[gst-devel] how to record mic input while processing via pocketsphinx

[Evren Esat Özkan]

Hi,

I think I've found a solution to my question. I've removed the "appsink  
sync=false name=appsink" part of the pipe and it worked. This is working  
pipeline:

pulsesrc  ! audioconvert ! audioresample ! tee  name=t  ! queue ! vader  
name=vad auto-threshold=true ! pocketsphinx name=asr t.  ! queue ! wavenc  
! filesink location=o.wav

Evren

On Sat, 05 Feb 2011 01:50:24 +0200, Evren Esat Özkan <sleytr at gmail.com>  
wrote:

> Hello,
>
> I'm trying to write a pipeline that records audio to disk while doing
> voice recognition via pocketsphinx (I want to stop the recording when  
> user
> say a predefined keyword). I've spend hours on various alternative
> pipelines but can't accomplish anything. The simplified python code can  
> be
> seen at the url bellow. It creates an empty wav file and speech
> recognition works only for once.
>
> http://pastebin.com/EvSgy0i7
>
> I'd appreciate any help with this.
> Evren


-- 
Using Opera's revolutionary email client: http://www.opera.com/mail/