mel_endpointdetection

科大讯飞语音识别最新Demo,国内优秀的语音识别平台(new demo)

VAD去噪算法，可以有效的过滤静音和无效的语音，同时消除了一些背景噪声。(VAD denoising algorithm, can effectively filter invalid mute and voice, while eliminating some background noise.)

基于MFCC的GMM的说话人识别，是很好的语音处理程序(MFCC of the GMM based speaker recognition, speech processing is a very good program)

语音信号分帧后，计算每帧信号的短时平均能量(After the speech signal is framed, the short-term average energy of each frame signal is calculated)

采用MFCC方法提取训练集样本或测试集样本语音特征参数的MATLAB程序，仿真时需先将voicebox语音处理工具箱添加到MATLAB软件中(The MFCC method is used to extract the MATLAB program of the training set sample or the test set sample speech feature parameters. The voicebox voice processing toolbox is added to the MATLAB software first.)

TD-PSOLA MATLAB编写语音合成模拟程序(The simulation program of combined speech signal beased on TD-PSOLA MATLAB method)

c++语言开发，用于LPCC特征参数提取的，在语言识别系统中应用广泛。(c language, LPCC Features for feature extraction, in language recognition system widely used.)

基于svm分类器的汉语声调识别，实现了对声调四声的识别，纯净语音识别率100 ，噪声环境下识别率90 (The svm classifier based Chinese tone recognition, the four tones of tone recognition, the pure voice recognition rate of 100 , 90 recognition rate in noisy environments)

speech file to be analyzed Format : binary file 16 bit-samples 240 samples per frame (speech file to be analyzed Format : 16 bit binary file-samples of 240 samples per fram e)

　提出了一种新的连续型隐马尔可夫模型(HMM ) 的概率密度函数, 并导出了一系列的参 数寻优迭代公式,(A new Continuous Hidden Markov Model (HMM) of the probability density function, and derived a series of parameters optimization iteration formula,)