Musical Genre Classification Using Support Vector Machines and Convolutional Neural Networks
Tharun R Ragav, B. S. Divya, A. Kavitha · 2023
Music is an escape to another dimension where everything remains infinitely better. It plays a vital role towards the establishment of cultural characteristics of a human society across the globe and the power of music can touch human lives in unimaginable ways. The creation of music is performed by composing the musical notes, which are the smallest and most basic elements of music. The significance of music has made it crucial to understand how musical notation works. Music signal processing is related to speech signal processing, because music can be closely correlated with speech as a lot of techniques used for the former can also be applied to the latter, with appropriate results. The genres of various audio files have been grouped into various classes through multiple training models. Starting with the extraction of spectrograms to its sub-windowing which would help us gain information more accurately. The training accuracy of MFCC images and spectrograms has been verified using Support Vector Machines (SVMs) and Transfer Learning Models in this process.