New psychoacoustic models for wavelet based audio watermarking

The new approach to apply psychoacoustic models to the wavelet based audio digital watermarking is proposed in this paper. In the scheme, the additive watermarking technique is used to embed a unique pseudo-random sequence, considered as a watermark, into the transformed domain of audio signal. The watermark strength is properly adjusted based on weighting factors derived from the proposed psychoacoustic models. The results show that at the equivalent quality of the watermarked audio, judged by the human hearing system, the robustness of the embedded watermark was increased by up to 97.1% and 21.3%, compared to the results obtained from the scheme with non-psychoacoustic model and the previous psychoacoustic model, respectively.