rohc team mailing list archive

[Didier Barvaux <didier@xxxxxxxxxxx>]

Hello,

>I would like to know your thoughts about an issue I am having, 
>
>Every time I started a rohc decompressor instance, after exactly 1246
>RTP packets (G.711a) received, the decompressor does the job but with
>huge latency, around 170 ms, after that it seems to work correctly but
>I cannot be totally sure since I haven't let it long time receiving
>compressed packets.
>The exact number of packet at where it fails make me think of some bug.

Yes, it looks strange to me too.

>I have a couple of logs, one in which I start sending the RTP packets
>for more than 45 seconds, in that test at packet #1246 it shows the
>latency in wireshark, but I also have a ms counter in code that confirm
>the latency during decompressor function call. 
>The second test I send packets for 10 seconds at a time, during them I
>never restart decompressor but compressor get restarted by calling
>force_contexts_reinit(), this test show same result of latency at
>packet #1246.

OK, good to know.

>I know the RTP stream is Ok since I have inspected under wireshark. I
>know the decompressor get packets on time and produce a latency after
>inspecting receiving end tcpdump capture (see attached image please)
>
>This is a extract of the log that rohc decompressor shows at packet of
>latency.

The log looks fine.

What library version do you use?

Did try to you reproduce the problem on another machine?

How large is the PCAP capture? If you find a way to send it to me, I may try to reproduce the problem on my system.

Regards,
Didier