资源列表
test-one
- 请录制语音信号集 (不能少于五个语音),然后根据最大似然估计方法,利用MATLAB设计一识别程序,并实现。-Please record a voice signal is set (not less than five voice), then maximum likelihood estimation method, using MATLAB to design a recognition program, and implementation.
SPEECH_ENHANCE.ppt
- speech processing technique
mp2enc
- 一个有关音频解码的详细代码,他支持各种采样和比特率,可以改动为自己需要的代码
studyandimplementonspeechenhance
- 语音增强方面的论文,来自优秀硕士博士论文库-Speech Enhancement papers, master' s doctoral dissertation from the excellent library
yuyinshibie
- 实现简单语音识别,标准的匹配文件是dataw.wav输入的语音生成data.wav。得出两者的匹配度-Simple speech recognition, matching the standard input of the voice file is dataw.wav generate data.wav. Drawn between the two matching
fft_test_show
- 这是一个用于对语音信号进行傅里叶变换,然后画图显示傅里叶变换的变换情况-This is a Fourier transform of the speech signal, and then drawing to show the Fourier transform of the transform
SR_speechrecognition
- 微软语音识别分两种模式:文本识别模式和命令识别模式.此两种模式的主要区别,主要在于识别过程中使用的匹配字典不同.前者使用的是通用字典,特点是内容多,覆盖的词汇量大,字典由sdk提供.适用于没有预定目标的随机听写之类的应用.同时因为词汇量大直接导致识别的精度降低,识别速度较慢.后者的字典需要 开发者自己编写,就是你们所说的xml文件.通过SR引擎读取xml实现特定词识别-Microsoft speech recognition points two modes: text recognition
Alize
- 说话人识别 语音识别的一部分 可以进行与文本无关的说话人识别-Speaker recognition can be a part of speech recognition and text-independent speaker recognition
HTK-3.4.tar
- htk因马尔科夫工具包 3.4用于连续语音识别 剑桥开发 开源-htk by Markov Toolkit 3.4 for Continuous Speech Recognition in Cambridge, the development of open source
fastica.rar
- 该算法为fastica的c++版本,有很好的分离效果。,Fastica the algorithm for the c++ version, there was good separation effect.
语音识别专辑五
- 语音识别专辑五- Speech recognition special edition five