Pre-AI Musical Style Analysis Via Their Spectral Distributions

Abstract

In this paper, we will revisit a pre-AI method of identifying musical genres. To do that, the spectrograms of a number of fragments of melodies belonging to various genres will be generated, with the aid of the spectrum analyser included in the Digital Signal Processing Toolbox from Simulink, Matlab, as well as the spectrum analyser of the audio editing and analysis program Audacity. Then, there will be an attempt at identifying the defining spectral features for each of the analysed genres, with the aim of identifying said genres via Fourier analysis. Included in this paper will be a revision of key theoretical concepts pertaining to the Fourier analysis of analog signals as well as digital, as well as an enumeration of the relevant musical theory concepts.