[asterisk-users] Better AMD module

[John Rose]

> -----Original Message-----
> From: asterisk-users-bounces at lists.digium.com [mailto:asterisk-users-
> bounces at lists.digium.com] On Behalf Of Trevor Benson

> What version of asterisk are you running.

1.6.0.28.

> What shortcomings are you experiencing in AMD?

Was thinking that an asynchronous AMD that runs longer and sends AMI UserEvents would be better. The existing AMD is synchronous and does a fast detection within a few seconds.

Also was thinking that generic answering machine beep detection would help, using an FFT or existing Asterisk dsp.c functions.

And AMD on a non-T.38 call won't report fax, modem or SIT answerers. (I realize SIT audio comes in before answer)

Also for some reason I need to play some audio out to the carrier to get RTP coming in otherwise AMD always reports an erroneous result.

> What type of tuning have you done or settings are you using with AMD?

I've messed around some. I haven't run it live yet across multiple carriers and am worried about accuracy. 

> What are you doing after you run AMD on the call? If the call is human
> you do X if its not you do Y?  Are these AGI's or Goto's or???
> 

FastAGI from C#. If AMD produces a MACHINE result I play a file that pauses if it is interrupted by received audio and restarts doing this x number of times until the message is laid onto the voicemail recorder. If HUMAN I assume human and prompt for DTMF etc..

So I am looking for a better AMD.c if there is one available...

Thanks,
John Rose
john at westfax.com


> --
> Trevor Benson
> dCAP, LPIC-1, CLA, Network+, MCP, CNA
> A1 Networks - Network Engineer
> DID (707)703-1041
> FAX (707)703-1983
> 
> 
> 
> 
> 
> 
> On May 26, 2010, at 8:24 AM, John Rose wrote:
> 
> 
> Has anyone written a better AMD than the default AMD? The existing AMD
> works great but it has a few shortcomings...
> 
> I do know about Sangoma but am just looking for a better AMD module.
> 
> Thanks,
> John
> 
> --