FFT-Based Audio Watermarking Using SVD and CPT

Abstract

In this chapter, we present a FFT-based audio watermarking method using SVD and CPT [22, 24]. In our proposed method, watermark information is pre-processed first using a Gaussian map in order to improve the robustness and enhance the confidentiality. Then, the original audio is segmented into nonoverlapping frames. FFT is applied to each frame and low frequency FFT coefficients are selected. SVD is applied to the selected FFT coefficients of each frame represented in a matrix form. The highest two singular values of each frame are selected. The selected singular values are assumed as the components of polar coordinate system and are transformed into the components of Cartesian coordinate system. Watermark information is embedded into each of these Cartesian components using an embedding function. Experimental results demonstrate that our proposed watermarking method resists various attacks such as noise addition, cropping, resampling, requantization, and MP3 compression. Moreover, it outperforms the state-of-the-art watermarking algorithms in terms of imperceptibility, robustness, and data payload. The SNR and MOS values of the proposed method range from 36.19 to 37.32 dB and 4.95 to 5.0, respectively. This is in contrast to the above state-of-the-art methods whose SNR and MOS values range from only 12.87 to 29.50 dB and 2.93 to 4.7, respectively. Moreover, the BER of the proposed method ranges from 0 to 2.9297, whereas the BER of the state-of-the-art methods ranges from 0 to 51.73. Furthermore, the data payload of the proposed method is 689.56 bps which is relatively higher than that of the state-of-the-art methods.