Help in converting transcription files

[narayana]

Hi
I am using aurora data base. I want to setup Sphinx Trainer and Sphinx decoder for that database.

I have checked the transcription files in that database they are in HTK format as follows.

#!MLF!#
"/FAC_13A.lab"
sil
one
three
sil
.
"/FAC_1473533A.lab"
sil
one
four
seven
three
five
three
three
sil
.
"/FAC_172A.lab"
sil
one
seven
two
sil
.
"/FAC_1911446A.lab"
sil
one
nine
one
one
four
four
six
sil
.
Can any one help me out how to convert these transcription files to sphinx format.Is there any existing command or program for it.

please help me .As i am new to Sphinx i am facing lots of problems.

Thanks
Narayana

[Nickolay V. Shmyrev]

On Linux you can use something like:

cat a.mlf | grep -v MLF | tr -d '\n' | sed -e 's/.lab//g;s:"*/::g;s:"::g;' | tr '.' '\n' |
awk '{for(i=2;i<=NF;i++)printf("%s ", $i);printf ("(%s)\n", $1)}'

Otherwise simple script or C program can do that.

[narayana]

Thanks for the help.

There is a small correction in the script.

cat a.mlf | grep -v MLF | tr '\n' ' '| sed -e 's/.lab//g;s:"*/::g;s:"::g;' | tr '.' '\n' |
awk '{for(i=2;i<=NF;i++)printf("%s ", $i);printf ("(%s)\n", $1)}'

First tr command.

--Narayana.