Using statistical analysis to predict pop hits

[bloodyjack]

Together in electric dreams

A computer program is changing the face of the music business by allowing record labels to predict a hit at the click of a mouse. Is this the death of pop as we know it, asks Jo Tatchell, or a new hope for unsigned bands everywhere?

Jo Tatchell
Monday January 17, 2005

Guardian

Martin and Ruth, aka Spike, the next big girl/boy duo (so they hope) add some synth and a new background vocal to the mix. He saves the song and she emails it to Polyphonic Human Media Interface who, within 24 hours, will tell them whether their song will be a hit. When the results arrive they hover over the 20in screen and click on the returned mail. There is a graph, showing a cluster of many dots, like a constellation, and somewhere in the cluster a red spot. The spot marks their song, not quite a bullseye, but still in the throng. "It's scored a seven," Ruth says, scanning down. "We're in. The record company will definitely meet us now." Their future suddenly looks a lot rosier.

Sounds unlikely? It shouldn't. Because, while no one's talking about it, it seems that the whole record industry is already using just this process. From unsigned acts dreaming in their garage, to multinationals such as Sony and Universal, everyone is clandestinely using a new and controversial technology to gain an edge on their competitors. And just as with athletes and performance-enhancing drugs, there is a remarkable reluctance to talk about it. But the secret is out: the record biz, once that bastion of wayward creative flair, is succumbing to the plain old-fashioned science of statistical analysis.

The magic ingredient set to revolutionise the pop industry is, simply, a piece of software that can "predict" the chance of a track being a hit or a miss. This computerised equivalent of the television programmer Juke Box Jury is known as Hit Song Science (HSS). It has been developed by a Spanish company, Polyphonic HMI, which used decades of experience developing artificial intelligence technology for the banking and telecoms industries to create a program that analysed the underlying mathematical patterns in music. It isolated and separated 20 aspects of song construction including melody, harmony, chord progression, beat, tempo and pitch and identifies and maps recurrent patterns in a song, before matching it against a database containing 30 years' worth of Billboard hit singles - 3.5m tunes in all. The program then accords the song a score, which registers, in effect, the likelihood of it being a chart success.

http://www.guardian.co.uk/arts/features/story/0,11710,1391951,00.html

[mark414]

1. if i was rich

the first thing i would do is start my own record label, where i would do the exact opposite of this bullshit

so much for allowing artists to grow and learn and expand

why the fuck does everything always have to be about money; why do people suck

[ladjf]

2. Unless atomic and sub-atomic particles are moving

randomly, all things could be predicted using the laws of physics as the basis of behavior.