Robust audio watermarking based non-uniform DCT

Targeted at security attack in digital audio watermarking, based non-uniform DCT audio watermarking algorithm was proposed. The watermark of algorithm was generated by chaos with key. The watermark was embedded in the mean of non-uniform DCT (NDCT) coefficients by quantifying the coefficients mean. NDCT frequency sampling positions were generated by chaos which was contralled by key and security of watermark system is better. Simulating experiments results demonstrated: the proposed algorithm was extremely robust to MP3 lossy compression, filting, re-sample, quantization etc signal process manipulations, the algorithm adapt to copyright protect.