Dolby Atmos

Dolby Atmos is a content creation, encoding, and delivery technology that provides immersive audio to consumers. Dolby Atmos adds overhead audio but more importantly renders an Atmos mix appropriate to the listeners device and/or speaker configuration. Atmos is not an audio codec, but can be delivered by multiple codecs. You can read our Dolby Audio Overview Tutorial.

There are two kinds of Dolby Atmos:

• Channel Based Immersive (CBI) is typically 5.1.4 (five floor channels, Low Frequency Effects, and four over head channels) and is used for live events.
• Object Based Immersive (OBI) that introduces the concept of audio objects and spatial coding. Object based immersive is used for post produced/file based content.

Object Based Immersive

Dolby Atmos content is comprised of bed channels (which correspond to specific speaker locations) and audio objects that have positional coordinates (x,y,z) and “size” metadata. Bed channels can be from LCR up to 7.1.2 with two center overhead channels. During content creation, up to 128 channels can be used. With a single 7.1.2 bed, this leaves up to 118 channels to be used for objects. If multiple beds are used to create stems for localization purposes, the number of channels available for objects decreases.

Content Creation and Mastering Tools and workflow

Atmos audio and panning metadata can be generated by DAW tools such as Avid Pro Tools, Steinberg Nuendo, and Black Magic Resolve.

During the mixing and mastering process, the bed channels, object audio, and object metadata are sent to a Rendering and Mastering Unit (RMU) running the Dolby Atmos Mastering Suite software. Audio is transported over MADI or Dante, panning metadata via ethernet, and LTC is used to chase source TC and set In/Out points. Alternatively the Dolby Atmos Production Suite can provide similar functionality running on the DAW.

Dolby Atmos for the Home

For home distribution, the delivery of 128 channels of audio and metadata is not practical. Spatial coding is employed along with additional metadata in Dolby Digital Plus to create a backwards compatible delivery that provide a cinematic home theater experience to consumers with Atmos-enabled equipment and 5.1 to those without.

Spatial coding is employed to reduce 128 bed and object channels to 12 or 16 elements or “clusters”. Actually, this is really 11.1 or 15.1 as the LFE doesn’t move. Spatial coding works by employing an algorithm to dynamically group audio into dynamic elements. Audio can move from cluster to cluster and the clusters themselves move as needed. While reducing the channel count from 128 to 16 sounds significant, keep in mind that the full number of audio bed channels or objects are rarely all active at the same time.

This image shows reducing/combining into "clusters" of audio elements

Once spatially coded, the audio can be rendered to 7.1 or 5.1 and encoded as Dolby Digital Plus JOC (Joint Object Coding) with each element’s Object Audio Metadata (OAMD) and JOC payload included in the bitstream. None of the audio is discarded. The minimum datarate is 384kbps (with 12 elements) or 448kbps (with 16 elements). Once this bitstream hits an Atmos enabled device, JOC is used to extract the audio elements and the Object Audio Renderer (OAR) uses the OAMD to render appropriately to the consumer’s listening environment. Non-Atmos devices can still use the 5.1 “core,” making it cross-compatible.

Atmos Master Files

There are four types of Atmos Master Files

DAMF (Dolby Atmos Master File)

The DAMF is a collection of 3 files created by the Rendering Master Unit (RMU). This is comprised of:

• filename.atmos
  • The .atmos file is an xml file describing the name of the other two files, how many beds and objects are utilized, the start time (offset), the FFoA (first frame of action, often the same as the starttime), the framerate, and other information.
• filename.atmos.metadata
  • The .atmos.metadata is an xml file with xyz and size coordinates for objects over time. These are large files.
• filename.atmos.audio
  • The .atmos.audio is a Core Audio File (CAF) of up to 128 tracks. These files are the largest.

ADM BWF (Audio Descriptive Model Broadcast WAV Format)

This file is an alternative to DAMF and is not proprietary to Dolby. It is a single file that is basically a broadcast WAV with a huge data chunk at the head containing the .atmos and .atmos.metadata information.

ADM BWF can also be imported back into Pro Tools 12.8 and up to recreate a Pro Tools session with all audio and panning metadata. ADM BWF files can be exported from the Mastering Suite or converted from DAMF using the Dolby Atmos Conversion Tool.

IAB/IMF (Immersive Audio Bitstream for IMF)

This is a frame based representation of the DAMF in a single file, which is specified for IMF (Interoperability Mastering Format). IMF mastering tools can be import and export IAB/MXF and master to IMF. IAB/IMF is an .mxf file with interleaved PCM and metadata.

9.1.6/7.1.4/5.1.4 PCM

The Dolby Atmos Mastering Suite and standalone Dolby Atmos Production Suite can render to 9.1.6, 7.1.4, 5.1.4. Rendering Object Based Immersive (OBI) to Channel Based Immersive (CBI) is a hybrid workflow. OBI is used during the creative process and CBI for encoding and delivery.

• The 9.1.6 is rendered by the OAR in the consumers home and with speaker layouts at 9.1.6 and below is identical in listening tests to the rendered OBI. While not technically delivering “real” OBI the use of 9.1.6 can simplify downstream workflows for editing, joining idents and ratings card, watermarking, etc.
• 7.1.4 output is provided to allow for immersive cut scenes and trailers in gaming engines to accompany realtime game play.
• 5.1.4 is provided to facilitate immersive post produced sound design for live playout CBI applications.

Loudness in Dolby Atmos

Loudness measurement and setting of dialnorm metadata takes place during the encode. The audio is rendered to 5.1 and this is used for measurement and to set dialnorm metadata properly during the Dolby Digital Plus JOC encoding process to ensure Atmos content is compatible with non-Atmos content.

Converting between Atmos Mezzanine Formats in Hybrik

See the sample jobs at the bottom of this page. Hybrik has the ability to convert between Atmos mezzanine formats, including:

• DAMF
• ADM BWF
• IAB in MXF

Trimming Atmos

Hybrik has the ability to trim Atmos content when encoding to DD+JOC using the dolby_audio_encoder task, and can also trim when converting between Atmos mezzanine formats. Currently, trim must be specified in seconds.

If inpoint_sec parameter is not specified, the encode start is determined automatically, using first frame of action (FFoA) metadata in the input file.

Frame Rate & Sample Rate Conversions

Hybrik is able to convert frame rates for Atmos ADM.

{
  "file_pattern": "{source_basename}_framerate_30.wav",
  "existing_files": "replace",
  "container": {
    "kind": "adm"
  },
  "audio": [
    {
      "frame_rate": "30"
    }
  ]
}

Hybrik is also able to convert sample rates for Atmos ADM, which can be useful when you need to process a 96kHz source asset.

{
  "file_pattern": "{source_basename}_samplerate_48k.wav",
  "existing_files": "replace",
  "container": {
    "kind": "adm"
  },
  "audio": [
    {
      "sample_rate": 48000
    }
  ]
}

Table of Terminology

Examples

Note: The following examples will require a license. Please speak to your Dolby contact.

Mezzanine to DD+JOC

• DAMF to DD+JOC
• ADM BWF to DD+JOC
• ADM BWF to DD+JOC with Trim

Mezzanine to Mezzanine

• DAMF to other Mezzanines
• IAB to other Mezzanines

Mezzanine Conversions

• ADM Frame Rate Conversion
• ADM Sample Rate Conversion

Mezzanine Trim

• ADM BWF with trim to ADM BWF
• DAMF to ADM, IAB, DAMF trimmed

Dolby Atmos Mezzanine Analysis

• See our Dolby Atmos Mezzanine Analysis tutorial for an explainer on Mezzanine Analysis