This is the G.727 standard documents

语音识别程序 可读出文本框中的文字,单词-a voise idetifying program,it can read string and word in word frame

语音识别原代码,VC++版!可以供研究研究!-Speech recognition source code, VC++ Version! Be available for research study!

验证过的最稳定的G.729A愿代码. 用于WINDOWS-Tested the most stable G.729A is willing to code. Used for WINDOWS

该程序是非特定人语音识别程序，原来在SPCE500A下运行，能识别并应答16个命令，存储空间为128K字节。 //我们将其进行了修改，使其能在SPCE061A下运行，由于存储空间为32K字，只能识别并应答9个命令。 //在这个例程当中，我们示范了如何使用语音识别引擎对8个已经经过训练的命令进行识别。 //第一个命令 "Snoopy"，可以作为使能命令，也就是说使用者只能在该命令之后发出其他命令。 //例如，用户只有在发出 "Snoopy"命令并且听到响应"Snoopy here"之后才能继续下达其他命令 //如"How are you", "I love you",等等。 //当"Snoopy"命令响应之后，程序将等待8秒钟，如果没有其他命令发出， //系统将跳回到第0组，并且再次等待触发命令"Snoopy"。当用户的命令无法被识别时， //程序将响应为"Execuse me?", "I can t hear you!"或 "Say again!"这时用户可以再一次下达语音命令。-the procedures were non-specific voice recognition procedure, the original run in SPCE500A can identify and Response 16 orders, storage space for 128K bytes.// We had to amending it so that it can run in SPCE061A, storage space for 32K words, recognize only nine responses and orders.// This routine, we have demonstrated how to use voice recognition engine has eight pairs of trained orders for identification.// The first order "Jack Champion", as it can make an order, users can only say in the order issued after other orders.// For exampl

隐马尔 柯夫模型( HM M ) 的传统 训练方法 ——Baum-W elch 算法只能 得到一个局 部最优 模型, 从而影 响 最终的识别率。对于 CHM M , 分段 K  平均方法来取得一个初始值可以解决这一问题, 但对 DHM M 却改进不大。 而 基于全 局搜索的 进化计 算的一 个重要 特点便 是可以 得到次 优解乃 至全 局最优 解。本文 把进 化计算 引入 到 DHM M 的训练中去, 提出一种改 进的进化训练方法, 实验结果表 明, 这种训 练方法初 具了全局搜 索和快速收 敛 的特点, 得到的模型优于传统方法和直接用进化计算所得的模型, 提高了系统的识别率。

converts text written by you to speech . require microsoft API.

E729国际标准的语音编码，非常有用的啊，大家一定要回啊-G729 annex E ITU speech vocoder, very usful for developer.

Peter Howell, Stevie Sackin, and Kazan Glenn Development of a Two-Stage Procedure for the Automatic Recognition of Dysfluencies in the Speech of Children Who Stutter: II. ANN Recognition of Repetitions and Prolongations With Supplied Word Segment Markers J Speech Lang Hear Res, Oct 1997 40: 1085 - 1096 -Peter Howell, Stevie Sackin, and Kazan Glenn Development of a Two-Stage Procedure for the Automatic Recognition of Dysfluencies in the Speech of Children Who Stutter: II. ANN Recognition of Repetitions and Prolongations With Supplied Word Segment Markers J Speech Lang Hear Res, Oct 1997 40: 1085- 1096

一个比较简单粗略的基于VQ矢量化模型的说话人识别程序-Rough a relatively simple model based on the VQ vector of speaker recognition procedures

语音信号预处理matlab程序，在7.0上运行通过，可以做实验参者代码。-Matlab voice signal pre-processing program, run through the 7.0, you can experiment with participation by the code.