基于量化技术的音频水印算法

针对数字音频作品版权保护的实际需要,提出一种利用二值图像作为版权标识的音频水印算法。利用伪随机序列对作为版权标识的二值图像进行置乱,同时对音频信号进行离散小波分解,然后根据置乱后的水印信息对所选取的小波系数进行量化处理,最后进行小波逆变换得到含水印信息的音频信号。实验表明,本算法对各种常见的水印攻击均具有较强的鲁棒性,可以有效解决数字音频作品版权保护等问题,具有一定的实用价值。

Audio Watermarking Based on Quantization Technology

This paper presents a novel audio watermarking algorithm by adopt binary image as a watermark. First, the binary image is disordered by the PN sequence. Then, audio watermark is constructed by quantizing the amplitude of DWT high frequency coefficients of digital audio signal. Final, watermarked audio signal is received by using IDWT of the quantized frequency coefficients. Experimental results show that watermark is inaudible and robust to many operate to digital signal, such as filter, re-sampling. This algorithm can resolve the problems of protect the digital audio signal's copyright so it has practical value.