Many years ago I had the opportunity to listen to famed music producer and ambient-music pioneer, Brian Eno, speak about the problems wrought by the emergence of music synthesizers. To paraphrase his thoughts, with an infinite variety of tonal colors available to musicians, how does one make a choice? Yes, there’s great power in an unlimited palette, but such a vast tonal landscape can be overwhelming.
I might suggest that with the advent of music distribution over the Web (and the accompanying breakdown of the corporate promotion/distribution scheme), listeners are now faced with a similar dilemma: As all of the world’s music becomes available to you, how do you choose what to listen to? Sure, the labels will continue to act as a filter by pushing the artists in their stables through radio, video, and online services, but their power to impose their will is likely to be diminished as more independent voices are brought to the fore.
Online music services such as the iTunes Music Store have attempted to address this conundrum through the “word of mouth” model—an increasingly popular scheme that posits that if your taste runs to Artist A, you’ll likely also enjoy the work of Artist B. You can see this reflected in the Listeners Also Bought links on album and artists pages as well as in your Shopping Cart page if you’ve elected to purchase music that way. Amazon.com customers see something similar in Amazon’s Customers Interested in This Item May Also be Interested in links.
But such filters are fairly broad—a couple of ill-advised purchases and before you know it, the iTunes Music Store tells you that you’re a perfect candidate for Doris Day’s rich catalog of chirpy hits.
Other outfits are attempting to offer greater personalization to these recommendations. One such entity is Soundflavor.com. I saw the service in action last week and it’s intriguing.
Getting the Flavor
The basic idea works this way: You create a playlist and publish it to the service. Recommendations for other music are created based on that playlist. For example, if you create a playlist that favors 70s funk, you’re likely to receive a lot of recommendations for tracks by James Brown, Tower of Power, and Parliament. You can preview most recommendations by clicking on a speaker icon, which then streams the tune to you from Streamwaves.com (Windows Media Player required). You can also get a list of suggestions based on recommended songs—effectively burrowing down into layers of recommended tracks. You can also rate recommended songs or the songs you’ve placed in your playlist. In addition, you’ll find links for buying the song from the iTunes Music Store or the CD from Amazon.com.
Soundflavor bases recommendations on both track analysis and user recommendations. Using algorithms to analyze music for certain factors—density of guitar, tempo, and style, for example—and the opinions of the member community, the service attempts to create playlists of music custom tailored to your tastes. You needn’t rely solely on the playlists the service generates, however. You can also access “trusted members” (other members of the community whose tastes fall in line with yours) recommendations.
Soundflavor is currently free and in beta. It’s an interesting way to explore new music and worth a long look.
Joining the Mob
Something Soundflavor lacks is the ability to upload a list of everything in your music library in order to compare it to others’ libraries. Thankfully, this service is offered by Musicmobs —a free service that compares the contents of your iTunes music library (and the popularity of songs in your library as determined by play count) to that of other registered users.
You just upload a file that contains a list of tracks in your iTunes library (using iTunes’ Export Library command to generate the proper XML file) and then click the Similar Users link to view a list of users whose libraries have similarities to yours. Click on a similar user’s link and the resulting window shows a list of the artists in that user’s library along with a number that reflects the number of times an artist’s tracks have been played by the user. Those artists that correspond to artists in your library are highlighted in pink. Musicmobs also offers lists of the service’s most popular artists as well as top listeners and most popular users.
A Good Start
They say there’s no accounting for taste and that’s fairly evident when you use each of these services. They do a reasonable job of pointing you toward music that you might not have otherwise sought out, but some of their recommendations are sure to fall flat. On the other hand, even in their infancy, they do a better job of targeting your tastes than the folks programming your local radio station or cycling videos through MTV. Check ‘em out and use the Comment link below to tell me what you think.