AC> On 3/5/07, Alexander Chemeris <[EMAIL PROTECTED]> wrote: >> On 3/2/07, logan <[EMAIL PROTECTED]> wrote: >> > I have applied your patch. >> From your cap-file I see that PhonerLite use 20ms packets. Try to remove >> 30ms patch and see if it will work. AC> Hum, your cap file look strange for me. May be I misunderstodd something. AC> PhonerLite sent SDP with no "mode=20" option and should use 30ms packets. AC> Your UA sent SDP with "mode=20" option, so should use 20ms packets. AC> And real RTP stream show that they send packets vice versa - PhonerLite use AC> 20ms packets, while your UA use 30ms packets.
PhonerLite sent ADP attribute ptime:20 stating that it requests 20ms frame, but it's not enough. It should still use "mode=20" with iLBC. sipX-based UA responded with "mode=20" because this logic is implemented within SdpBody for GIPS codec (it does so in response to ptime:20). Nevertheless it uses my 30ms patches, so it sends 30ms frame. My iLBC code originally used codec ID different from GIPS one, so there was no problem with the special handling of GIPS codec in SdpBody. Alexander changed the codec ID to be identical with GIPS, and this is the source of erratical behaviour. If the codec IDs remains unchanged, sipX would respond with "mode=30", forcing PhonerLite to use 30ms frame or drop iLBC from "common set" of codecs. Best Regards, Andrzej Ciarkowski _______________________________________________ sipxtapi-dev mailing list [email protected] List Archive: http://list.sipfoundry.org/archive/sipxtapi-dev/
