The recent work “Measuring the Evolution of Contemporary Western Popular Music” [1] focuses on measuring large-scale trends in music over time. The authors analyzed acoustic features calculated from about 500,000 pieces of popular music spanning across 54 years. Their approach first subsamples the dataset so that the same number of tracks will be analyzed for each year. They then use various quantization techniques to represent each feature vector as a codeword. Rank-frequency distributions and large weighted networks are constructed out of these generated codewords. Based on the codeword distributions and various statistics of the networks, they draw conclusions about the characteristics of the music from each year. Unfortunately, a software implementation of the experiments in [1] is not publicly available. Furthermore, the use of network analysis is uncommon in the field of Music Information Retrieval. As a result, the reasoning behind the conclusions made about the evolution of popular music is not clear. Motivated by these issues, we recreated the experiments which involved measuring largescale changes in pitch, harmony, and melody. Our implementation gives further insight into the techniques used and allows for future comparative research to be carried out. The rest of this report is structured as follows: First, we describe the sampling and quantization steps used to create the dataset. In the second section, we summarize the statistics and techniques used to draw conclusions about musical evolution. In Section 3, we describe the specifics of our reimplementation of these experiments. Finally, we compare our results and give further insight into the techniques used.
[1]
Emilia Gómez,et al.
Transposing Chroma Representations to a Common Key
,
2008
.
[2]
K. Sneppen,et al.
Specificity and Stability in Topology of Protein Networks
,
2002,
Science.
[3]
Gábor Csárdi,et al.
The igraph software package for complex network research
,
2006
.
[4]
Eric Jones,et al.
SciPy: Open Source Scientific Tools for Python
,
2001
.
[5]
John D. Hunter,et al.
Matplotlib: A 2D Graphics Environment
,
2007,
Computing in Science & Engineering.
[6]
B. Granger.
Ipython: a System for Interactive Scientific Computing Python: an Open and General- Purpose Environment
,
2007
.
[7]
Thierry Bertin-Mahieux,et al.
The Million Song Dataset
,
2011,
ISMIR.
[8]
Marián Boguñá,et al.
Measuring the Evolution of Contemporary Western Popular Music
,
2012,
Scientific Reports.
[9]
Petri Toiviainen,et al.
MIDI toolbox : MATLAB tools for music research
,
2004
.
[10]
Brian E. Granger,et al.
IPython: A System for Interactive Scientific Computing
,
2007,
Computing in Science & Engineering.