HIT SONG SCIENCE

A Barcelona company says it has “developed an artificial intelligence application that could analyzed song and determine its potential to become a hit,” reports Neil Strauss in The New York Times. Really. Executives of the company, called Polyphonic HMI say that when they “loaded every song in the Billboard Top 30 over the last five years into the program, they found that “lo and behold, they all had something in common.”

Of course. They all sucked. They need a computer to determine this? Polyphonic sees things a bit differently, naturally. They think their killer app, which they call Hit Song Science, will help “raise the bar for music,” as suggested by Tracie Reed, vp of Polyphonic’s North American office. “We empathize with the people who buy a CD and only like two songs. Let’s get better music on more CD’s,” she says. Originally, the company thought their software could be used to “offer CD shoppers recommendations based on their favorite songs.” Now, however, they are promoting its use among record companies as a way to determine whether or not a song should be released. Several major labels, including Sony Music, RCA and Universal UK, “are either using it or considering the option.”

Jordan Berliant of Tenth Street Entertainment says he doubts a machine could measure the emotional content of a song, “in particular the lyrics.” Jaron Lanier, a computer scientist and musician, meanwhile dismissed “the science side of the application” as “sloppy.” He added: “As for the music side of things, I doubt pop music could get any worse, so using even a meaningless tool like this might result in some improvement.”

[Reveries]