fairfolio
Saved Projects

Genre Differentiation Using Timbral Analysis

CSEF · 2016 Mathematics & Computer Science

Ilene J. Hoffman · Rhoades (The) School, Encinitas, San Diego County

Overview

Objectives/Goals Music has always fascinated me - especially genres. What classifies a particular song into a certain genre or subgenre? Do these subgenres actually differ? I decided to try to determine whether there was any mathematical difference between often confused genres using MATLAB. I hypothesized that when subgenres were analyzed, there would be no clustering by genre, revealing that the classifications are subjective. Methods/Materials I analyzed over 200 songs from 10 different genres, divided into two groups. Group 1 (learning) was made up of rock, modern pop, 80's pop, EDM, blues, and classical music. Group 2 (testing) was made up of indie rock, alternative rock, emo, and pop-punk. I extracted the MFCCs of 250ms windows throughout entire songs, recording the Multivariate Likelihood Estimate mean vector and covariance matrix of these MFCC matrices into separate files. I then read these files into another script, which compiled all of these into struct arrays by genre. I then plotted the clustered struct arrays, using the first parameter of the .m vectors for the x coordinate and the second parameter for the y coordinate. This way, I was able to plot the clustering. Results My hypothesis that subgenres were classified subjectively rather than mathematically was supported by the resulting plots. All four subgenres in Group 2 were overlapping and indistinct since they were clustered very closely together. However, I was surprised to find the true roots of the umbrella genres in Group 1, finding rock and 80's pop completely intermingled, while classical music was completely separate, with blues somewhere in the middle. Conclusions/Discussion For further comparative analysis, my recommendations would be to implement timbral vector voting and convert my algorithms into a compiler language so that they will run faster. These algorithms could be put to use in softwares such as Pandora or iTunes.

Summary statement

I used MATLAB to analyze & plot 200 songs from various genres using the Mel frequency cepstrum to try to find the mathematical differences between musical subgenres.

Help received

I wrote the algorithm myself with minimal help from Mrs. Gontar, a senior designer at Via Telecom. I received a lecture on k-means clustering from Professor Gontar of Ben-Gurion University.

Competition history

  • CSEF 2016 Mathematics & Computer Science · Entry J1410

Resources

Related projects

Closest projects by meaning, across every fair and year in the corpus.

Browse more like this

Source: California Science & Engineering Fair public projects

Save projects to your library

Sign in with Google to keep track of projects you find interesting, organized into folders. An account also raises your daily allowance for “Has this been done?”, and lets you create a key for the MCP server with a much higher limit than anonymous use. Browsing stays public.

Continue with Google