Imperial News

Volunteers needed to chart fifty years of ch-ch-changes in pop music

by Simon Levey

Music lovers are invited to take part in an online experiment to explore the ways that popular music has evolved between 1960 and the present day.

Music lovers are invited to take part in an online experiment to explore the ways that popular music has evolved between 1960 and the present day.

The Descent of Pop experiment is a computer program that borrows theories from Darwinian evolution to evaluate musical characteristics like tempo and pitch.

Does the Top 50 become more or less diverse with time? It's an obvious one for an evolutionary biologist to ask.

– Professor Armand Leroi

Descent of Pop researcher from Imperial's Department of Life Sciences

The program has analysed music from the last 50 years of chart music taken from the Billboard Hot 100 listings and over 15,000 digital MP3 tracks from music sharing site Last.fm.

Now the scientists running the experiment need real people to listen to and compare similarities in pairs of tracks and rate the quality of the audio, tasks that are hard for a computer to do reliably.

Descent of Pop researcher, Professor Armand Leroi from Imperial's Department of Life Sciences, said "One of the questions we've been asking is this. Does the Top 50 become more or less diverse with time?  It's an obvious one for an evolutionary biologist to ask.  Paleontologists have charted the diversity of animal forms through time: Is that how it was for the Charts?  Did pop begin in uniformity and then explode into diversity? Or was it the other way around?  When was pop’s Golden Age?"

"We need at least 30,000 voters to listen to the sample music and participate in the research project to understand how chart music has evolved from 1960 to 2010," said project leader, Dr Matthias Mauch from Queen Mary’s Centre for Digital Music.

"The data can help us answer lots of interesting questions, such as have musical styles changed more swiftly now than they did in the 1970s, or do music buyers in one year have an influence of what musicians will compose in the next year?. But before we can start, we need clean, reliable data. First, we want to exclude all bad recordings to ensure that what we observe are genuine musical characteristics. And because humans still outdo computers at spotting a bad recording, we need help from a lot of listeners."

Building on previous research, including the Darwintunes project that analysed artificial music, the team are now digging deep into the rich data set to unravel what drives innovation and imitation in popular music.

To participate in the Descent of Pop study:  http://descentofpop.eecs.qmul.