ETSI发布的用于研究分布式语音识别的前端降噪算法

应用背景他是一个程序 ；VoIP ..是一个很好的项目 ；VoIP关键技术VoIP  ；（H323）源，G.723.1 G729 G728 G711音频算法的实现

对数运算实现是32位的运算包含有指数和对数的求法 -LOG

一种HMM训练的源程序，调试可以通过，说明比较详细-A HMM training source, debugging can explain in more detail

对不同特征进行融合，得到更好的语音特征，识别率高。-Integration of different features, better voice features, a high recognition rate.

Dialogic源码，包括FAX，TAPI，Voice。-Dialogic source, including FAX, TAPI, Voice.

Levison-Durbin 语音信号处理中的线性预测编码LPC 理论、格型滤波器以及求解现行预 测方程的算法，可以实现对语音信号重要元素的分析、合成甚至识别。 基于现有的实验平台，我们可以利用 Matlab 函数来获得几个固定语音元素（如元音） 的模型系数，LPC 得到的系数组成 IIR 滤波器。利用冲击脉冲 序列作为输入，我们就可以得到原来的语音。这是一种简单的语音合成功能。-Levison-Durbin speech signal processing in linear predictive coding LPC theory, lattice filters, as well as the current prediction equation solving algorithm, can achieve an important element of the speech signal analysis, synthesis or recognition. Based on the existing experimental platform, we can use Matlab function to obtain the number of fixed-voice elements (such as vowels) model coefficients, LPC coefficients are the composition of IIR filters. Shock pulse sequence used as input, we can get the original voice. This is a simple voice synthesis.

一个用vs2005开发的、简单的TTS文本转换语音程序，是.net版的TTS，有助于语音合成初学者。-One with the vs2005 developed a simple TTS Text speech translation procedure is. Net version of the TTS, speech synthesis will help beginners.

拷贝文件的java源代码 -copy of java source code

采用G.729的语言实时通信DLL(含测试源代码)-G.729 language used real-time communications DLL (including test source code)

ITU-T G.723语音压缩编解码算法源码和测试语音，内有标准附件和使用说明。-ITU-T G.723 voice compression codec algorithm source code and test voice, within the standard accessories and use.