Re: [reSIProcate] rtp ringing

[Robert Sparks <rjsparks@xxxxxxxxxxx>]

One other caveat while you're exploring this idea.

If you send your ringing from one place (ip address+port) and then  
the media from
the other phone from another (directly from the phone), you could run  
into a problem
with some endpoints that "latch" onto the source of the first media  
packet that they receive
in an attempt to deal with multiple streams coming in as a result of  
a forked invite.

Probably not an issue for you since you said you've built an SBC, but  
I wanted to make
sure you knew it was there.

RjS

On Nov 20, 2007, at 5:01 PM, John McBride wrote:

> For the G711 ring tone - you can alter the frequencies slightly to
> make the difference 100Hz,
> e.g. 445Hz and 345Hz - which will give 100Hz separation and will
> equate to an integer number
> of 8kHz samples. From memory in the UK the frequency tolerance  
> allowed is 5%.
>
> I have done this in the past and it has worked well.
>
> John
>
>
> On 21/11/2007, Adam Roach <adam@xxxxxxxxxxx> wrote:
> > For what you describe, a full RTP stack would be overkill. You can
> > probably get by just sending your packets with an appropriately  
> > encoded
> > RTP header every 20 ms. You'll also need to listen for incoming RTP
> > packets and discard them. (Ideally, you'd calculate statistics  
> > based on
> > these packets and generate proper RTCP; however, in practice, things
> > will pretty much work just fine if you don't).
> >
> > Also: while what you're proposing will work _mostly_ fine for codecs
> > that don't rely on prior frames for encoding (like G.711), you are  
> > very
> > likely to run into difficulty coming up with a single G.729 or G.723
> > packet to represent "ringing." Even with G.711, you'll run into  
> > problems
> > because the standard North-American 350+440 ringback tone doesn't  
> > repeat
> > neatly on 20 ms boundaries.
> >
> > You might try sending an RFC4733 representation of ringback (e.g., if
> > you're targeting the US market: modulation = 0, volume = -6 dBm0,
> > frequency = 350, frequency = 440, duration = 2 seconds; sent every 6
> > seconds or so). Support for this approach may be more limited, so  
> > make
> > sure you test it against all your target devices.
> >
> > /a
> >
> > On 11/20/07 10:37 AM, David Thompson wrote:
> > > Quick question if anyone has an idea for me:
> > >
> > > I have built an sbc using the reciprocate stack. (THANK YOU to  
> > > all who
> > > have worked on this stack)
> > >
> > > I want to provide my own "ringing" early media rtp for comfort
> > > reasons. I know reciprocate does not offer an RTP stack but I was
> > > wondering if someone could point me in the right direction on  
> > > where I
> > > could look.
> > >
> > > Basically I would already have the ringing file encoded in g729,  
> > > g723,
> > > g711 and would simply want to issue a 180 early media from the sip
> > > stack then start transmitting the rtp in the proper codec. Once the
> > > call completed I would then let the media pass from end point to
> > > endpoint directly.
> > >
> > > Any thoughts would be appreciated……
> > >
> > > David
> > >
> > > -------------------------------------------------------------------- 
> > > ----
> > >
> > > _______________________________________________
> > > resiprocate-devel mailing list
> > > resiprocate-devel@xxxxxxxxxxxxxxx
> > > https://list.resiprocate.org/mailman/listinfo/resiprocate-devel
> >
> > _______________________________________________
> > resiprocate-devel mailing list
> > resiprocate-devel@xxxxxxxxxxxxxxx
> > https://list.resiprocate.org/mailman/listinfo/resiprocate-devel
>
>
> --
> John McBride
> Mobile: +61 430 062 005
> _______________________________________________
> resiprocate-devel mailing list
> resiprocate-devel@xxxxxxxxxxxxxxx
> https://list.resiprocate.org/mailman/listinfo/resiprocate-devel