Hi, I¹m resending this e-mail to the list, because I suspect that it was never actually distributed when I sent it a week ago. (Can¹t find it in the archives). It might be because I attached 2 audio files, which might have caused it to (invisibly) bounce. Please excuse me if some of you did receive the original e-mail.
--this is what I sent a week ago -- In one of my previous post I described the service we built with SEMS. You can find an archived version of this post here: http://lists.iptel.org/pipermail/sems/2013-May/004171.html We recently found out that in some cases we experience poor quality of the relayed media in the B2ABUASession. I¹ve attached 2 audio files, which I¹ve captured using tcpdump and converted to audio files using wireshark. The naming of the files describes the direction. I¹m not sure if the attachments will be distributed on the mailinglist, so I¹ve uploaded them to SoundCloud as well: https://soundcloud.com/tvdgeer/from-a-to-sems https://soundcloud.com/tvdgeer/from-sems-to-b NB. There is another set of files which I did not upload/attached (from_B_to_SEMS.au and from_SEMS_to_A.au) and they have the same problem. So the problem is in both directions. You can clearly hear that the audio coming from the A leg is OK, but the audio from SEMS to the B leg is distorted. Does anybody recognise this type of distortion? It¹s not a PCMU/PCMA incompatibility. It sounds more like a timing problem, like every sample/packet is slightly corrupted. The stream analyses of Wireshark do not display any problem. It might be good to mention that the problem seems to appear and disappear out of nowhere. At one time the problem disappeared by restarting SEMS. On one occasion it disappeared overnight (without restart) but the next day it was back (again without restart). In our setup we run SEMS 1.5 on a virtual system (VMWare hypervisor). Could this be the cause of this problem? Any help is appreciated. Thanks in advance! Best regards, Tom van der Geer
_______________________________________________ Sems mailing list [email protected] http://lists.iptel.org/mailman/listinfo/sems
