Professional Metadata (PMD)

Professional Metadata or PMD is a new professional audio metadata format created by Dolby. PMD enables a new set of use-cases aimed at supporting next generation audio features such as

  • Immersive audio
  • Dialogue enhancement
  • Multi-Language and audio-description support
  • Commentary/Dialogue selection and adjustment
  • Audio-object movement
  • Headphone virtualization

Typically, PMD is authored during production and then carried through to an emission encoder. PMD flows alongside the audio, occupying a spare audio channel within the SDI, MADI or SMPTE ST 2110 streams. The emission encoder will apply or inject the various metadata parameters, as appropriate.

PMD is an open format and a license is not required to use it. The payload specification is disclosed in SMPTE RDD49. The method for carriage of PMD over SMPTE ST 337 and AES3 is standardized in SMPTE ST 2109.


Documents


Sample Streams

This zipfile contains 4 PMD test streams in three different formats: XML, PCAP and WAV

Each PMD stream covers a different use-case or scenario:

  • pmd_51main - 5.1 Main with 1 presentation
  • pmd_51eng_and_spa - 5.1 M&E plus English and Spanish dialogue tracks with 2 presentations
  • pmd_514_eng_spa_vds - 5.1.4 M&E plus English and Spanish dialogue and VDS tracks with 4 presentations
  • pmd_514_4objects - 5.1.4 M&E plus 4 dialogue objects in different positions with 4 presentations

Code

Source code to create PMD binary payloads from XML is available on github under an open-source license. Typically a user-interface would be used to set the various metadata parameters. The resulting XML can then be converted into a binary payload that can be inserted into a metadata stream using KLV and the SMPTE ST 2109 standard. This metadata stream can be carried over AES3/SDI using SMPTE ST 337 or IP using SMPTE ST 2110-31.


Audio Metadata Authoring Tool

The Audio Metadata Authoring Tool is a Windows based application that allows the user to easily author metadata in both PMD, ADM and Serial ADM formats. The application supports the following features:

  • Multiple audio objects and presentations
  • Selection of audio object type (Dialog, VDS etc.)
  • Selection of audio object position gain and divergence
  • Language and presentation name language selection
  • HTTP post to specified URL
  • XML file output with update trigger to enable downstream streaming