Mufin, an automated music recommendation engine that actually works (most of the time), has released a new native media player in public beta. The player includes a powerful recommendation engine based on technology developed at the Fraunhofer Institute, allowing users to quickly generate playlists similar to any song they have in their library. The player is for Windows only at the moment, but Mac users can still try out the recommendation technology at Mufin.com, which launched to the public in November. You can see a full list of the new player’s features here
Mufin’s core technology is based on finding recommendations based on knowing a few songs that you like. Unlike Apple’s Genius feature, which creates recommendations based on aggregate data compiled from user listening habits, Mufin actually analyzes the sound file itself, ‘listening’ to 40 audio characteristics as it tries to recommend similar songs. In my testing the service has usually worked surprisingly well, though there are sometimes a few bizarre results. Oftentimes these apparent errors actually do share characteristics with the original song I used to gather recommendations, but they are very obscure and sometimes in a completely different genre (which is both the blessing and the curse of using such audio-based recommendation engines).
Mufin’s new player incorporates this recommendation technology, alongside a standard set of features that you’d expect from a modern audio player. Beyond creating the recommended song playlists, Mufin’s Player also brings a new twist to music management by allowing users to sort their songs by the way they sound (versus by song title or artist name). But aside from that I’m not sure if there’s a compelling reason to use it over iTunes at this point, especially since you can download a Mufin plugin to integrate the site’s audio recommendations into Apple’s music player (though the plugin is limited to creating playlists, and doesn’t allow users to sort by sound).