media-metadata-examples

This is an exploratory page to be used for storage of various multimedia metadata profiles currently in use around the web. As this is a very, very early exploration, we should include as many types of multimedia as possible in the opening discussion - but please be sure to place your profiles in the correct portion of this page.

Purpose of this exploration (or : The Problem)

The purpose of the studies on this page is to determine the feasibility and demand for a flexible, semantic markup format for providing metadata alongside linked multimedia files. Typically, this metadata is stored within the header of the media file itself - which has massive implications for any application where metadata is to be indexed, searched, or made externally accessible. Of course, even if you could easily access the correct portions of a media file remotely, you'd still have to cope with a multitude of open and proprietary metadata formats, each with it's own distinct fieldset, nomenclature and storage method.

This study aims to make a start at solving this problem.

Still image

• EXIF
• RDF for self-describing images
• An Indexing and Querying System for Online Images Based on the PNG Format and Embedded Metadata
• Describing and retrieving photos using RDF and HTTP

Audio

• ID3v2

• iTunes
  • Album : String
  • Artist : String
  • Beats Per Minute : Number
  • Bit Rate : Number
  • Comment : Blob
  • Composer : String
  • Genre : String or Foreign ID
  • Disc Number : Number
  • Kind : Proprietary - could be implimented as MIME type
  • My Rating : Number < 5
  • Sample Rate : Number
  • Size : Number
  • Time : Number
  • Track Number : Number
  • Year : Number
  • Additional metadata used internally by the application : Date Added (Timestamp), Date Modified (Timestamp), Equalizer (foreign ID), Play Count (Number), Grouping (Internal)
  • Noteable absences : Tags, License, Copyright year. Also note common complaints about ID3 and classical music.

Podcasts

• iTunes podcast: http://phobos.apple.com/static/podcast_specifications.pdf (local copy, in HTML, updated version from Apple Chapter Tool documentation)

iTunes release 4.9 was widely heralded as an advance for mainstream acceptance of podcasting. However, while some advocates resist even the fragment of a trademarked name in the label for the phenomenon, the use of "itunes:" in the namespace is one of many design decisions in Apple's original specification that became focal points of debate. While Apple has a tradition of working on breakthrough features very quietly, they have been open to public input after its initial release. Many other bloggers have chronicled some of the feedback, as well as some face-to-face discussions.

A mailing list may be coming soon...

• Media RSS: http://search.yahoo.com/mrss

Media RSS has been in the works longer, but doesn't have the overnight-adoption advantage of iTunes. However, it has a much broader ambit, including video; and a much broader community of interest.

A Podcasting Microformat?

The best starting point for understanding these two proposals is W3C's summary table, by Karl Dubost.

There are (at least) two paths to consider when 'porting' these proposals into Microformats:

1. Standalone. What is the most straightforward rendering of each proposal into XHTML? This keeps the interests of the developers of consuming applications foremost: how can the migration be made as painless as possible?
2. Refactored. Given the existing core of Microformats, what is the minimal (!) necessary to add on to RelLicense, RelTag, hReview, and so forth? This favors (we'd hope) content publishing applications: how can the marginal migration costs of adding media be made as painless as possible.

Mapping the existing specs requires following the data definitions set forth in each specification. In summary:

Comparing existing proposals

Aspect	Apple	Yahoo!	Microformat?
Categorization	<itunes:category> <itunes:keywords>	@@	RelTag
Rating	<itunes:explicit>	?	?
Description	<itunes:subtitle> <itunes:summary>	?	?
Authorship	<itunes:author> <itunes:owner> <itunes:name> <itunes:email>	<media:credit>	hCard
Metadata	<itunes:duration> <itunes:image>	<media:thumbnail>	?
Licensing	<itunes:block> <copyright> <link>	?	?

Video

• EXIF - (Yes, the spec deals with video taken by still cameras. -- RyanKing)
• MPEG-7 - MPEG-7, formally named "Multimedia Content Description Interface", is a standard for describing the multimedia content data that supports some degree of interpretation of the information meaning, which can be passed onto, or accessed by, a device or a computer code. (Very powerful, but you have to pay for documentation -- ChristopherA 01:53, 29 Jun 2005 (PDT))
• SMIL - Synchronized Multimedia Integration Language (SMIL, pronounced "smile") enables simple authoring of interactive audiovisual presentations. (Not exactly video, but has a lot of useful video-related features in it. -- ChristopherA 01:55, 29 Jun 2005 (PDT))

Interactive

• SVG
• SMIL
• QTVR?

Offline Media

(I'm not sure if you want to open the scope up this much.)

Printed Publications (books/magazines)

• Amazon

Movies

• IMDB