Salford Analytics and Data Mining Conference 2012

Insight For Data Enthusiasts • San Diego, CA • May 24-25
Training May 21-23 • Welcome Reception May 23

You are here:Home»Conference»Sessions»Quantitative Assessment of Classical Music Style and Music Score Attributions

Quantitative Assessment of Classical Music Style and Music Score Attributions

Hector Bellmann (Griffith University) and
Ray Duplock (Queensland University of Technology)

We conducted an exploratory study to investigate if the main conceptual dimensions of tonal music style could be quantitatively assessed for classical music composed in the period of common practice (1700 to 1950). The identified main dimensions of music style can be construed as the basis for a scientific theory of tonal music style, and as a musicological tool for authorship studies for this period.

Using a database of 297 keyboard scores for 27 composers from Handel to Shostakovich, the study measured scalar features of Tonality - such as frequency of use of Scale Degrees, Modulations and Modes, Rhythm as well as the vector variables from Harmony and Melody, and Entropies were also measured. In total, 412 continuous variables were assessed by means of purposely written software.

Unsupervised CART and Random Forests (Salford Systems software), effectively reduced this large list of predictors to just a handful of useful variables, the application of which is effective to categorize music style and can be used to make accurate attributions about the authorship of new keyboard scores.

back to top