Mapping Chordal Progressions from the Magnitude Spectrum to Symbolic Notation

12 Pages Posted: 24 Jan 2023 Last revised: 13 Apr 2023

See all articles by Erin Ollewagen

Erin Ollewagen

University of the Witwatersrand

Ritesh Ajoodha

University of the Witwatersrand

Date Written: November 23, 2022

Abstract

This paper addresses the problem of automatic identification of chordal progressions from audio by presenting the application of Machine Learning techniques to map chordal progressions from the magnitude spectrum to symbolic notation. Five different machine learning techniques are used to map the chordal progressions, namely: Naive Bayes, Support Vector Machines, Random Forests, Logistic Regression Classifiers and Multilayer Perceptrons. The results showed that the Multilayer Perceptron achieved the highest accuracy of 99.48 percent, and Naive Bayes achieved the lowest accuracy of 50.15 percent. This paper shows the capability of real-time machine learning models to identify chordal progressions given noisy wav files.

Keywords: Algorithmic Music Composition, Signal Processing, Content-based, Machine Learning, Symbolic Notation

Suggested Citation

Ollewagen, Erin and Ajoodha, Ritesh, Mapping Chordal Progressions from the Magnitude Spectrum to Symbolic Notation (November 23, 2022). Available at SSRN: https://ssrn.com/abstract=4332856 or http://dx.doi.org/10.2139/ssrn.4332856

Erin Ollewagen (Contact Author)

University of the Witwatersrand

Ritesh Ajoodha

University of the Witwatersrand ( email )

Do you have a job opening that you would like to promote on SSRN?

Paper statistics

Downloads
13
Abstract Views
136
PlumX Metrics