[Asterisk-Users] SIP to IAX - forcing codec pass thru
Mark Phillips
g7ltt at g7ltt.com
Mon May 22 04:01:15 MST 2006
Hi Peter,
I don't see any codec allow=blah statements. If your end user has
something like
[gradwell]
disallow=all
allow=gsm
Then you'll be forced to send them a GSM coded call.
Why not force the codec at your end by only supporting one? If the
customer then transcodes the call when it gets forwarded to his handset
there's not much you can do about that but at least you'll have handed
the call off in the best format you can source.
Mark
On Mon, 2006-05-22 at 09:57 +0100, Peter Gradwell wrote:
> hi
>
> We take calls inbound via SIP from our Cisco PSTN gateways, and pass it
> to customers using IAX (they run their own asterisk servers).
>
> We've noticed that asterisk is transcoding the call into a different
> codec, if the customer prefers a codec different to that which our cisco
> gw prefers. As such, the quality of the call can degrade.
>
> We'd rather asterisk just passed through the RTP stream and maintained
> the same codec, so that all asterisk did was signalling conversion.
>
>
>
> sip.conf...
>
> ---
>
> [sip-router-1.gradwell.net]
> context=sip-inbound
> type=peer
> host=sip-router-1.gradwell.net
>
> [sip-router-2.gradwell.net]
> context=sip-inbound
> type=peer
> host=sip-router-2.gradwell.net
>
> ---
>
> iax.conf...
>
> [general]
> bandwidth=high
> disallow=lpc10
> jitterbuffer=yes
> dropcount=2
> maxjitterbuffer=500
> maxexcessbuffer=80
> minexcessbuffer=10
> jittershrinkrate=1
> tos=lowdelay
>
>
> ---
>
> when a call comes in, we dial something like this, in our dial plan:
>
> -- Executing Goto("SIP/213.166.5.134-118f5310",
> "sip-users|7770002|1") in new stack
> -- Goto (sip-users,7770002,1)
> -- Executing Dial("SIP/213.166.5.134-118f5310",
> "IAX2/user:pass at customeripaddress/441376350002") in new stack
> -- Called user:3l3phant at customeripaddress/441376350002
> -- Call accepted by customerip (format alaw)
> -- Format for call is alaw
> -- IAX2/customerip:4569-23 answered SIP/213.166.5.134-118f5310
>
> thanks
> peter
>
More information about the asterisk-users
mailing list