MP3音乐是目前广为流行的数字音乐形式，它能够在音质丢失很小的情况下把文件压缩到更小的程度。由于MP3具有高压缩比、音质优美、便于携带等优点，所以研究和实现MP3音频数据解码技术，也就具有极为重要的意义。
通过对MPEG-1 Audio 压缩编码标准和基于心理声学模型的音频数据压缩原理的研究，了解和掌握MP3音频数据格式的特点。本次设计是在Visual C++ 6.0的编程环境下完成的。基于对MP3音频数据文件的研究，实现了合成子带滤波器和逆改进型离散余弦变换等关键模块的设计。并且在此基础上，实现了对MP3音频文件的解码工作，完成了帧头信息的输出功能和PCM文件的存储功能。
论文介绍了MP3音频文件编解码的一些关键技术，并且根据这些技术，利用面向对象的编程方法实现了MP3音频文件的解码。
关键词：MP3解码，心理声学模型，逆改进型离散余弦变换，子带滤波器
ABSTRACT
MP3 music is one of the most popular digital music at present, which can compress files to a smaller extent while lose little of timbre. Because of the characteristics such as high compress ratio, good timbre, and portable, it is valuable to research and implement of the digital audio encoding technology of MP3.
From the investigating work of standard of MPEG-1 Audio compress coding and principles of digital audio compress based on psychologic acoustic model, MP3 audio characteristics is acquired. This thesis is study of MP3 digital audio files, and implement some key modules such as synthetical subband filter and Inverse Modified Discrete Cosine Transform under Visual C++ 6.0. Based on the above work, function of decoding MP3 files, output of information of frame head and storing of PCM files is implemented.
Key technologies of decoding MP3 audio file are introduced in this paper. Based on these technologies, object oriented method is used to implement the function of decoding MP3 audio files.
