One of the new features in the iPhone and iPod Touch 2.1 firmware is the ability to create Genius Playlists on the fly, and right on the device. iTunes 8 introduced the concept of Genius Playlists, which like Pandora and Last.fm, use algorithms based on audio profiling, metadata matching, and (anonymously acquired) end user tastes, crunched in the Apple cloud. Basically, they take a song you like, what you typically like with it (what playlists you include it on), what others who like it also like (what playlists they include it on), and what music mathematically (tempo, tone, etc.) fits in with it. Then they try to predict other songs you might also like — in this case already on your iPhone — and whip you up a near-instant playlist on-demand.
How good is it? That depends entirely on how much music you have on your iPhone (the pool from which it can draw), how closely the metadata matches what’s in iTunes (or it won’t recognize your music — try tweaking the fields if you have trouble), and how many other users have contributed their data to the cloud (because the engine will supposedly get better and better the more information it’s fed).
Leave a comment