Hi,
I have some trouble configuring Sphinx4 to enable speech verification.
What I would like to setup is an enviroment where a wave-file plays a 5 digit number (TIDIGITS) thru my sound card (line out) and into a mobile phone that has a call established.
The recieving phone recieves the information (ie numbers) and sends it into back into the computer (sound card line in: microphone) and then to sphinx for recognition.
My configuration works with a microphone, but not when a radio enviroment is downsampling the information sent. The radio network is a 3G network with a frequency up to ~6kHz.
If I record the file recieved, I could then run it on the wav-file demo that has no trouble understanding the information.
The file has then been recorded as signed, 22050Hz, 16bit little-endian...
All help is very appriciated!
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Hi,
I have some trouble configuring Sphinx4 to enable speech verification.
What I would like to setup is an enviroment where a wave-file plays a 5 digit number (TIDIGITS) thru my sound card (line out) and into a mobile phone that has a call established.
The recieving phone recieves the information (ie numbers) and sends it into back into the computer (sound card line in: microphone) and then to sphinx for recognition.
My configuration works with a microphone, but not when a radio enviroment is downsampling the information sent. The radio network is a 3G network with a frequency up to ~6kHz.
I've tried configuring my microphone with:
<property name="bigEndian" value="false"/>
<property name="sampleRate" value="6000"/>
<property name="signed" value="false"/>
<property name="bitsPerSample" value="8"/>
If I record the file recieved, I could then run it on the wav-file demo that has no trouble understanding the information.
The file has then been recorded as signed, 22050Hz, 16bit little-endian...
All help is very appriciated!