A chord recognition method based on joint feature of MFCC and PCP

Combined with music and signal process theory, the paper proposes a new chord recognition approach which utilizes the MFCC(Mel Frequency Cepstral Coefficent) reflecting auditory perception properties jointly with the traditional PCP(Pitch Class Profile) as the combined feature and SRC(Sparse Representation classification) to improve the recognition rate. The experimental results show that the recognition rata of the new method is much better in average than that of the traditional methods.