yuputu

基于hmm的识别程序 识别数字0-9，识别率全对(speech recognition based on hmm number 0-9 all right)

一个对有噪声的音频信号进行自相关滤波的c程序，用的turbo c 3.0工具(One pair of noisy audio signal autocorrelation filtering c program using turbo c 3.0 Tools)

完整的基于HMM的语音识别系统，时间序列数据分析中的梅林变换工具，最终的权值矩阵就是滤波器的系数。( Complete HMM-based speech recognition system, Time series data analysis Mellin transform tool, The final weight matrix is ??the filter coefficient.)

语音增强算法综述，还可以，基本的谱减法，维纳滤波法都有(Overview of speech enhancement algorithm, can also basic spectral subtraction, Wiener filtering method has)

基于espeak的一个文本转语音的代码,可以把文本转成WAV文件(basic espaeck to translate txt to wav file)

语音信号的时域、频域分析,包括短时能量分析、短时平均过零率、自相关函数、短时平均幅度差函数等。(Time-domain speech signal, frequency domain analysis, including short-term energy analysis, the average short-term zero-crossing rate, autocorrelation function, such as short-time average magnitude difference function.)

语音唤醒就是在设备待机状态下，用户说出特定的语音指令（唤醒词）使设备进入工作状态或完成某一操作。设备为了实现语音唤醒功能，就需要设备实时监听，也就是实时的录音并分析有没有唤醒词来唤醒设备。评价一个语音唤醒系统的标准有三个：唤醒正确率、误报率、功耗。一般情况下，唤醒正确率越高 ，误报率也越高。好的系统就需要唤醒率高，误报率低，功耗低。(Speech wakeup is when the device is in standby mode, the user speaks a particular voice command (wake up word), so that the device enters the work state or completes an operation. In order to realize the function of speech wakeup, the device needs real-time monitoring, that is, real-time recording and analyzing whether or not wake words are used to wake up the equipment. There are three criteria for evaluating a voice wakeup system: wakeup accuracy, false positive rate, and power consumption. Generally, the higher the correct rate of arousal, the higher the false positive rate. A good system requires high wake-up rate, low false positive rate and low power consumption.)

基于倒谱短时部分反映了语音的声道特性，先用汉明窗取一帧语音，然后经变换得到语音倒谱，将倒谱短时部分取出，进行正交反变换后将得到声道的对数谱，即得到语音频谱的包络。将频谱包络和频谱画在一张图上，有很好的对比效果。获取的包络效果十分好。(Based Cepstral partly reflects the short channel characteristics of the speech, first take a Hamming window with a frame of speech, and speech cepstrum obtained by converting the cepstrum short segment out inverse orthogonal transform to obtain channel will of the spectrum, i.e. to obtain the envelope of the speech spectrum. The spectral envelope and spectral painted on a chart, there is a good contrast. Get the envelope effect is very good.)

语音识别中的两种特征提取方法lpcc和mfcc，还有一个是文本无关的识别算法dtw，另外还有一个是预处理消噪部分的。共享一下，这些都是我调试过的，好用。(Speech Recognition two feature extraction methods and mfcc lpcc. There is a text-independent recognition algorithm dtw, in addition to a pretreatment is part of the noise source. Sharing that those are off the debugging and ease of use.)

这是基于HMM的数字语音识别，实现了0到9 的识别，并带有GUI界面，可以进行录音和播放，预处理，识别结果准确，并附有音频(This is a digital speech recognition based on HMM identification of 0-9, with a GUI interface and recording box player, there are pretreatment, accurate identification, together with audio)