EVRC

利用MATLAB来实现语音的分帧及语音端点检测(Using MATLAB to achieve voice segmentation and voice endpoint detection)

matlab程序，从语音中提取lpcc，可以修改维数，带详细中文解释，方便阅读。(matlab program, extracted from voice lpcc, you can modify the dimension, with a detailed explanation of the Chinese, and easy to read.)

实现多个人的说话人识别，基于gmm模型，用mfcc参数训练(voice conversion based on gmm)

HMM语音识别的matlab程序，举了1-9几个数字的例子，对语音识别的程序进行了验证(HMM Speech Recognition program， cited the example of figures 1-9, the speech recognition process was validated)

一个HMM的Matlab实现方法，可实现孤立词语音识别(Matlab implementation of a HMM methods, an isolated word speech recognition can be achieved)

优化小波阈值，对硬阈值和软阈值的改进，通常用于语音信号增强，去噪(Optimization of wavelet threshold, the hard threshold and soft threshold improvement, commonly used in speech signal enhancement and denoise)

熵值越大则每个符号包含的平均信息量越大。有研究发现，在有噪声的语音信号中，语音信号的熵和噪声信号的熵存在着较大的差异，对噪声信号来说在整个频带内分布相对平坦，熵值小，语音信号集中在某些特定频段内，熵值大。因此利用这个差异可以区分噪音段和语音段。(The greater the entropy is, the greater the average information of each symbol is. It is found that, in noisy speech signals, the entropy of speech signals and the entropy of noise signals are quite different. For noisy signals, the distribution is relatively flat in the whole frequency band, and the entropy value is small. The speech signal is concentrated in some specific frequency bands, and the entropy value is large. So the difference can be used to distinguish the noise segment and the speech segment.)

利用自相关或平均幅度差提取基音原理，帧长160，帧移80。求能量E、过零率Z和基音周期pitch；采用自相关法或平均幅度差法求取基音周期，基音周期范围（80~400hz）；画出能量、过零率和基音曲线。(Pitch principle is extracted by means of autocorrelation or mean amplitude difference)

通常语音信号在增强时会出现混响现象，演讲者为了消除背景混响，不得不频繁地偏转头部的方向，这样会造成脉冲响应的不断改变。我们结合盲解卷法和频谱消去法来提高逆滤波器的滤波效果。我们利用输入语音信号间的相关系数矩阵计算出稳定、精确的室内脉冲响应的逆滤波器，而这些输入信号无需测量室内的脉冲响应就能被观测到。逆滤波能够消除早期的反射，这些反射包含混响中的绝大部分能量。之后，用频谱消去法来抑制逆滤波后的信号的尾部混响。本方法在实际适应性方面的表现通过具体的实验进行了验证，结果表明盲解卷法和频谱消去法的结合相较于单独使用一种方法，能够提供一个更优越的演说环境。(Usually voice signal will be enhanced when the reverberation phenomenon, the speaker in order to eliminate the background reverberation, and had to frequently deflect the direction of the head, this will cause the impulse response of changing. We combine the blind deconvolution and spectral elimination method to improve the filtering effect of the inverse filter. We use the correlation coefficient matrix between the input speech signal to calculate the stable, accurate inverse filter of the indoor impulse response, and these input signals do not need to measure the room impulse response can be observed. Inverse filtering to eliminate early reflections, these reflections contains most of the energy in the reverberation. After the spectrum elimination method to suppress the tail reverberation of the signal after the inverse filtering. Performance in the actual adaptation of the method by specific experimental validation results show that the blind deconvolution and spectrum to eliminate )

（附带有文档，文档中有方案，有程序有结果，设计完后对一段语音进行处理）给定滤波器的规一化性能指标（参考指标,实际中依据每个同学所叠加噪声情况而定）通带截止频率wp=0.25*pi。 采用窗函数法设计低通、高通、带通型FIR滤波器，对叠加噪声前后的语音信号进行滤波处理，绘出滤波器的频域响应，绘出滤波后信号的时域波形和频谱，并对滤波前后的信号进行对比，分析信号的变化；在相同的性能指标下比较各方法的滤波效果，并从理论上进行分析((With the document, document, program, program results, the design after a voice) to set the filter rules of performance indicators (reference, practice, each student adding noise may be) passband cutoff frequency wp = 0.25* pi. Window function method to design low pass, high pass, band-pass FIR filter, filtering the voice signal before and after adding noise, draw the filter frequency response plot the time-domain waveform and spectrum of the filtered signal, and filtering the signal before and after comparison and analysis of signal changes the filtering effect of the various methods in the same performance indicators, and the theoretical analysis)