VoIP Mailing List Archives
Mailing list archives for the VoIP community |
|
View previous topic :: View next topic |
Author |
Message |
helmut.kuper at ewetel.de Guest
|
Posted: Fri Jul 31, 2009 8:28 am Post subject: [Freeswitch-users] pocketsphinx |
|
|
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Hello,
I spent a few days working on my problem. I played around with
voxforge's data, read some of the train docs of CMU, and debugged the
sphinxbase, pocketsphinx and mod_pocketsphinx.
Results:
- -I found a way to use the voxforge data as training data for creation of
a german language corpus.
- -I enabled the logging of pocketsphinx to stderr (Dirty, but easy way to
see what went wrong when FS loads grammar, mdef, etc and simply stopped.
Very helpful!
- -I had to add a "dictcase" parameter to pocketsphinx.conf.xml resp
mod_pocketsphinx.c to allow case sensitive dictionaries (like the german
dictionary from voxforge).
FS starts up with german language model and detects the words as
expected. But it's not so reliable as I want to ... I guess this is
caused by the very small amount of training audio data. I used 4000 of
19000 audio files provided by voxforge due to the reason that voyforge's
training fileid-list contains only 4000 files ... I have to create new
fileid-list and transcription-lists containing all audio I have
downloaded from voxforge.
Quite complex the whole thing ...
regards
helmut
On 10.07.2009 14:53, Helmut Kuper wrote:
Quote: | Hi,
I try to change pocketsphinx's grammar from default (english) to german.
I found this archive
(http://www.repository.voxforge1.org/downloads/de/Trunk/AcousticModels/), which
contains similar files like those which can be found in
grammar/model/communicator directory.
Unfortunately FS crashed without writing a core file nor logfile enries
as soon as as pizza demo trys to detect speech.
Any Ideas? Maybe someone has already working grammar/model files for
german language?
regards
helmut
| -----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.7 (MingW32)
iD8DBQFKcu8z4tZeNddg3dwRAiVKAKCJdAvZyYXLClWp+kQO4MhsPVzbSQCgoX/T
7NosCujHwpkCXHzxkEY/N3M=
=6Bu8
-----END PGP SIGNATURE-----
_______________________________________________
FreeSWITCH-users mailing list
FreeSWITCH-users@lists.freeswitch.org
http://lists.freeswitch.org/mailman/listinfo/freeswitch-users
UNSUBSCRIBE:http://lists.freeswitch.org/mailman/options/freeswitch-users
http://www.freeswitch.org |
|
Back to top |
|
|
brian at freeswitch.org Guest
|
Posted: Fri Jul 31, 2009 8:44 am Post subject: [Freeswitch-users] pocketsphinx |
|
|
On Jul 31, 2009, at 8:18 AM, Helmut Kuper wrote:
Quote: | -----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Hello,
I spent a few days working on my problem. I played around with
voxforge's data, read some of the train docs of CMU, and debugged the
sphinxbase, pocketsphinx and mod_pocketsphinx.
Results:
- -I found a way to use the voxforge data as training data for
creation of
a german language corpus.
|
Care to document the process?
Quote: |
- -I enabled the logging of pocketsphinx to stderr (Dirty, but easy
way to
see what went wrong when FS loads grammar, mdef, etc and simply
stopped.
Very helpful!
|
I can't recall if there is a logger callback we can register for
this... Last I checked you couldn't this is something we should make a
config option for up in the mod if possible.
Quote: |
- -I had to add a "dictcase" parameter to pocketsphinx.conf.xml resp
mod_pocketsphinx.c to allow case sensitive dictionaries (like the
german
dictionary from voxforge).
|
What do you mean? Can you put this on jira please.
Quote: |
FS starts up with german language model and detects the words as
expected. But it's not so reliable as I want to ... I guess this is
caused by the very small amount of training audio data. I used 4000 of
19000 audio files provided by voxforge due to the reason that
voyforge's
training fileid-list contains only 4000 files ... I have to create
new
fileid-list and transcription-lists containing all audio I have
downloaded from voxforge.
Quite complex the whole thing ...
regards
helmut
|
_______________________________________________
FreeSWITCH-users mailing list
FreeSWITCH-users@lists.freeswitch.org
http://lists.freeswitch.org/mailman/listinfo/freeswitch-users
UNSUBSCRIBE:http://lists.freeswitch.org/mailman/options/freeswitch-users
http://www.freeswitch.org |
|
Back to top |
|
|
msc at freeswitch.org Guest
|
Posted: Fri Jul 31, 2009 1:54 pm Post subject: [Freeswitch-users] pocketsphinx |
|
|
Helmut,
Your hard work is appreciated. Like Brian said, we'd all be interested in knowing more. Please feel free to put this on the wiki or see me off list and we'll discuss further how to document it for the good of the FS community.
Thanks!
-MC
On Fri, Jul 31, 2009 at 6:18 AM, Helmut Kuper <helmut.kuper@ewetel.de (helmut.kuper@ewetel.de)> wrote:
Quote: | -----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Hello,
I spent a few days working on my problem. I played around with
voxforge's data, read some of the train docs of CMU, and debugged the
sphinxbase, pocketsphinx and mod_pocketsphinx.
Results:
- -I found a way to use the voxforge data as training data for creation of
a german language corpus.
- -I enabled the logging of pocketsphinx to stderr (Dirty, but easy way to
see what went wrong when FS loads grammar, mdef, etc and simply stopped.
Very helpful!
- -I had to add a "dictcase" parameter to pocketsphinx.conf.xml resp
mod_pocketsphinx.c to allow case sensitive dictionaries (like the german
dictionary from voxforge).
FS starts up with german language model and detects the words as
expected. But it's not so reliable as I want to ... I guess this is
caused by the very small amount of training audio data. I used 4000 of
19000 audio files provided by voxforge due to the reason that voyforge's
training fileid-list contains only 4000 files ... I have to create new
fileid-list and transcription-lists containing all audio I have
downloaded from voxforge.
Quite complex the whole thing ...
regards
helmut
On 10.07.2009 14:53, Helmut Kuper wrote:
Quote: | Hi,
I try to change pocketsphinx's grammar from default (english) to german.
I found this archive
(http://www.repository.voxforge1.org/downloads/de/Trunk/AcousticModels/), which
contains similar files like those which can be found in
grammar/model/communicator directory.
Unfortunately FS crashed without writing a core file nor logfile enries
as soon as as pizza demo trys to detect speech.
Any Ideas? Maybe someone has already working grammar/model files for
german language?
regards
helmut
| -----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.7 (MingW32)
iD8DBQFKcu8z4tZeNddg3dwRAiVKAKCJdAvZyYXLClWp+kQO4MhsPVzbSQCgoX/T
7NosCujHwpkCXHzxkEY/N3M=
=6Bu8
-----END PGP SIGNATURE-----
_______________________________________________
FreeSWITCH-users mailing list
FreeSWITCH-users@lists.freeswitch.org (FreeSWITCH-users@lists.freeswitch.org)
http://lists.freeswitch.org/mailman/listinfo/freeswitch-users
UNSUBSCRIBE:http://lists.freeswitch.org/mailman/options/freeswitch-users
http://www.freeswitch.org
|
|
|
Back to top |
|
|
helmut.kuper at ewetel.de Guest
|
Posted: Wed Aug 05, 2009 10:00 am Post subject: [Freeswitch-users] pocketsphinx |
|
|
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Hi Michael,
today I put my documentation on FS wiki describing the steps to get a
8kHz sample rate acoustic model basing on voxforge's data for german
language. It's not complete, yet.
You can found it here: http://wiki.freeswitch.org/wiki/Mod_pocketsphinx
regards
Helmut
On 31.07.2009 20:45, Michael Collins wrote:
Quote: | Helmut,
Your hard work is appreciated. Like Brian said, we'd all be interested
in knowing more. Please feel free to put this on the wiki or see me off
list and we'll discuss further how to document it for the good of the FS
community.
| -----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.7 (MingW32)
iD8DBQFKeZ2e4tZeNddg3dwRAlwuAKCXK6b/f3J7tRmcev0/EPAUFGZBbgCfXMQW
B8MAREKeR82dTFnYyFeutig=
=1/Jr
-----END PGP SIGNATURE-----
_______________________________________________
FreeSWITCH-users mailing list
FreeSWITCH-users@lists.freeswitch.org
http://lists.freeswitch.org/mailman/listinfo/freeswitch-users
UNSUBSCRIBE:http://lists.freeswitch.org/mailman/options/freeswitch-users
http://www.freeswitch.org |
|
Back to top |
|
|
msc at freeswitch.org Guest
|
Posted: Wed Aug 05, 2009 5:16 pm Post subject: [Freeswitch-users] pocketsphinx |
|
|
On Wed, Aug 5, 2009 at 9:56 AM, Helmut Kuper <helmut.kuper@ewetel.de (helmut.kuper@ewetel.de)> wrote:
Quote: | -----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Hi Michael,
today I put my documentation on FS wiki describing the steps to get a
8kHz sample rate acoustic model basing on voxforge's data for german
language. It's not complete, yet.
You can found it here: http://wiki.freeswitch.org/wiki/Mod_pocketsphinx
regards
Helmut
|
Thanks!
I'd like to ask the community members who are interested in ASR and PocketSphinx to please review Helmut's page and add to it as well as offer feedback.
-MC |
|
Back to top |
|
|
|
|
|
You cannot post new topics in this forum You cannot reply to topics in this forum You cannot edit your posts in this forum You cannot delete your posts in this forum You cannot vote in polls in this forum
|
Powered by phpBB © 2001, 2005 phpBB Group
|