The evolution of pop
We have been analyzing the historical dynamics of US and UK pop music from an evolutionary point of view. By this we mean understanding the patterns of change over time and the processes responsible for them. We have used MIR tools to extract harmonic, timbral and crude rhythmic features from around 30,000 songs that entered the US Billboard Hot 100 and UK Official Top 40 Charts, 1960-2010. We then discretize and cluster these features using a technique called Topic analysis borrowed from text mining. Using the resulting topic frequencies we estimate the musical diversity of the charts, and show that diversity is largely constant apart from a dip in the 1980s associated with the rise of drum machines and synthesizers. We also quantitatively identify the major revolutions in pop and identify 3 around 1964, 1983 and the early 1990s. Finally, using various time-series techniques we investigate the forces that shape the US charts. Surprisingly, we find evidence for strong long-term stabilizing equilibria.