Home / A Short History of Auto Tagging Music

A Short History of Auto Tagging Music

Please Share...Print this pageTweet about this on TwitterShare on Facebook0Share on Google+0Pin on Pinterest0Share on Tumblr0Share on StumbleUpon0Share on Reddit0Email this to someone

The music files on your computer contain much more information than just the music itself. They contain meta-data about the file contents; information such as the names of the tracks, artists, albums, and much more. All of this information is stored adjacent to the audio data itself, in data items called tags.

Unfortunately, the presence of this useful data is entirely optional, and more often than not is either incomplete, or worse, inaccurate. There are many causes for this; perhaps the program that ripped the files from the disc failed to copy the meta data, or perhaps it wasn’t supplied at all. Problems can also arise when people convert between the many different audio file formats. The data is often missing in audio files downloaded from the internet.

This can be a major inconvenience, because music player programs use this meta-data to organise your music collection. Whenever your music player offers features such as browsing your collection based on genre or artist, it relies on your tags being accurate and complete. If this information is left blank, all your music will be lumped together into one group.

If you want to fix this problem, you need to add the correct tags to your music files. This is something that can be done manually, if you’re a patient person with a lot of time to spare; this would be the modern day equivalent of writing all the details on the cassette sleeve of a mixtape. Unfortunately, this approach quickly becomes infeasible for people with the size of music collections people tend to have these days.

The earliest attempted solution to automatic tagging was the Compact Disc Database (CDDB). This provided a lookup database for disc ripping programs to look up basic album information over the Internet, by recognising the unique features of the CD’s “table of contents” (basically, the position of the tracks on the CD). This was only a useful approach for copying commercial discs, making it entirely useless for custom discs or individual tracks.

More modern approaches try to tackle the problem more intelligently, by trying to actually recognise individual tracks by analysing the audio signals. This process is called audio fingerprinting, and it can be used to match together tracks that sound similar or identical to the human ear. The algorithms behind these approaches were originally closed source and expensive to use; however, open source alternatives have recently emerged, making the technology widely available and significantly cheaper.

In summary, tags contain useful information about your music, which is used by music players to arrange your collection. Tags are often left incomplete when the music files are copied from discs or the Internet, which means that music programs don’t have the information needed to organise your collection properly. If you have the time and patience, you can manually tag your files, but for even modestly sized collections, this isn’t feasible. These days, music tagging can be done automatically by programs that analyse the audio files and recognise your music, which will fill in all the missing details for you.

Powered by

About Dan Gravell