Please use this identifier to cite or link to this item: https://dl.ucsc.cmb.ac.lk/jspui/handle/123456789/4241
Title: Automated Accompaniment Generation for Vocal Input
Authors: Jayaweera, M. J. N.
Issue Date: 27-Jul-2021
Abstract: Music composition has become a popular area in the present. Melody of music is considered as the key element and forms the theme of the music. Therefore usually, a composer initially comes up with an idea for a melody, and then the music accompaniment is created in order to form up a piece of music or a song. However, creating appropriate accompaniment with proper chord progressions is definitely a complex task for non-musicians. Thus, automated accompaniment generation for vocal melodies can be very useful. A system which is capable of generating accompaniment for vocal melodies is presented in this thesis. The vocal input is taken as the input to the system via a microphone and it undergoes several modules in order to detect the proper chord progression for the melody. Initially, the vocal melody is divided into chunks based on the tempo in beats per minute selected by the user. A tone detection algorithm is used in order to detect the notes being sung and the associated energies of those notes. A dataset is prepared by analyzing popular songs that use simple chord progressions. The data set is trained by using a multi target learning and evaluation module which is used to predict the proper chord progression. Once the chord progression is identified, a simple accompaniment is assigned at each chord detected area. The experiments are carried out by obtaining vocal melodies from different users for a specific song. The results of experiments are evaluated by comparing the original simple chord progression of the selected song, with the chord progression generated by the system for each vocal melody. These findings are useful to develop an accurate system for accompaniment generation and to enhance the functionalities of existing systems related to the music industry.
URI: http://dl.ucsc.cmb.ac.lk/jspui/handle/123456789/4241
Appears in Collections:2018

Files in This Item:
File Description SizeFormat 
2015MCS036.pdf1.07 MBAdobe PDFView/Open


Items in UCSC Digital Library are protected by copyright, with all rights reserved, unless otherwise indicated.