利用小波包变换实现音频压缩算法

构造了一个音频编码系统，用ＦＶ－ＭＬＴ（Ｆｒｅｑｕｅｎｃｙ Ｖａｒｙｉｎｇ Ｍｏｄｅｌａｔｅｄ Ｌａｐｐｅｄ Ｔｒａｎｓｆｏｒｍｓ）＾［１］实现小波包的变换，进行对输入信号的时频变换。并利用时频变换输出的ＦＶ－ＭＬＴ系数计算蔗蔽门限，提出了基于小波变换的心理声学模型算法。对整个编码器进行了软件仿真，并与ＭＰＥＧ的标准编解码器进行了音质与压缩率方面的比较。在音质接近的情况下，该编码器的平均码率比Ｘｉ

An Audio Compression Algorithm Using a Wavelet Packet Transform

In this paper an audio coding system is implemented. FV-MLT(Frequency Vary- ing Modulated Lapped Transforms)1],a kind of wavelet packet transforms, is used to realize the time-frequency transform of the input PCM signal. Meanwhile, the coefficients of the FV-MLT are also used to calculate the masking threshold, and a new computing algorithm of the psychoacous- tic model based on the wavelet packet transform is developed. The whole audio coding system is compared with the standard encoder of MPEG on audio quality and compression ratio through software simulation. Given equivalent quality,the average bit rate of the new system is about 10% to 20% lower than that of Xing AudioCatalyst 2.1 encoder, which is based on the MPEG stan- dard.