Jason Freeman, a professor and musician from Georgia Tech, wrote a Java application which scans your iTunes library and creates your own “signature” out of small audio samples from your songs. From reading his technical docs, it’s more sophisticated than you’d think — the program finds your most played songs and then uses an FFT to merge spectrally similar audio, mixing down your audio smoother than a baby’s bottom.
It takes a little while to render and you have to sign a Java security license, but the results are brilliant! Our music collections can pretty accurately describe who we are and the ITSM does an amazing job at boiling me down to 47 seconds of glory. Something tells me that posting your own audio signature is the next big thing to hit
MySpace…
Ed: I know Jason from a while back, and this sounds like just the wacky development I’d expect from his past efforts — cool, Jas– er, Dr. Freeman!
See also Jason’s auralization of Gnutella searches, which creates total sonic chaos out of peer-to-peer files, algorithmic audio mixing, and even an interactive Net instrument you can play in your browser. (The others can be downloaded.) Thanks for the heads-up, Jordan! -PK