基于压缩域的MP3音频数字水印算法

摘要

结合MP3编码特性,提出一种用于MP3版权保护的音频数字水印算法.该算法直接在压缩域中进行水印信息嵌入与提取,通过修改MP3比特流中的比例因子嵌入水印信息,无需MP3编解码过程.实验结果表明,该算法嵌入水印时不改变音频文件大小,提取水印时无需原始音频文件,其能在保证较好感知透明性的同时,具有较低的计算复杂度和较好的实时性.%This paper combines the characteristics of Moving Picture Experts Group Audio Layer 3(MP3) encoding, and presents a digital watermarking algorithm for MP3 audio files.It goes on watermark embedding and extraction directly in the compressed domain.The watermark is embedded in MP3 by slightly modifying the scale factors in bit stream, the method does not need encode the sound data into an MP3 file or decode the MP3 file.Experimental results show that the size of the audio file is not changed in the embedding process, and it does not need original audio file when watermark is extracted.The algorithm has advantages of higher real-time, lower computational complexity and excellent imperceptibility.