A perception-based audio watermarking scheme

In this paper, a perceptual audio watermarking scheme for MPEG scalable lossless audio that operates directly in the compressed data is proposed. The watermarking scheme is implemented in the compressed stream by modifying the quantized coefficients to satisfy the requirements of robustness, security and imperceptibility. The watermarking procedure directly exploits perceptual frequency masking of the human auditory system (HAS) to guarantee that the scheme is inaudible and robust. The watermark is constructed by selecting and quantizing the perceptually significant audio segments and coefficients. Extensive experimental results show that the high robustness and transparency of watermarked audio can be achieved simultaneously.