsing/reconstruction before reaching that boundary.
Identify irreversible combinations
A downmix or ordinary channel render combines source components for a chosen output. A later processor cannot uniquely infer the original objects and paths from those sums.
A generic E-AC-3 encoder also does not recreate an Atmos scene merely because the input came from Atmos. It needs the relevant scene representation and a supported encoding workflow; ordinary channel PCM is a different input.
An existing compatible core can keep a soundtrack audible while leaving the immersive extension unused. That is a particular component-selection result, not proof that the file's spatial information vanished from storage.
Verify the route using the active renderer
For an Emby session, record the source form, delivered presentation, decoder's output type, and where rendering occurs. If the receiver is intended to render, confirm its detected immersive input instead of only its upmix/listening mode. If the player renders headphones, evaluate that output as the finished ear presentation.
A basic channel-count report and height-speaker activity are insufficient evidence because a compatible channel mix and an upmixer can produce those observations. The boundary ledger establishes which spatial instructions actually remained available to the renderer making the final listening presentation.