Voice and Video GSoC Project
cmaiku at gmail.com
Tue May 27 12:16:31 EDT 2008
On Tue, May 27, 2008 at 9:05 AM, Pekka Riikonen <priikone at iki.fi> wrote:
> I haven't checked out all your changes yet and I don't know farsight API
> yet but does the new media API/farsight allow me to get the actual RTP
> data into the protocol specific code? You see, in SILC not only the
> multimedia setup procedure is protocol specific but the actual multimedia
> data, be it voice or video, is encapsulated into SILC protocol packets.
> This means that I'd need to get the data into the SILC prpl. If you check
> the code above in voice.c you'll see that I set up the src and sink
> callbacks to receive and send the actual RTP data. O nce I have the data I
> encapsulate/decapsulate it as defined in SILC protocol, encrypt/decrypt it
> and send it its merry way.
> Ie. SILC provides the transport for the multimedia data. In the end, it
> can be UDP or TCP, but they are always SILC packets, not RTP packets.
> So, is this in anyway possible with the new media API or with farsight or
> with GST the way the Pidgin media API now uses it? If it isn't, is there
> any way to get support for setting such a callbacks? Or do you have an
> idea how it should be done? Without that, or without some other kind of
> mechanism to encapsulate the multimedia data into SILC packets we can't do
> vv support for SILC protocol.
I believe this would be possible with Farsight. Farsight uses plugins
for the different types of protocols, fsrtpconference for RTP
conferences etc. It also uses transmitter plugins to actually send the
data, such as nice for ICE and rawudp for raw udp. In order to support
SILC, if I'm understanding you correctly, plugins will need to be
written specifically for SILC.
More information about the Devel