Audio watermarking in transform domain based on singular value decomposition and quantization

Abstract

In this paper, we propose an audio watermarking method in transform domain based on singular value decomposition (SVD) and quantization for copyright protection of audio data. In our proposed method, initially the original audio is segmented into non-overlapping frames. Discrete wavelet transformation (DWT) is applied to each frame and detail coefficients are formulated. Discrete cosine transformation (DCT) is performed on the detail coefficients and the obtained DCT coefficients are reshaped. SVD is applied on the DCT coefficients and watermark information is then embedded into the highest singular value by quantization. Watermark is extracted by comparing the largest singular value of original and attacked watermarked DCT coefficients obtained from DWT sub bands of each audio frame. Experimental results indicate that the proposed watermarking method resists various attacks such as noise addition, cropping, re-sampling, re-quantization, and MP3 compression. Moreover, it outperforms conventional methods in terms of imperceptibility and robustness. Our proposed method achieves signal-to-noise ratio (SNR) values ranging from 38 to 40 dB, in contrast to conventional methods which achieve SNR values ranging from only 10 to 26 dB.