VC编写的音频频谱柱状图可视化显示界面

基于SPCE061的语音（乐音）辨识研究-SPCE061 based voice (music) Identification

对HMM基础知识的了解很有帮助，有发展历史，应用等-The basis of HMM knowledge helpful, there are the history of the development, application, etc.

语音识别源代码，可以对固定的某些语音实行识别，对于开发语音识别的朋友，具有很好的参考用途-Speech recognition source code, can be fixed to implement some voice recognition, speech recognition for the development of friends, an excellent reference

音频信号的短时分析工具箱，包含处理音频信号的常用命令文件等-Short-time audio signal analysis toolbox, including the common audio signal processing command files, etc.

MFCC 参数和 LPCC 参数是说话人识别中两种最常用的特征参数，研究了 MFCC 和 LPCC 参数提取的算法原理及差分倒谱参数的提取方法，采用MFCC、LPCC 及其一阶、二阶差分作为特征参数，通过 k 均值算法与三层BP 神经网络来进行说话人识别。实验结果表明，该方法可以有效提高识别率，同时也验证 MFCC 参数的鲁棒性优于 LPCC 参数

编译环境：vc++.net2003 名称：语音合成软件 作者：yxifu-Program environment: VC++.net2003 Name: program of speech synthesize Programer: yxifu

本程序的压缩编码是G.723的编解码程序，其的算法是基于ACELP的算法基础上的编码。用在语音通信中，例如手机中-the procedure coding is the G.723 codec procedure, the algorithm is based on ACELP algorithm based on the coding. Used in voice communications, such as mobile phones

PCM编解码源程序 很好的程序。谢谢大家进行参考。-PCM codec good source procedures. Thank you for your reference.

录音部分是参照vckbase的录音api代码,结合了一小段socket(TCP)就可以用来单方说话录音了,程序分两部分一部分是(录音机+网络发送代码),一部分是(接收数据+播放波形音乐代码),由于程序只是为了自己用,很多地方没有注意错误的识别,结构也比较乱,本程序当初最头疼的就是控制损耗内存,结果用了双缓存来存储波形数据来交替的存储/清除. -part of the recording light vckbase recording api code combination of a small socket (TCP) can be used for recording a unilateral words, part of the procedure is divided into two parts (recorders network code), part of the (receiving data waveform music broadcast code), as a result of procedures only for their own use. many places no attention to the wrong identification, the structure was chaotic, most of the first procedure is a headache for the control of memory loss, Results of the double-caching to store waveform data to the turn of the storage/removals.

这里主要我自己成果，是关于语音编码的，个人认为还可以，里面程序上传前经本人调试确认完全可用．有对多级矢量量化的程序（这个网上很少有），还有就是对MELP多帧参数的提取（这些网上也少见），还有就是统计子带清浊音概率的，还有一个我搜集的认为比较好的关于语音编码的英文文档．-Here the main results of my own, on the speech coding, and personally think that can also be inside the procedure before From my debugging can be used to confirm complete. There are of multi-stage vector quantization process (the online few), there is a multi-frame MELP parameters of the extraction (which are online also rare), there is statistical subband audio Qingzhuo probability, there is one I have collected think of a better speech coding on the document in English.