A Robust Speech Watermarking Based On Quantization Index Modulation And Double Discrete Cosine Transform

With the development of digital technology, there are many problems in voice protection and authorization. As a solution, digital watermarking can solve these problems. The principle of this method is to detect tampering by embedding the watermark in the original signal. In this paper, a speech watermarking method based on Quantization Index Modulation(QIM) and double Discrete Cosine Transform(DDCT) is proposed. Firstly, the original signal is segmented into frames and DCT is applied to all these frames. The AC coefficients which contain the details of the signal are chosen for DCT again to reduce the distortions caused by watermarking. The QIM technique is employed for embedding watermarks into the AC coefficients after the DDCT, and we select the most appropriate quantization step so that the modifications of signal will not cause serious distortion in speech quality. Each AC coefficient of one frame is embedded with one-bit watermark, and the majority voting is used to extract the watermark for better robustness. In terms of robustness and fragility, we evaluate the proposed approach. The experimental results show that the proposed method can not only provide satisfactory robustness and tamper resistance to signal processing, but also meet the requirements of non audible knowledge.