Abstract: In order to protect copyright, an adaptive digital audio watermarking method based on Discrete Cosine Transform (DCT) and lifting wavelet transform was proposed, which was based on computing power of lifting wavelet and the largest auditory tolerance of DC coefficients after DCT transform. Firstly, the original audio signal was decomposed into low-frequency and high frequency component after lifting wavelet transform, then DCT was performed on the low-frequency component and watermarks were embedded into the DC coefficient sequences according to the largest DC auditory tolerance after DCT coefficients transform. Meanwhile, considering the balance between the robustness and the imperceptibility of the audio watermarking, the adaptive adjustments were used for embedded watermarking. The experimental results show that the proposed method not only has low computational complexity, but also is robust to malicious replace operation and common signal attacks like noise adding and lowpass filtering.

Key words: audio watermarking, lifting wavelet transform, Discrete Cosine Transform (DCT), adaptive

摘要： 为了达到版权保护的目的，利用提升小波变换计算速度快、离散余弦变换（DCT）后直流系数的听觉容差性强的特点，提出了一种在提升小波域进行DCT的自适应音频水印算法。原始音频信号经提升小波变换后分解为低频子带和高频子带,对其低频子带进行DCT，将水印序列嵌入到DC系数上。考虑到水印音频信号的不可感知性和鲁棒性之间的平衡问题，采用了水印序列自适应调整嵌入。实验结果表明，该水印算法计算复杂度低，且对噪声、 低通滤波等常见信号攻击及恶意替换操作均表现出很强的鲁棒性。

关键词: 音频水印, 提升小波, 离散余弦变换, 自适应