As part of the emergence of “web 2.0″ websites a few services that let the user discover new music have been developed. Let´s look at three of them and compare them.

Pandora

The probably most well known music discovery service is Pandora, available at pandora.com. It´s based on the music genome project that records the different musical attributes of songs. The user enters a few interprets and songs he likes and Pandora starts playing music with similar attributes. Rating the songs with “thumbs up” or “thumbs down” refines the result. The user can create up to 100 stations. The services database can also be accessed via a “backstage” area where it is possible to search for, discover and preview new songs.

Limitations:

• Currently not available outside the USA
• Six songs can be skipped per hour
• It´s not possible to select and play a specific song

Last.fm

Last.fm is not restricted to a specific country. The approach is similar to Pandora, but songs similar to the users taste are found according to “tags” and what other users with the same taste hear. When first using last.fm the user is asked to install a program called “the scobbler” on his computer and listen to his favorite music. The user then submits songs he likes to the system together with some tags of his choice (for example: metal, sad, piano, female singer). The system then recommends and plays songs with similar tags. Each song that plays can be “loved” (which then puts it on a list of favorites) or “banned” (this song will not play anymore). The system provides a list of users with similar music taste and lets you listen to their music collection.

Additional functions:

• Statistics over which songs you listen to most often
• Listen to your favorite songs
• Social networking functions: friends, groups, comments, messages, blogs, events

Limitations:

• Playing songs can´t be paused
• No playing of specific songs
• Favorite songs channel only available by paying a small monthly fee

The Filter

The Filter is a pure music and movie discovery service with no radio function. The service learns the users music taste by recording which songs are played on the users computer (via a small plugin) and asking the user to rate songs and recommended songs. Songs can be filtered according to popularity and obscurity (how much this factors into the recommendations can be chosen by the user).