[asterisk-dev] SIT detection in pre-answer audio - guidance for rewriting for trunk

[Kevin P. Fleming]

On 11/25/2010 12:09 AM, D Tucny wrote:
> Hi Folks,
> I had a need to detect SITs, at the time, just on SIP calls in early
> media, but, expected that the same requirement could be present on any
> channel driver... So... I modified app_dial, added an option to enable
> inband progress detection, plugged in a dsp call and during early media
> forced transcoding to slin for the dsp's consumption, switching
> transcoding back on answer... This was with Asterisk 1.2... It's been
> running pretty successfully for the past 6 months, so I want to
> contribute this functionality back.
> What I could use some feedback on is that the way I've implemented this,
> while fine for my needs, probably isn't the 'right way' and does cause
> problems with early media being bridged elsewhere when the dsp is
> attached due to forcing the transcoding, so while looking to rewrite for
> trunk I'm first looking at doing it the right way ...
> My thoughts on this are that perhaps duplicating the frames would make
> sense and feeding one half through transcoding leaving the original
> audio frames intact, but, also wondering whether I'd even need to do
> this with additions since 1.2 such as transcode_via_sln and audiohooks
> perhaps usable to do the main work there...

Yes, an audiohook would be the way to go here.

-- 
Kevin P. Fleming
Digium, Inc. | Director of Software Technologies
445 Jan Davis Drive NW - Huntsville, AL 35806 - USA
skype: kpfleming | jabber: kfleming at digium.com
Check us out at www.digium.com & www.asterisk.org