本程序支持实时语音的控制,处理

模式识别k-means法与k-l变换做的特征提取-K-means pattern recognition method and kl transform to do the feature extraction

Microsoft Speech SDK提供关于语音（Speech）处理的一套应用程序编程接口SAPI（Speech Application Programming Interface）。SAPI提供了实现文字-语音转换（Text-to-Speech）和语音识别（Speech Recognition）程序的基本函数，大大简化了语音编程的难度，降低了语音编程的工作量。

FIR滤波器设计中往往是线性相位，但不是最小相位，所以它的零点有部分在单位圆外。把单位圆外的零点映射到单位圆内，构成最小相位滤波器。-FIR filter design is often the linear phase, but it is not minimum phase, so it has some of the zeros outside the unit circle. To zero outside the unit circle is mapped to the unit circle, the minimum-phase filter constituted.

本文系统的介绍了语音识别的技术及其实现方法-This article introduces the technic and implement method of speech recognise

lpc10-15为美军2400bps语音压缩标准的C语音源代码。-lpc10-15 for the U.S. 2400 voice compression standard C source code voice.

基于GMM的概率神经网络PNN具有良好的泛化能力，快速的学习能力，易于在线更新,并具有统计学的贝叶斯估计理论基础，已成为一种解决像说话人识别、文字识别、医疗图像识别、卫星云图识别等许多实际困难分类问题的很有效的工具。而且PNN不但具有GMM的大部分优点，还具有许多GMM没有的优点，如强鲁棒性，需要更少的训练语料，可以和其他网络其他理论无缝整合等。-GMM based probabilistic neural network PNN good generalization ability, the ability to learn fast, easy online updates, and with the Bayesian statistical theory based on estimates, and has become a solution as speaker recognition, text recognition, medical image recognition, satellite images and other real recognition when difficulties classification of very effective tool. But GMM PNN is not only the most advantages, but also has many advantages GMM not as strong robustness, require less training corpus, and other networks to other theories, such as seamless integration.

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.

一种噪声环境下连续语音识别的快速端点检测算法，格式ｖｉｐ-a noisy environment for the rapid identification voice endpoint detection algorithm format vip

这是一个语音程序-This a procedural voice!

基于最大信噪比的语音分离程序，有需要的朋友可以借鉴-Based on the maximum signal to noise ratio of the speech separation process, there is a need of a friend can learn from