Watermarking for digital audio based on adaptive phase modulation

Abstract

This paper proposes a novel blind watermarking method for digital audio based on adaptive phase modulation. Audio signals are usually non-stationary, i.e., their own characteristics are time-variant. The features for watermarking have not been selected by combining the principle of variability, which affects the performance of the whole watermarking system. The proposed method embeds a watermark into an audio signal by adaptively modulating its phase with the embedded bit ‘0’ or ‘1’ using two all-pass filters. The frequency location of the filter pole-zero which characterizes the transfer function of the filter is adapted to distribution of signal power spectrum. The filter pole-zero locations are adapted in such a way that phase modulation causes the least distortion in watermarked signals to achieve the best sound quality. The experimental results show that the proposed method could embed inaudible watermark into various kinds of audio signals and correctly detect watermark without the aid of original signals. The proposed method has the ability to embed watermark into audio signals up to 150 bits per second with the bit error rate of less than 10 %.