Audio Fingerprint based on Power Spectral Density and Hamming Distance Measure