In this paper, we will propose an automatic music genre classification approach based on modulation spectral analysis of audio features. The audio features are Mel-frequency cepstral coefficients (MFCC) and octave-based spectral energy ratio (OSER). Modul