This is a tool for auto align a corpus using HTK toolkit.I impelemented it for my research work,and I'd like to share it to help those who are not familiar with this. PS:It is for mandarine,if you want to use it for other languages,you need to change the dict file in general directory for your language.
HTK http://htk.eng.cam.ac.uk/
Perl5.8
GNU make
wav :wav files
script : perl script to do aligning
general : dictionary transcription and other general files
cofig: config files for HTK
step 1(preparation):
1 Edit line 2 in script/align.pl,specify your HTK PATH
2 Put your wav files in wav directory
3 Creat a file named words.mlf in general directory,this file should include the transcription of your utterances and in this format:
#!MLF!#
"/201.lab"
jiu
shi
xia
yu
ye
qu
.
"/202.lab"
wo
ma
shang
na
lai
.
you can generate this file with shell, perl or other script.
step 2(run):
execute make in project directory ,and then wait...
$ make