Dolby Atmos

Audio, position metadata, and a renderer

Our REAPER mix was channel-based

ReaSurroundPan distributed audio among eight channels
Automation changed that distribution over time
The exported 7.1 WAV stored those channel signals
An Atmos object keeps audio and position metadata separate

Bed format and speaker layout

Format	Channels	Role in this course
7.1	8	Our REAPER mix and room playback
7.1.2	10	Our Logic Atmos bed
7.1.4	12	A monitoring layout with four height channels

The bed format and monitoring format are separate choices.

Beds and objects

Bed: tracks combine into a channel-based submix
Object: audio stays separate, with position and size metadata
Either can contain stationary or moving sounds
A stereo object track uses two object inputs

Two paths into the renderer

Simplified course diagram. Eventide routing reference · Logic signal paths

Rendering for the playback system

Bed channel signals and object metadata enter the renderer
The renderer creates outputs for the selected layout
Headphones require binaural rendering
Different playback systems still need listening checks

Ambisonics and Atmos

	Ambisonics	Atmos
Stores	Encoded sound-field components	Bed audio, object audio, metadata
Positions	Encoded into component signals	Object positions stored separately
Playback	Decode for speakers or headphones	Render for speakers or headphones

Both can preserve height and adapt to different playback layouts.

Recall Project 3: source > ATK encoder > B-format mix > decoder > outputs. The four channels of full-sphere first-order B-format are sound-field components, not four speaker feeds. Higher orders use more components for greater spatial detail. Channel order and normalization must match the decoder; our ATK workflow uses FuMa conventions, while AmbiX uses a different ordering and normalization.

For Atmos: bed audio plus separate object audio and position metadata > renderer

outputs. An object channel is also not a speaker feed. Neither format requires one fixed playback speaker count, and both can produce binaural headphone audio. Playback flexibility does not guarantee equal localization on every system.

Ambisonics is not just a microphone technique: you can encode and automate individual sources, as in Project 3. Atmos is not just a larger surround layout. The distinction here is how the spatial scene is represented and retained.

One bird, one background

Ambisonics: encode the bird and traffic into a shared field
Atmos: keep the bird as an object; place the traffic in the bed
Rotate an Ambisonic mix and the whole field turns
Move the Atmos bird object and its background can stay put

What would you need to move only the bird in a finished Ambisonic mix?

Use the bird and traffic example from the Ambisonics introduction. In the Ambisonic session, the bird can move independently at its encoder before it joins the shared field. But a single mixed B-format master no longer retains that bird as a separately addressable source. Return to the source session or keep separate Ambisonic stems to revise it independently; rotating the finished field rotates the traffic too. Directional processing of a field is possible, but it is not the same as retrieving an isolated bird track.

In the Atmos master, an object retains separate audio and metadata, so a compatible authoring tool can revise its placement independently of the bed. Sounds already combined in the bed are not separately recoverable either. A source’s independent editability depends on what you preserve in the master or stems, not whether it was originally recorded with a particular microphone.

This is a workflow comparison, not a ranking. Ambisonics suits captured or constructed sound fields and field transformations; Atmos suits bed/object authoring and its delivery ecosystem. Both can mix recorded ambience with separately recorded sounds and both can animate spatial placement.

Listen to one placement change

Hear a short excerpt with one object held in place
Hear it again with one deliberate movement
Compare a 7.1 speaker render with a binaural preview
Describe what stayed clear and what changed

Two panners in Logic

Logic Surround Panner feeding the bed beside the 3D Object Panner sending object position metadata.

Choose a path and explain why

A lead vocal that stays in front
A sound effect that moves overhead
A diffuse ambience or reverb return

Which would you put in the bed or an object? What control do you need?

Binaural monitoring

Dolby Renderer modes: Off, Near, Mid, Far
Off includes the sound without distance modeling
LFE is fixed to Off in Logic
Apple Renderer has Music and Movie previews

A master and a listening file

ADM BWF: bed audio, object audio, and metadata
Rendered WAV: a particular channel layout or binaural preview
Playback encode: a service or device delivery format
Project 4 requires the Logic project and ADM BWF

References: routing and monitoring

References: masters and delivery

Appendix: other Atmos tools

Pro Tools Studio and Ultimate: integrated or external renderer
Nuendo: internal authoring and external-renderer support
Dolby Atmos Composer : adds Atmos authoring to other DAWs

Appendix: other immersive formats

MPEG-H Audio: channels, objects, scenes, and listener interactivity
IAMF: an open, codec-agnostic specification; Eclipsa Audio is based on it
ASAF and APAC: Apple’s production format and delivery codec