Multi-pitch Detection Using Complex Morlet Wavelet-Based Technique

In recent trends of music signal processing, different components of music signal like pitch, harmonics, tempo, rhythm, melody, beats, musical texture (whose combined effect give music a way to have good audibility) can be detected separately using various techniques. In music signal, pitch plays a very important role and hearing system of human is highly sensitive to pitch of any music signal. Therefore, a very high importance is given to pitch of music signal. Detection of multi-pitch makes it possible to synthesis the music signal also. This paper presents a novel complex Morlet wavelet-based technique for detecting multi-pitch of music signal where inharmonicity of musical instruments is taken under consideration. The proposed algorithm is compared in terms of accuracy, complexity, and root mean square error (RMSE) value with nonlinear least-squares algorithm and MUltiple SIgnal Classification algorithm to highlight its effectiveness in multi-pitch detection. The proposed scheme outperforms the baseline schemes in terms of these parameters.