Music Signal Recognition System for Music Sheet

Boey, Kai Lung (2017) Music Signal Recognition System for Music Sheet. Final Year Project (Bachelor), Tunku Abdul Rahman University College.

Text Boey Kai Lung.pdf Restricted to Registered users only Download (5MB)

Abstract

This project aims to perform music signal recognition and to identify the song. It generates a music sheet for the unknown songs. This project creates a system used to discover the song titles of the music from the surrounding environment. The system does this function by performing searching and matching from the music database. The main software used for the system is MATLAB. The library used to perform this recognition is MFCC. On the other hand, Vector Quantization is used for the matching and recognition of the results. These functions enable the system to perform extraction and recognition process as well as identification purpose. During the testing process, a song was played using a keyboard. The song was then instantly recorded by the MATLAB program. The software will then recognize the song title and display the musical notes. This software was specialized for the music instrumental systems of piano and musical keyboard. Electronic piano was used because it is more stable and easy to be handled on doing sound recognition project. The system achieved the accuracy of about 95%. This project is definitely useful for the musicians who wish to identify song titles using song audio signals.