[Asterisk-Users] Voice detection

[Christian Hecimovic]

dsp.c has silence detection that works quite well for detecting end-of-voice 
silence. It is used to allow only a certain amount of silence at the end of 
voicemails, for instance. See app_voicemail2.c on how to use it, specifically 
the function play_and_record(). Note that the silence threshold (how 
sensitive you are to silence) is read in from the voicemail.conf file. 

Since the silence detection stuff has a nice public API, you can use it for 
any app you write. See app_skel.c for a basic "shell", and follow something 
like app_voicemail2.c. Read in the acceptable values for threshold and so 
forth from a configuration file (there is a nice Asterisk API for this, 
also), and you're set. For your purposes (playing a file after detecting 
silence in a remote voice stream), such an app should be quite simple.

Christian

On Friday 03 October 2003 10:25, Brad Waite wrote:
> Steve,
>
> I don't have any real experience in DSP methodologies, although I have
> picked up on the high-level theories in my research.  However, I am *very*
> strong-willed in the "Where there's a will, there's a way" category.  :)
>
> Here's my current thought:
>
> Sphinx is an open source STT library that can work in real-time
> (specifically sphinx2).  Could we not pipe the called party's audio into it
> and then look for a given period of time with no text output?
>
> I also found this site,
> http://www.cs.wpi.edu/~claypool/courses/525-S01/projects/proj1/ where a
> prof's got a project for students that fits perfectly with what I'd like to
> do.  He mentions Rabiner and Sambur's algorithm (from 1975) for detecting
> isolated speech endpoints.
>
> Brad
>
> Steve Underwood wrote:
> > Hi Brad,
> >
> > If you want to detect that a sound is voice, rather than something else,
> > it isn't easy. There is information around on the Internet about
> > methods, but I have never tried them and don't know how well they work.
> > Unless you have some understanding of DSP I wouldn't bother trying. On
> > the other hand, if you do have some DSP expertise it might be a fun
> > thing to try.
> >
> > Regards,
> > Steve
> >
> > Brad Waite wrote:
> >> Does anyone know if there's public voice detection algorithms
> >> available?  I've scoured the net for the last hour or so, and I can't
> >> come up with anything except a few proprietary or embedded solutions.
> >>
> >> I know dsp.c uses goertzel algorithms for DTMF detection, but how does
> >> one detect voice?
> >>
> >> I dunno, maybe detecting voice isn't the way to go.  I want to begin
> >> playback of a file after a phone/answering machine has answered.
> >
> > _______________________________________________
> > Asterisk-Users mailing list
> > Asterisk-Users at lists.digium.com
> > http://lists.digium.com/mailman/listinfo/asterisk-users
>
> _______________________________________________
> Asterisk-Users mailing list
> Asterisk-Users at lists.digium.com
> http://lists.digium.com/mailman/listinfo/asterisk-users