Design and Realization of Music Recognition Based on Speech Recognition

Abstract

This paper propounds a new music recognition method which is based on singing splitting, MFCC characteristic parameter and BP Neural Network, and the origin of this idea comes from speech recognition. Firstly, fundamentals and key techniques of speech recognition are presented. In what flows is a brief introduction to a music recognition system that is designed and implemented by comparing speech recognition and music recognition. Then, we illustrate applications of BP neural network in music recognition. Finally, 10 English songs are chosen as experiment subjects and are inputted into the music recognition system. The experiment results show that this music recognition method which combines speech recognition can effectively recognize music signals.