Report on MP3 Searching
MP3 files are the most common form of digital music served on the Web at this time. In addition to the encoding of music itself, this file format has provision for a small amount of metadata about the file. This textual metadata is indexable and searchable if the search engine knows about the file format.
The original MP3 file format didn't have any way to store metadata about itself. One early adopter, Eric Kemp, decided to add that information by creating a simple tag format, ID3. This, tacked on to the end of the MP3 files, provided space to store the song title, artist, album, year, a comment and a one-byte code for genre (Blues, Classic Rock, Country, Dance, Disco, etc.). This was soon extended to ID3v1.1 by Michael Mutschler who added a field for the track number of the song on the CD. For more informatin, see ID3 Made Easy [v1] and the short history of tagging at the id3.org web site.
ID3v1 is quite limited in scope, so in 1998 a group of people, led by Martin Nilsson, decided to take the idea and run with it. ID3v2 is much more flexible and extendible, adding a concept of "frames" (like programming objects) which can contain many kinds of data. Unlike ID3v1, it is stored at the beginning of the MP3 file, so the metadata can be downloaded without waiting for the entire song. The standards group has been quite ambitious, so the frames can include anything from encryption methods through Music CD identifiers, lists of the people involved in the music, synchronized lyrics, legal information and attached pictures.
Using ID3 In Search
Napster recognizes the ID3 tags, as does the MediaFind online search engine. Most MP3 players and editors allow people to edit these tags, and they are becoming very common. It's easy to download and upload them from the CDDB database (now developed by Gracenote).
Several site search engines and services do index one or both versions of the ID3 tags: see below for a listing.