Audio fingerprinting and music recognition system to aid in recognition and verification of local music in Uganda

Abstract

The objective of this project was to develop a song recognition system that can be used for accurate music recognition using various features of a song signal. This system works by first recording a sample of a playing song using the microphone of the device on which the system is loaded, the system then processes the sample and returns the meta-data such as the name and artist corresponding to the sample. The systems operation can be broken down into two parts that is first fingerprinting the audio files and storing the hashes and meta-data to the database and then recognizing unknown sample audio from the microphone. In this report, the working of the system will be explained in detail and also results from the testing of the system explained and evaluated.