The problem with e.g. G.726 is that Wireshark gives those packets RTP type 102 which afaik is an error code ("unknown encoding").
I would need to know where and how Wireshark maps dynamic payload types (negotiated via SDP) to internal static ones. Above that RFC3551 notes that
static G.726 payload types are obsolete, and afaik there aren't even (obsolete) static payload types for all G.726 variants, so Wireshark would need to
take care of that by using some (more or less arbitrary) internal static type numbers.
I will do my best to provide a patch once I have fully integrated all codecs (currently only G.726-32 has been implemented as proof of concept, but since
this is working adding more is no big deal).
Getting G.726 to work was a bit of a pain btw because of the weird frame sync calculation in rtp_player.c::play_channels() as this function seems to
assume 1:1 relationships of decoder input and output stream sizes and thus simply halves the decoder output batch sizes to determine whether frames
are properly sync'd. This doesn't work for compressed audio. To compensate, my decodeG726_32() function doubles the number of bytes returned
(as it has a 1:2 relationship of input and output buffer sizes). Before it did that, lots of silence frames were inserted and half of the audio data was
dropped by the player.
Dietfrid
> From: jaap.keuter@xxxxxxxxx
> Date: Tue, 25 Jan 2011 16:54:25 +0100
> To: wireshark-dev@xxxxxxxxxxxxx
> Subject: Re: [Wireshark-dev] G.722 and G.726 decoders for Wireshark
>
> Hi,
>
> That would be interesting. Can you put the code in a patch on bugzilla?
>
> Can't work on it right now, but would be nice to have.
>
> btw: their are already static RTP types assigned for both codecs. The dynamic types should come in through protocols like SDP, or a dissector preference.
>
> Thanks,
> Jaap
>
> Send from my iPhone
>
> On 25 jan. 2011, at 16:07, Dietfrid Mali <karx11erx@xxxxxxxxxxx> wrote:
>
> > Hi,
> >
> > using spandsp, I have added G.722 and G.726 decoders to Wireshark.
> >
> > Currently this is a bit of a hack job, particularly regarding inclusion of the spandsp lib, and I could need a bit help to properly integrate it into Wireshark's automake hell (configure.in).
> >
> > There also isn't a proper Wireshark signature for that RTP type (I am simply reacting to RTP type 102, which actually is an error code), so some help getting this straight and introducing proper codec types would be appreciated, too.
> >
> >
> > ___________________________________________________________________________
> > Sent via: Wireshark-dev mailing list <wireshark-dev@xxxxxxxxxxxxx>
> > Archives: http://www.wireshark.org/lists/wireshark-dev
> > Unsubscribe: https://wireshark.org/mailman/options/wireshark-dev
> > mailto:wireshark-dev-request@xxxxxxxxxxxxx?subject=unsubscribe
> ___________________________________________________________________________
> Sent via: Wireshark-dev mailing list <wireshark-dev@xxxxxxxxxxxxx>
> Archives: http://www.wireshark.org/lists/wireshark-dev
> Unsubscribe: https://wireshark.org/mailman/options/wireshark-dev
> mailto:wireshark-dev-request@xxxxxxxxxxxxx?subject=unsubscribe
|
