Transparent high capacity audio watermarking in wavelet domain

This paper suggests a novel high capacity audio watermarking technique by using the high frequency band of the wavelet decomposition, for which the human auditory system is not very sensitive to variation. The high frequency band, DD, is divided into four parts and, then, in each step, a single sample from each part is selected. The sum of the selected wavelet samples is used for embedding the secret bits. The robustness and imperceptibility of the proposed scheme are adjustable by varying a tuning parameter. The outstanding feature of the suggested technique is the capacity to limit the changes of the wavelet samples, which guarantees the transparency of the embedding method. The experimental results show that the technique has very high capacity (about 2756 bps), without significant perceptual distortion (ODG in [-1, 0] and SNR about 42.1 dB) and provides robustness against MPEG compression (MP3).