资源列表
MATLABYUYIN
- MATLAB 语音信号分析与处理 武汉理工大学课程设计 -MATLAB MATLAB
microsoft_ASR
- 微软内部人员关于语言识别领域将在中国大陆的研究计划和展望。对从事语音识别领域的研究人员来说不可多得。-a microsoft company s overlook about the progress the speech recognition technology will make in next a few years.
Speechunderstandingsystem
- windows语音识别系统可以识别语音系统分析语音内容实现相应功能
massey_speech_project
- This the source code to the Massey School of Engineering Speech project.It has the Alice Artificial Intelligence engine included and only needs a Microsoft Agent character to work. The program will work without MSagent as well however. The program
Lecture02
- voice box contains mel function and gaussian func for voice recognition/ verification
music_synthesis
- 改程序实现了根据简朴音符合成音乐的功能,并且对不同乐器的声音进行傅里叶分析,可以合成具有不同音色的音乐-Reform process to achieve a synthesis of musical notes based on simple features, and the voices of different instruments to Fourier analysis, can be synthesized with different sound of music
voice_processor
- The purpose of this project is to create a different system for processing voice signals and from which we draw useful conclusions about these brands. More specifically, the system will first filter the voice signal with a pass bandwidth (chebyshe
PSOLA
- 基于C++的语音歌唱合成系统,核心转换为psola转换-C++-based singing voice synthesis system, the core is converted to psola conversion
shiyan
- 使用VC++ 基于MFC的小程序 能实现基本的语音识别 和语音合成功能,载入xml文件-VC++ based on the use of small MFC program to achieve the basic speech recognition and voice synthesis functions, loading xml document
LBG
- this is lbg program in v techniqu
asterisk-1.4.24.1.tar
- All Speech codec s source
HMM
- :为了使应力变异在顽健语音识别系统中能够达到较好的识别效果,研究了基于隐马 尔可夫模型(HMM)的自适应技术,提出了将最大后验概率(MAP)和最大似然回归方法(MLLR)用 于应力变异语音的自适应中。实验结果表明,与基本系统相比,两种方法均有效地提高系统识别 率。以SD为初始模型的最大后验概率方法在150个训练样本时识别效果最好,可以达到90.4% 。-: In order to stress variation in the robustness of speech recogni