adpcm_encoder

说明：  光谱预处理auto 光谱预处理center(Spectral pretreatment)

说明：  线性预测编码（LPC）是主要用于音频信号处理与语音处理中根据线性预测模型的信息用压缩形式表示数字语音信号谱包络（en:spectral envelope）的工具。它是最有效的语音分析技术之一，也是低位速下编码方法高质量语音最有用的方法之一，它能够提供非常精确的语音参数预测。(linear predictive coding)

可以对处理后的语音进行一个客观上的评判，提供客观评价因子。(It is possible to make an objective evaluation of the processed speech and provide an objective evaluation factor.)

语音压缩 自适应差分脉冲调制编码 即ADPCM 用matlab开发 (Adaptive differential pulse-modulated voice compression coding using matlab development that ADPCM)

假设一段语音信号，用其作为序列，编程实现LBG算法产生码书。(Assuming a speech signal, use it as a sequence, programming the LBG algorithm to generate the codebook.)

是FM调频立体声的调制解调的matlab仿真代码，实现了对双声道语音的调制和解调，非simulink方式，完完全全的matlab代码(This is for FM stereo radio signal modulation and de-modulation code. it is not for matlab simulink, it is completely matlab code, you can follow it and for understanding FM stereo signal generate very well.)

说明：  MeLPe Codec 1200 bps

语音信号的矢量量化，其中vector_quantization.m为读取音频文件，转换成语音帧，求lpc及lsf参数以及进行矢量量化的主程序，a_lsf_conversion.m为主程序调用的lpc转lsf参数的子程序，INLS1.wav为音频文件。(Vector quantization of speech signals, wherein vector_quantization.m to read audio file, converted into speech frames, seeking lpc and lsf parameters and vector quantization of the main program, a_lsf_conversion.m main program calls a subroutine parameter lsf lpc turn, INLS1.wav as an audio file.)

基于麦克风阵列用盖尔圆法进行信源数目估计。输入数据位麦克风采集的信号，阵元数目为10.(Based on microphone array，the number of sources is estimated by Gerschgrid method. Input data is obtained by microphone array and the number of array elements is 10.)

双通道语音增强算法，消除环境噪声。采用归一化自适应方法，噪噪声抵消10dB，语音保持较好可懂度。(The dual-channel speech enhancement algorithm to eliminate ambient noise. Using normalized adaptive noise noise offset by 10dB, and the voice to maintain good intelligibility.)