[Freeswitch-dev] VAD, TALK and NOTALK events

[KMTDBOURSE]

Hi,

   Thanks a lot for describing the vad algo it seems rather a basic 
algo and I am surprised it is not more sophisticated. I hear also 
from speex users (must be same algo?)  that the vad algo is very poor 
and should be improved. That is what I am doing, try to improve vad 
and echo cancellation algorithms. I would appreciate a link where I 
can find data files (audio) to download on which I can perform tests. 
Of course I will feedback the freeswitch community in case of improvment.

best

Zelda



At 15:09 15/02/2008, Anthony Minessale wrote:
>The algorithm used is not so great and could stand to be replaced
>by a better one if we get the chance.
>
>basically we decode the audio to signed linear 16bit samples
>for each packet we receive and iterate the samples and add up the
>absolute values then compute the average and compare it to the
>acceptable level considered "talking" then we use a counter to make
>sure it's in the same state several times in a row before actually
>changing the status of the channel to talking or not talking.
>
>
>A real vad algorithm would also be able to filter out the voice from
>the background noise and judge the energy based only on that.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.freeswitch.org/pipermail/freeswitch-dev/attachments/20080216/0622b2ea/attachment.html