On Fri, Oct 1, 2010 at 7:27 PM, P.Muge Ersoy <muge.ersoy at gmail.com> wrote: > Hi Kabil; Hi again M?ge, > Could you explain more about what you are trying to do.. > I mean as a first look at your mail i suspect that you are trying to build > an rtp stream from one point to another ... > you may use only rtp libraries such as ortp .. > Because my application involves actually SIP, rather than isolating SIP, I needed to modify whole pjsua and underlying pjmedia and etc. Currently any SIP client can connect to another for m=application, m=audio (also m=video, but there is no codec for video) (multiple lines of media is supported.) > Does sip involve with your application ? > So yes. > May be it is better to think about what you really need? Modifying pjsips > rtp and media management would not be an answer for your application since > pjsips rtp and audio media bounds are really tight as far as i see.. > As you said, stream, codecs, jitter is really very tightly bound. But I already modified upper levels (pjmedia, pjnath, pjsip...) on 1.6 and will migrate to 1.8. So the latest step is to decide on either removing encoder/decoder, jitter or introduce dummy encoder/decoder and circular buffer. But first is rather easier than the latter, just bypassing codec things. I preferred this way because I will use all pjsua features. > regards > best regards > muge > > (I tried to reach you after your vacation (by guessing the date), but I was unsuccessful.) > 2010/10/1 Kabil Akp?nar <kabilakpinar at gmail.com> > >> So anybody? >> >> 2010/10/1 Kabil Akp?nar <kabilakpinar at gmail.com> >> >> Hi all; >>> I am trying to give m=application support in pjsip, and close to the end. >>> But before I proceed, I need to double check the "stream" issue. >>> For m=audio stream has encoder and decoder attached and a jitter buffer. >>> In m=application case, incoming data does not need any encoder or decoder >>> for the moment, so incoming/outgoing must just pass without any >>> modification. In this case I need to either remove enc/dec parts for >>> m=application and modify get_frame/put_frame or I should introduce a dummy >>> encoder/decoder, but removing seems easier, latter is not that hard in fact. >>> Also the buffer used for storing incoming/outgoing data is not necessary, >>> simply the data can be immediately passed to upper level by callback >>> mechanism.. >>> >>> So what would be your very valuable suggestions? >>> >>> -- >>> Kabil Akp?nar >>> >>> >> >> >> -- >> Kabil Akp?nar >> >> >> _______________________________________________ >> Visit our blog: http://blog.pjsip.org >> >> pjsip mailing list >> pjsip at lists.pjsip.org >> http://lists.pjsip.org/mailman/listinfo/pjsip_lists.pjsip.org >> >> > > _______________________________________________ > Visit our blog: http://blog.pjsip.org > > pjsip mailing list > pjsip at lists.pjsip.org > http://lists.pjsip.org/mailman/listinfo/pjsip_lists.pjsip.org > > -- Kabil Akp?nar -------------- next part -------------- An HTML attachment was scrubbed... URL: <http://lists.pjsip.org/pipermail/pjsip_lists.pjsip.org/attachments/20101001/a21bf1f2/attachment.html>