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