Introduction - General Information about ADM

The ADM standard (Audio Definition Model) describes the technical properties of audio using metadata. This provides a high degree of abstraction and makes it possible to depict complex audio data streams. In addition to the actual audio data, additional properties can be encoded within the audio file:

• 3D Audio: Sound sources that are called ADM objects in ADM terminology can be statically placed in a room or moved. Their positions are specified directly and converted by the renderer into volume components for the respective loudspeaker channels during playback (in accordance with the ITU-R BS.2051 standard).

  ADM objects should not be confused with Sequoia objects. These are different concepts with different meanings. In this help chapter, the term “objects” always refers to ADM objects.

• 3D audio as a multi-channel surround mix can be integrated as a direct speaker object (Dolby Atmos: Bed) (background audio). Depending on the speaker configuration used, a downmix or upmix takes place in the target system during playback by the renderer. Other audio types such as ambisonics, matrix or binaural audio (headphone output in which the spatial information is encoded into the stereo signal using filters and phase shifts) are also possible.

• Content (contents or programs) can be provided with a language setting and the finished file can contain several alternative language versions. Objects can also be marked as interactive, allowing the audio data to be moved in 3D space during playback, the volume to be adjusted separately, or even muted completely.

• Data streams can be marked as having different levels of importance so that a decision can be made as to which data streams are omitted when transmitting content with limited capacity.

The metadata is defined independently of the code (codec agnostic). This means that ADM audio data is described using the model. The renderer is free to decide which of these properties are evaluated and mapped and how. Since not all renderers currently support every feature of the ADM specification, a target renderer must be specified for export and preview (monitoring). Currently, Sequoia supports the renderers EAR (EBU ADM Renderer), and Dolby Atmos.

The exported ADM file is a broadcast wave file in which all audio data is stored in an unstructured format. It also contains a section with the ADM metadata in XML format. This metadata defines both the logical structure of the file, such as the grouping and movement of the sound sources, as well as the technical specifications including sample format, number of channels, and the layout of audio streams.

Structure of the ADM metadata

The ADM metadata has a hierarchical structure and can contain the following elements and sub-elements:

• Program: A complete audio production with all content played during a certain period of time. (The term “programs” comes from the ADM standard and refers to a complete audio unit, not a TV or radio program). An ADM file can contain several programs, for example for different languages. A program contains one or more contents.

  Note: The Dolby Atmos renderer only supports one program.

• Content: Part of a program, e.g. a dialog or the background music. Typical metadata assigned to the content includes the language of the dialog and the types of content (e.g. dialog or music). A content contains one or more objects.

• Object: The actual audible components of the content. An object is either

  • An Object: Any audio content with a specific position in 3D space. The surround mix according to the perceived position in the room is left to the ADM renderer, which realizes this according to the loudspeaker configuration used.

  • A Direct Speaker Object: Channel-based surround audio with a specific speaker configuration, for example the audio output of a surround bus.

  Objects can be marked as interactive. This makes it possible to mark objects as mutually exclusive (complementary objects) - for instance dialogs in different languages - or to specify objects that can be adjusted separately in position and volume during playback by the listener. In addition to the ADM objects, there are other specific object types that are specialized for particular applications.

A further section of the ADM metadata defines how the audio data within the wave file is grouped into the individual objects (tracks, streams, channels, packs). These specifications correspond to the conventional format parameters for audio data, including the number of channels and speaker assignment (for channel-based objects) as well as sample rate, bit depth, and codec. In Sequoia, this information is derived from the format of the source material and the channels of the Sequoia objects, tracks and surround busses used in the project and therefore does not need to be edited directly.

Overview of the ADM workflow

When creating a 3D audio mix with the ADM editor, the basic procedure is as follows:

1. Create a new project with 48 or 96kHz sample rate.

2. Assign the audio to the ADM objects on the tracks in the project. Import all the required audio files onto separate tracks. When importing an audio file that is already available in a surround multi-channel format, e.g. for use as a bed in a Dolby Atmos mix, a corresponding surround submix folder is created.

3. Open the ADM Editor and create programs and content according to the planned structure of the ADM and create the corresponding objects for all audio content: For the surround folders as a direct speaker object, for the other audio tracks as simple objects.

   For a simple standard layout with a single program and one content, as the standard used in the Dolby Atmos format, you can use the AutoFill button. An object is automatically assigned to each simple track, while surround submix busses or folders are converted into the corresponding direct speaker objects.

4. Place the objects in 3D space using the surround panner of the corresponding tracks.

   If necessary, automate the object movements. In all tracks in which objects have been linked, the standard panorama slider is replaced by a special version of the surround editor.

   For acoustic monitoring of the panning: Activate ADM monitoring. This creates a surround bus into which the render plug-in for the corresponding output format is inserted. In the Monitoring settings dialog, set the output format that matches your speaker system.

5. Add the required information for interactivity or language variants (e.g. attributes, complementary objects, interactivity) to the objects (only for ADM).

6. Export the ADM or Dolby Atmos audio file. If Dolby Atmos was selected as the target format, the file can subsequently be converted to the desired distribution format using the appropriate encoder.