fdfdesd
所属分类:语音合成
开发工具:matlab
文件大小:78KB
下载次数:16
上传日期:2008-10-08 20:50:39
上 传 者:
haipi
说明: 语音识别中试程序,包括几个测试用的WAV文件和TCL脚本程序,用来测试语音类别
(Speech Recognition test procedures, including several test WAV files and TCL scripts to test voice categories)
文件列表:
project1\dtw.c (10439, 2006-04-04)
project1\genvec.tcl (2157, 2006-04-04)
project1\input1.txt (13695, 2003-04-07)
project1\input2.txt (8452, 2003-04-07)
project1\input3.txt (20382, 2003-04-07)
project1\no_template.txt (7522, 2003-04-07)
project1\no_template.wav (16034, 2002-03-27)
project1\no_validation.txt (8495, 2003-04-16)
project1\no_validation.wav (17872, 2003-04-16)
project1\yes_template.txt (8554, 2003-04-07)
project1\yes_template.wav (17996, 2002-03-27)
project1\yes_validation.txt (9111, 2003-04-16)
project1\yes_validation.wav (19284, 2003-04-16)
project1 (0, 2007-06-13)
Notes:
1. The .txt files specify values for 20 features at every 10-msec frame in
ASCII format. The structure of the file is:
...
...
...
...
where = the number of frames in the file
= the number of features per frame
= the target value (not used for this project), = 1
= the first value of the first frame
= the last value of the first frame
= the last value of the last frame
2. at each frame, features are computed according to the following process:
1. pre-emphasis
2. Hamming window
3. log power spectrum
4. mel-scale warping to 20 frequency bands
5. scaling by 0.10
近期下载者:
相关文件:
收藏者: