Abstract:

A novel multiple audio watermarking algorithm based on vector quantization technique and DCT transformation is proposed. To accomplish the purpose of copyright protection and content integrity authentication at the same time, the robust and fragile watermarks are both embedded into the carrier audio signal. The robust watermark is embedded into the middle frequency coefficients of DCT during the codeword labeled vector quantization process, and the fragile watermark is embedded into the high frequency DCT coefficients by using the index constrained method. Blind extraction can be achieved both for the robust and fragile watermarks. The simulation results indicate that the induced distortion by watermark embedding is small, and the robust watermark can resist to most of the common attacks. At the same time, the fragile watermark can accomplish the effective location to the random substitute falsify.

Key words: labeled vector quantization, multiple digital watermarking, audio signal