Mufin launches new music discovery engine beta, Download Squad readers are invited
As a music lover, one of my favorite innovations of the last five years or so has come in the form of music discovery services. Services like Last.fm, Pandora, Rhapsody and iTunes 8 have made it easy to find artists or songs that are similar to your musical tastes and preferences. The only problem with these services, is that they are based on limited databases (like the iTunes Music Store, for iTunes's Genius Mode), reviews or ratings by other users and overall popularity. That means that the music discovered might indeed be similar, but it really limits results to mainsream songs and generic genre distinctions. This is great for an automatic playlist, but not so great for really finding new bands or artists.
This is why Mufin's new music discovery engine is so unique; it analyzes the actual structure of the song, not just the genre meta-data. Mufin's discovery engine just entered private beta, but Download Squad readers can get in on the action by using this invite link: http://beta.mufin.com/start?ic=e75eecf85a4a547ca9379d6f8b6c23bb.
Co-developed at the Fraunhofer Institute (the creators of the MP3 format), Mufin uses audio recognition technology to analyze the actual musical characteristics of a song. Mufin creates a unique "fingerprint" for each song, using 40 characteristics like tempo, instruments, rhythm structure and sound density. Then, when you search for a song in Mufin's database, an alogrithm compares the fingerprint of that song against the database and presents you with results of songs that are similar in structure.
Right now, the Mufin database has more than 4 million songs, and you can submit your own music to the database to be analyzed and to appear in the search results. You can then listen to the results and you have the option to download or buy the track. Mufin has also launched a MySpace widget -- Facebook is getting an app very, very soon -- that lets you search Mufin's database from within MySpace and then attach results to your profile, share songs directly with friends and generate custom playlists.
PC users (software for Mac and Linux users is coming soon) can even run the analyzer on your own music library if you want to find similar tracks within your own collection. In the next month or so, Mufin will be releasing an iTunes plugin that will work with your own music collection and bring you recommendations and the ability to search the Mufin databse within the program itself.
I talked to Petar at Mufin last week, and just from our demo and my own tests, I was really impressed with both the technology and the structure of the site. The site has a similar layout to Songza, meaning it is easy to navigate. The MySpace app has an almost identical interface (a step-up from lots of MySpace apps), and adds the ability to share and embed in a user profile. Because the algorithm is based on song content and not an musical popularity, there is a tremendous opportunity for music lovers to discover new bands, and for new bands to get exposure to more users.
Use our link to try out the beta!