I intend to build a hight precision IVR for italian language that contain a very small number of words, numbers (less than 100), but precision should be over 95%.

Please, someone guide me with exact steps to follow and to indicate avaible resources for italian language. I whant to contribute with work in this direction.

I think that a better precision can be reached if for each user will be builded his own acustic model (this is not a problem).

> I intend to build a hight precision IVR for italian language that contain a very small number of words, numbers (less than 100), but precision should be over 95%.

Tahnk you! I'll give a try.

hi are you realized your IVR?

I'm trying to follow this guide
http://code.google.com/p/unimrcp/wiki/PocketSphinxPlugin

but i can't find the pocketsphinx.xml in on my disk.

could you help me

Hi

> but i can't find the pocketsphinx.xml in on my disk.

Hi nsh

i followed this guide (but probably isn't the right way)

1) http://code.google.com/p/unimrcp/wiki/InstallationGuide

2) http://code.google.com/p/unimrcp/wiki/asteriskUniMRCP

everything run well also the installation of this step:

3) http://code.google.com/p/unimrcp/wiki/PocketSphinxPlugin

I used pocketsphinx-0.8 and sphinxbase-0.8

but in the plugin folder i have this scenary

root@dantavo-PC:/usr/local/unimrcp/plugin# ls
demorecog.a         demosynth.so        demoverifier.so.0.1.0
demorecog.la        demosynth.so.0      mrcprecorder.a
demorecog.so        demosynth.so.0.1.0  mrcprecorder.la
demorecog.so.0      demoverifier.a      mrcprecorder.so
demorecog.so.0.1.0  demoverifier.la     mrcprecorder.so.0
demosynth.a         demoverifier.so     mrcprecorder.so.0.1.0
demosynth.la        demoverifier.so.0

Well, build problems are usually easy to fix. You just need to read build logs. Upload build logs somewhere if you need help on this issue.

To build pocketsphinx plugin you need to pass configure option

 --enable-pocketsphinx-plugin

To unimrcp. I'm not sure if you did that.

Thank you

you are great ..

now I have:

root@dantavo-PC:/usr/local/unimrcp/plugin$ ls
demorecog.a     demorecog.so.0.1.0  demosynth.so.0      demoverifier.so        mrcppocketsphinx.la        mrcprecorder.a     mrcprecorder.so.0.1.0
demorecog.la    demosynth.a         demosynth.so.0.1.0  demoverifier.so.0      mrcppocketsphinx.so        mrcprecorder.la
demorecog.so    demosynth.la        demoverifier.a      demoverifier.so.0.1.0  mrcppocketsphinx.so.0      mrcprecorder.so
demorecog.so.0  demosynth.so        demoverifier.la     mrcppocketsphinx.a     mrcppocketsphinx.so.0.1.0  mrcprecorder.so.0

and

root@dantavo-PC:/usr/local/unimrcp/conf# ls
client-profiles  logger.xml  pocketsphinx.xml  umcscenarios.xml  unimrcpclient.xml  unimrcpclient.xsd  unimrcpserver.xml  unimrcpserver.xml.bkp  unimrcpserver.xsd

now i set in pocketsphinx.xml the right path of my acustic model of my dic and grammar, but my dubt is if is right to follow the guide at http://code.google.com/p/unimrcp/wiki/asteriskUniMRCP to realize a simulation for my IVR or is better follow another way.

Thanks for your help