Cepstrum-Based Audio Watermarking Algorithm against the A/D and D/A Attacks

Owing to the broad application value of audio watermarking in remote transmission and mobile extraction, it is very important to study watermarking approaches against A/D and D/A conversion. Based on wavelet decomposition and cepstrum technology, an audio watermarking algorithm is proposed where the low-frequency wavelet coefficients are chosen for cepstrum transforming. Then, the obtained cepstrum data shows stable for watermark hiding by the methods of edging data cutting and statistical mean modification. The synchronization is not required in the algorithm. Experimental results show that the proposed algorithm can resist on the attack of A/D and D/A successfully with features of large capacity and zero bit error. It will be helpful for the mobile audio in secure broadcast and management in the future.