Posts Tagged ‘mufin’

Many music recommendation engines have sprouted in the wake of Pandora’s success. None, however, have been as accurate and fool proof. Part of the reason is that Pandora relies of HUMANS to help the process along.

TechCrunch is high on Mufin, a new service that opens to the public today. Developed by the folks who invented and patented the ever-present MP3 file, the site aims to help you find music you like via “digital fingerprinting technology.” That means the system actually “listens” to music , analyzing dozens of song characteristics, including tempo and sound density. Read more

Mufin, the music search and recommendation engine that suggests music based on an algorithmic analysis of music, has launched its public beta today. Of course, the biggest point of interest is the lack of human editing when it comes to Mufin’s recommendations, setting it apart from the likes of Last.fm and Pandora. Even though many systems such as web search queries have been after the all mighty algorithm that works as well as (if not better than) a human, none have really come close to replacing the human element.

Granted, Mufin’s recommendations do work rather well. Perhaps this is because matching music can be more readily turned into a mathematical representation than say a search query for the best restaurants in Union Square. But Mufin seems to lack the human touch in its entirety. There’s no tweaking you can do for the recommendations Mufin offers, indicating how well of a match a song really is. And the social implications on Mufin’s main site are minimal, limited to a top playlist of songs that have been viewed, searched or played by others.
Read more