资源列表
Ec_test
- 改程序是本人在vs2008环境下编译通过的自适应滤波器的回声消除程序,运用了NLMS算法消除回声,效果还不错!-Change the program I compiled in vs2008 environment through the adaptive filter echo cancellation procedures, the use of the NLMS algorithm to eliminate echo, the results were pretty good!
Japanese-vowels
- 数据仓库与数据挖掘的应用,对9中来自不同人的语音的识别-Data warehouse and data mining applications, on the 9 people from different speech recognition
pujianfa
- 基于谱减法的简单消噪程序,但是在低信噪比情况下无法达到满意效果,适用于新手入门学习(A simple denoising program based on spectral subtraction)
freetts-1.2.1-tst
- speech synthesizer with JSAPI(Java Speech API). Source:http://sourceforge.net/projects/freetts
Wave
- 实时采集语音信号;显示语音信号的时域波形;重放采集的语音信号。-Real-time voice signal acquisition display time-domain waveform of the speech signal voice signal reproduction acquisition
TMMFCCdeCh
- 工程包括声音文件的读取、预处理、MFCC参数的提取、最后的的聚类函数,对于做语音识别的人帮助很大,已通过测试。 -The works will include the sound files to read, pre-treatment, MFCC parameters extracted, the final clustering function, speech recognition for people who do great help to the full source co
SpeechSignalProcess
- 语音信号处理方面的书籍。书籍的附录中有多个算法的C语言实现。超星图书需要使用超星浏览器打开-Speech signal processing books. Books have more than one algorithm in Appendix C language. Superstar Superstar books need to use the browser to open
MFCCdeC
- 工程包括声音文件的读取、预处理、MFCC参数的提取、最后的聚类函数,对于做语音识别的人帮助很大-The works will include the sound files to read, pre-treatment, MFCC parameters extracted, the final clustering function, to do speech recognition for the great help of people
HTK-Speech--recognition
- 运用HTK工具箱搭建了一个语音识别系统,可直接运行,参考价值大-HTK toolkit application to build a speech recognition system can be run directly reference the large value
test
- 自己编写的语音信号的采集,fft变换(两种),以及信噪比的计算!希望对大家有所帮助!-I have written the speech signal acquisition, fft transform (two kinds), and the calculation of signal to noise ratio! We want to help!
yate4.tar
- yate 开源sip服务器,支持sip和多种语音和视频编解码-yate opensipserver
jiyuneirongdeyinpinjiansuoyan
- 本文根据上述的研究,采用基于Mel倒谱系数特征的隐马尔可夫模型对音 乐进行分类。在音乐特征提取方面,以感知特征和Mel倒谱系数组成特征向量 在音乐分类方面,以隐马尔可夫模型作为分类器,对音乐进行聚类和分类。通过 有监督的学习方式进行聚类,分类时将测试样本归入似然值最大的类别,对同一 音频抽取若干样本,对样本识别结果采用投票法,判定该音频的音乐类别,使分 类的准确率得到进一步的提高。根据上述方法进行了仿真实验,并对实验结果进 行了分析。本文将音频数据分为5类,对4种分类器