摘要
充分结合MP3的编码特性,提出一种MP3压缩域的音频水印算法来对MP3音乐进行版权保护.在MP3压缩过程中,通过寻求m个能量较大的比例因子频带,然后分别修改每个频带中最大修正的离散余弦变换(modifieddiscrete cosine transform,MDCT)系数的符号来进行水印的嵌入.在MP3解码过程中对水印信号进行提取,提取不需要原始载体.嵌入过程和编码过程同步完成,而提取过程和解码过程亦是同步完成.实验证明,该算法具有很好的听觉透明性和很强的鲁棒性,可以有效地抵抗攻击者对比特流解码后再重新压缩的攻击.
This paper proposes a new scheme of audio watermarking based on MP3 compression for protecting copyright of MP3 music. During MP3 encoding, energy of each scale band is computed to get m scale bands having relatively high energy. The largest modified discrete cosine transform （MDCT） coefficient is found in each scale band. The watermark bits are embedded by modifying the sign of these coefficients. The watermark can be extracted in the decoding process without using the original audio signal. The presence of watermark is undetectable by human hearing, and the scheme is robust against attack which is performed by decoding the watermarked MP3 bit stream and then re-coding it.
出处
《上海大学学报（自然科学版）》
CAS
CSCD
北大核心
2008年第4期331-335,共5页
Journal of Shanghai University:Natural Science Edition
基金
国家自然科学基金资助项目(60502039)
关键词
MP3
音频水印
版权保护
比例因子频带
修正的离散余弦变换系数
MP3
audio watermarking
copyright protection
scale band
modified discrete cosine transform （MDCT） coefficient
