Robust Audio Watermarking Based on MDCT Coefficients

Multimedia contents can easily be replicated and spread in the Internet. For the verification of copyright of digital audio, digital watermark may be embedded in the audio. Compared with traditional watermarking technologies, zero-watermarking technology does not result in quality degradation, and thus can avoid the conflict between imperceptibility and robustness. In this paper, we calculate the bi-level characteristic sequence of audio signal by using the MDCT coefficients. Coupled with bi-level watermark image and the secrete key provided by an impartial third party, a private key is generated for subsequent intellectual property right verification of the copyright owner. To evaluate the robustness of the proposed zero-watermarking scheme, this study measured the accuracy of reconstructed watermark from audio signal after watermark attacks by various signal processing methods. The experimental results indicate that the proposed zero-watermarking method has better robustness as compared with methods proposed in literature. The proposed scheme is efficient and can be applied in the protection of intellectual property rights of audio signals.