Well, it appears that the all to sip_rtp_read dies when I call
ast_read(p->owner) for some reason..

See what I don't get is that when I call ast_set_read_format, I give it a
channel and a format.

when I call ast_dsp_call_progress I give it a DSP and a ast_frame

what makes the ast_frame be in the right format?

needless to say, I get:
dsp.c:1154 ast_dsp_call_progress: Can only check call progress in
signed-linear frames

any ideas?



On 1/17/07, Leo Ann Boon <[EMAIL PROTECTED]> wrote:

Brett List wrote:
> My SIP Provider (A large major carrier) passes me calls this way. And
> no, they won't disable early media and provide an appropriate
> supervision code.
>
> In fact, all carriers we've tested have transmitted the call this way.
> Even my cell phone gives me the tri-tone without "connecting"
In my part of the world, the telcos will drop the call immediately and
return the cause code.

>
> Yes, chan_sip would need to be modifed. That is the work I am
requesting.
The problem with goertzel detection in G.729 is not trivial. Too much
spectral energy was lost in the transcoding. I'm speaking from
experience with DTMF detection in G.729.
>
> For what it's worth, I've attempted to modify chan_sip in the
> sip_rtp_read func to include adding a dsp with ast_dsp_new, converting
> to SLINEAR with ast_set_read_format and then passing it to
> ast_dsp_call_progress. But it doesn't seem to be doing anything at
> all.. I've been performing my test on G.711.
Your logic is sound. Did you dump out the return values?

Leo

_______________________________________________
--Bandwidth and Colocation provided by Easynews.com --

asterisk-biz mailing list
To UNSUBSCRIBE or update options visit:
  http://lists.digium.com/mailman/listinfo/asterisk-biz

_______________________________________________
--Bandwidth and Colocation provided by Easynews.com --

asterisk-biz mailing list
To UNSUBSCRIBE or update options visit:
   http://lists.digium.com/mailman/listinfo/asterisk-biz

Reply via email to