A digital audio watermarking algorithm based on block encoding

Firstly we introduce Hearing Masking Properties, the principle of cepstrum and backward and forward kernels, and then we propose a new scheme. The proposed scheme improve recovery rate at the expense of lowing the maximum communication rate. What's more, we compare the conventional system with our proposed system by adding white noise, filtering, re-sampling and also altering attenuation coefficient, the delay time, and the communication rate. We also do experiments on audio quality. Experiments demonstrate that the proposed scheme is suitable in the situation in which we need high the recovery rate.