Hi,
i have a question but i have very low knowledge of sphinx itself...
I try to do my best to describe my problem.
And maybe not all terms i use are correct in the way you (if you are a sphinx
expert) would use them.
There is a server on which my asterisk phone application runs. Now it is
supposed to handle speech recognition.
One guy installed (pocket)sphinx and he created two directories. One is
entitled -de (german) and one is entitled -en.
In either directory there are startup scripts that start the german or the
english version.
Now my problem is of course: I want to have an application that can handle
multi-languages.
Thus my questions are:
1. Is it correct to have two instances of (pocket)sphinx?
1a. Can you run multi languages in a single (pocket)sphinx instance?
2. If it is correct to run two instances - how do you open them as far as ports are concerned? They are configured to open port 10069. If i try to run the second server it says of course "Error setting listen port". If i change the port for the second server to 10068 it still says "Error setting listen port".
3. If i somehow managed to set up two instances - how do i create the speech objects in my asterisk application? SpeechCreate(Sphinx) is what i call now. Where and how are the instances named?
Any help is very much appreciated!
Thanks in advance.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
One guy installed (pocket)sphinx and he created two directories. One is
entitled -de (german) and one is entitled -en.
In either directory there are startup scripts that start the german or the
english version.
Hello, it's not actually clear what exact connector package do you use. It
might be UNIMRCP or Zanzibar or some other package. Once you'll mention that
it would be easier to give you a hint
Is it correct to have two instances of (pocket)sphinx?
It can be either way, but proper way might be different. I think it's easier
to have one MRCP speech recognition server which will be able to work with
multiple languages
If i somehow managed to set up two instances - how do i create the speech
objects in my asterisk application?
Proper way would be to use multiple languages in signle server at once. But
this proper way will require some tweaks on the server. Then you could just
specify language to use in the grammar or use multilanguage grammar. Both
tweaks require some coding.
I think you just need to ask the guy who set you things to finish his work.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Hi,
i have a question but i have very low knowledge of sphinx itself...
I try to do my best to describe my problem.
And maybe not all terms i use are correct in the way you (if you are a sphinx
expert) would use them.
There is a server on which my asterisk phone application runs. Now it is
supposed to handle speech recognition.
One guy installed (pocket)sphinx and he created two directories. One is
entitled -de (german) and one is entitled -en.
In either directory there are startup scripts that start the german or the
english version.
Now my problem is of course: I want to have an application that can handle
multi-languages.
Thus my questions are:
1. Is it correct to have two instances of (pocket)sphinx?
1a. Can you run multi languages in a single (pocket)sphinx instance?
2. If it is correct to run two instances - how do you open them as far as ports are concerned? They are configured to open port 10069. If i try to run the second server it says of course "Error setting listen port". If i change the port for the second server to 10068 it still says "Error setting listen port".
3. If i somehow managed to set up two instances - how do i create the speech objects in my asterisk application? SpeechCreate(Sphinx) is what i call now. Where and how are the instances named?
Any help is very much appreciated!
Thanks in advance.
Hello, it's not actually clear what exact connector package do you use. It
might be UNIMRCP or Zanzibar or some other package. Once you'll mention that
it would be easier to give you a hint
It can be either way, but proper way might be different. I think it's easier
to have one MRCP speech recognition server which will be able to work with
multiple languages
Proper way would be to use multiple languages in signle server at once. But
this proper way will require some tweaks on the server. Then you could just
specify language to use in the grammar or use multilanguage grammar. Both
tweaks require some coding.
I think you just need to ask the guy who set you things to finish his work.