Easy Semantic Markup of Audio Recordings on the Web

It is difficult for a browser to extract semantic information about an audio recording described on a web page. Metadata such as speaker, musician, publisher, label, title of the work, release date, download link, related image artwork and tags provide relevant context for the audio recording, but are currently (2008) inaccessible to computers.

Having such information marked up in a semantic way can provide a number of benefits to the viewer. If a web browser understands that a particular web page contains a song performed by an artist, it can produce richer interactions. For example, specific searches may be performed for artists and songs via general search services such as Google and Wikipedia. Specific search services may also be queried such as MusicBrainz, The Internet Archive, FreeDB, or Bitmunk. Additionally, classification by crawlers can become more accurate. If there are 20 tracks found on a page performed by the same artist, and that content consumes a significant portion of the page, it can be assumed that the page is not only about music, but also about a particular artist.

In order to enable and encourage the sharing, distribution, syndication, and aggregation of audio content, the authors propose the Audio RDF vocabulary, an open standard for distributed audio metadata. The authors have researched both numerous examples in the wild and earlier attempts at audio metadata formats, and have designed the Audio RDF Vocabulary around a simple minimal schema for audio content. Feedback is encouraged and open participation is desired from all that are interested in contributing to the open exchange of audio metadata on the Internet.

The Audio Vocabulary

The easiest way to use the Audio RDF Vocabulary is to embed semantic audio metadata in your web pages using RDFa. For examples on how to do this, please visit the Audio RDF Vocabulary section on the RDFa wiki.

Base Vocabularies

The Audio RDF Vocabulary relies heavily on aggregating functionality in previously defined vocabularies. Vocabulary term re-use is of primary importance. What follows are the vocabularies on which the RDF Audio Vocabulary depends:

xhv
The XHTML Vocabulary namespace which defines the license keyword.
dcterms
The Dublin Core Terms Vocabulary is used for describing various aspects of the audio recording such as title, description and contributors.
media
The Media Vocabulary is used for describing apects of the audio recording that are common to all types of media.
commerce
The Commerce Vocabulary is used for describing the financial costs of acquiring audio recordings.

Classes

audio:Recording
Status stable
Description A recording of acoustic signals.
Subclass of media:Recording
audio:Album
Status stable
Description A collection of one or more audio recordings.
Subclass of

Properties

dcterms:title
Status stable
Description title of the audio recording
Datatype plain literal
dcterms:creator
Status stable
Description The primary creator or performer of the audio recording.
Datatype (pick one)
dcterms:contributor
Status stable
Description An entity that helped create the audio recording.
Datatype (pick one)
dcterms:published
Status stable
Description The date the audio recording was published.
Datatype (pick one) xsd:date using ISO-8601
dcterms:description
Status stable
Description The description associated with the audio recording.
Datatype (pick one)
media:position
Status stable
Description The position of the audio recording in an album, LP, playlist, top 10 list, podcast history or other ordered list of audio recordings.
Datatype (pick one)
media:sample
Status stable
Description A link to a sample file or stream of the audio recording.
Datatype URL
media:download
Status stable
Description A link to a complete audio file representation of the audio recording.
Datatype URL
media:depiction
Status stable
Description An image summary of the audio recording.
Datatype URL
dcterms:type
Status stable
Description A genre that should be associated with the audio recording.
Datatype plain literal
xhv:license
Status stable
Description The license under which the audio recording is distributed.
Datatype URL via rel="license"
media:duration
Status stable
Description The length of the audio recording encoded in ISO-8601 time duration format.
Datatype xsd:duration using ISO-8601
media:contains
Status stable
Description Relationship for linking songs to albums, audio snippets to podcasts and singular recordings to collections of recordings.
Datatype URI using rel="media:contains"
commerce:payment
Status stable
Description A link to a method of paying for a full version of the audio recording.
Datatype URL using rel="commerce:payment"
commerce:costs
Status stable
Description A link to a method of paying for a full version of the audio recording.
Datatype commerce:Price (containing both commerce:currency and commerce:amount)

Scope

Audio content consistently share several common fields. Where possible the Audio RDF Vocabulary has been based on this minimal common subset.

Out of scope

Fields that are type-specific have been omitted from the Audio RDF Vocabulary. It is important that this vocabulary be kept simple and minimal from the start. Additional features can be added as deemed necessary by practical implementation experience.

The concept of a universal audio identifier, that is, how to identify the same audio album, song, speech, or podcast across different music and audio sites, though something very useful to have, is outside the scope of this format.

Extending the Vocabulary

This vocabulary is the embodiment of a large group effort lead by Digital Bazaar, Inc. and pulls resources from the World Wide Web Consortium, the RDFa community and the Microformats community. It is a world-wide effort in semantic metadata standardization.

If you would like to extend this particular vocabulary, we ask that you engage the community in doing so. History has shown that if there is a wide-spread need for a vocabulary extension, it will be discussed, and agreed upon by the larger community before making it's way into the vocabulary. This process can take as little as a week. Work with us before forging out on your own, the knowledge of the masses are your ally.

Copyright and Patents

This document is copyright 2008 Digital Bazaar, Inc. and is licensed under a Creative Commons Public Domain Dedication.

All ideas and patentable material outlined in the document are hereby dedicated to the public domain. It is our intent that all future additions, modifications and contributions will be dedicated to the public domain as well.

We have released our copyright and control of this vocabulary for the good of the community and the betterment of the world in the name of open standards.

We kindly ask that proper credit is given when using the vocabulary. A line like the following is sufficient:

The Audio RDF Vocabulary is an initiative lead by Digital Bazaar, Inc. and collaborated on by a number of people from the Web at large, the World Wide Web Consortium, the RDFa community and the Microformats community.

Acknowledgements

Many thanks to the various individuals that did research and proposed ideas and discussion related to media info and audio info in general. Among the many participants are RodBegbie, Dean Hudson, Tantek Çelik, Mary Hodder, Joshua Kinberg, ChrisMessina, and Lisa Rein. Others include Martin McEvoy, Alexandre Van De Sande, Michael Johnson, Dave Longley, Brian Suda, Ben Wiley Sittler, Scott Reynen, Frances Berriman, James Craig, David Janes, Andy Mabbett, Danny Ayers, Rudy Desjardins, Edward O'Connor, Ryan King, Chris Griego, Brad Hafichuk, Colin Barrett, and Joe Andrieu.

Many thanks as well to Mike Kaply, Ben Adida, Mark Birbeck, Ralph R. Swick, Shane McCarron, Michael Hausenblas, Steven Pemberton, and Mike Linksvayer for guidance with regards to RDFa.

Editor

Manu Sporny, Bitmunk - Digital Bazaar, Inc.

Vocabulary Contributors

Manu Sporny, Martin McEvoy, Alexandre Van De Sande, Michael Johnson, Dave Longley, Brian Suda, Ben Wiley Sittler, Scott Reynen, Toby A. Inkster, Frances Berriman, James Craig, David Janes, Andy Mabbett, Danny Ayers, Rudy Desjardins, Edward O'Connor, Ryan King, Chris Griego, Brad Hafichuk, Tantek Çelik, Colin Barrett, Joe Andrieu, Michael Smethurst, Chris Newell, Julian Stahnke, Justin Maxwell, Paul Wilkins and David I. Lehn.