Hello! We with the colleague have carried out the detailed analysis of the traffic between Asterisk and sipXtapi based softphone.
Result: http://aspxws.workline.ru/public/sipXtapi_rtp.zip Sequence of actions: 1) softphone_1 calls abonent_1, waits when the abonent_1 will respond 2) softphone_1 calls abonent_2 (here works softphone_2), waits when the abonent_2/softphone_2 will respond 3) softphone_1 transfer a call to abonent_2/softphone_2 (using sipxCallTransfer(...)) As a result, softphone_2 responding to an incoming call receives at first one rtp-stream (conversation with softphone_1), and then, after call transfer, receives the second rtp-stream (conversation with abonent_1), and here there is a divergence. Streams different (different seq, timestamp and SSRC), but sipXtapi them tries to process as one, and certainly there is a RTP-sequence failure. I attach a bright example. Captured traffic by WireShark in which it is visible that comes two rtp-streams (with different SSRC). And sipXtapi debug-log where it is visible, that at first there has been begun processing of the first stream, and then at once the second, though SSRC at streams the different. Why sipXtapi does not look on SSRC? After all if in one conversation streams with miscellaneous SSRC it is necessary to switch at least to new sequence seq/timestamp? Or I am not right? -- Best regards, Alexey [ [EMAIL PROTECTED] ] _______________________________________________ sipxtapi-dev mailing list [email protected] List Archive: http://list.sipfoundry.org/archive/sipxtapi-dev/
