ISSN:1005-3026

EXTRACTION OF VARIOUS HIGH- AND LOW-LEVEL FEATURES FROM MUSIC AUDIO SAMPLES USING DIGITAL SIGNAL PROCESSING, CEPSTRAL ANALYSIS AND ONSET DETECTION

Atul Kumar Rai1, Sachi Gupta2,Gaurav Agarwal3

1Department of Computer Science &Engineering (DS), ABES Institute of Technology, Ghaziabad

2* Department of Information Technology, IMS Engineering College, Ghaziabad

3Department of Computer Science & Engineering, KIET Group of Institutions, Ghaziabad

atulrocks@gmail.com, shaurya13@gmail.com, , gaurav13shaurya@gmail.com

 

Abstract

Music is a form of art that has a significant role in the life of all humans. Music helps in the development of the mind, makes the listener feel better and is even a source of energy and motivation for many. Moreover, music having certain characteristics is found more suitable to listen to than others in certain cases. Inspired by the downfall in the quality of recommendations given by various music playback platforms due to the use of a user-metadata system, the aim here has been to understand the effect of music on the human psychology and realize how music emotionally affects a person. Research has been done in Neuromusicology and light has been shed on the various kinds of effects of music and the various ways that music can affect the brain. Various kinds of features were extracted from a number of audio signals (songs) and analysed on the basis of the knowledge gained.For music is not just lyrics that we interpret or beats that we tap to, music has a great impact on our brain, perception, memory, emotions and nervous system.

Keywords: Music Information Retrieval; Human Perception; Neuromusicology