Adaptive spread spectrum audio watermarking based on data redundancy analysis

摘要

In this paper an adaptive spread spectrum method is proposed for effective embedding and detection of audio watermarking. The method introduces audio data redundancy analysis at the watermark receiver to achieve enhanced extraction performance. In particular, audio signals tend to have large dynamic range, and data redundancy exists between neighboring audio samples. By removing data redundancy of the host audio signal, the host interference is suppressed. Furthermore, we introduce a double-sequence modulation technique at the watermark embedder. The modulation coefficients are adaptively selected to reduce the negative effect of perceptual shaping. The advantages of our method are verified through experiments.