tibaoluo

这是开源语音合成espeak的源代码，对研究TTS的同学很有帮助。(This source code is open source speech synthesizer espeak, helpful studies TTS classmates.)

提供了segmantal SNR及IS的计算代码，用于语音增强后评价用。(Provided segmantal SNR and IS calculation code for speech enhancement with post-evaluation.)

最经典的基于ICA实现的语音信号的采集、随机混合，再通过盲分离将混合后的语音信号分离(The most classical ICA-based speech signal acquisition, random mixing, and then the mixed speech signal is separated by blind separation.)

基于LMS算法和NLMS算法的自适应声回波消除器AEC，实现的主要功能是：在用户对电视发出的指令混杂了电视扬声器的声学回声的情况下，滤掉电视声音得到较为纯净的用户语音指令信号。(The main function of the adaptive filter based on the LMS algorithm and the NLMS algorithm is mixed: in the instruction issued by the user on the television TV speakers acoustic echo case, the TV sound was removed by filtration to obtain relatively pure user voice command signal.)

利用所给的中心削波,计算自相关/AMDF代码,选取一段语音,使用Praat预估其基音周期范围,合理设置预估的极值点距离等参数,计算并画出语音的基音频率曲线(By using the given center clipping, the autocorrelation /AMDF code is calculated, a speech is selected, the range of the pitch period is estimated by Praat, the estimated distance of the extreme point is set up reasonably, and the pitch frequency curve of the speech is calculated and drawn.)

基于ICA和PCA相结合实现的语音信号的采集、随机混合，再通过盲分离将混合后的语音信号分离(Speech signal acquisition and random mixing based on ICA and PCA, and then the mixed speech signal is separated by blind separation.)

c＃文件处理类库 C # file handling class library(C # File processing Class Library C # file handling class library)

实现对一个余弦信号的均匀量化和μ律非均匀量化，并将结果输出在图形界面上。(The uniform quantization and non uniform quantization of a cosine signal are realized and the results are output on the graphical interface.)

gmm模型下的语音情感识别系统，GMM只是一个数学模型，只是对数据形态的拟和，但是和你所看到的数据分布存在出入也是正常的，因为用EM估计GMM的那些参数时，一般假设我们所得到的数据是不完备的（也就是说假设我们看到的数据分布不是真正的分布，它在运算时把那部分丢失或者叫隐藏的数据“补”上了）(gmm model speech emotion recognition system, GMM is a mathematical model, but fitting the data form, but you can see there is access to the data distribution is normal, because those with the EM estimated GMM parameters, the general assumption we get the data is incomplete (that is, assuming we see the distribution of the data distribution is not real, it is part of the operation at the time that is lost or hidden data is called " fill" on the))

本程序介绍了一种改进的LMS算法-均一化的LMS算法，与传统的LMS算法比较收敛性更好，并由程序给出了比较图(This procedure introduced an improved LMS algorithm- homogenization of the LMS algorithm, with the traditional LMS algorithm convergence better, by the procedures given comparison)