Hi,
I am playing with voice quality statistics for my sip calls and I need to clear something strange to me.
I have RTCP in both directions but after the call hangup I can see that rx jitter is always reported as 0.000000
In the h extension I am calling a macro which is using ${CHANNEL(rtpqos,audio,all)} to report all needed parameters to calculate MOS.
Executing [s@macro-rtcp3:3] NoOp("SIP/PS1-0000015a", "RTPAUDIOQOS: ssrc=474481140;themssrc=2176743560;lp=0;rxjitter=0.000000;rxcount=466;txjitter=0.000038;txcount=465;rlp=1;rtt=0.251255") in new stack
Executing [s@macro-rtcp3:4] NoOp("SIP/PS1-0000015a", "RTPAUDIOQOSBRIDGED: ssrc=1048430821;themssrc=1723414528;lp=0;rxjitter=0.000000;rxcount=465;txjitter=0.000140;txcount=466;rlp=0;rtt=0.215690") in new stack
I don’t think rjxitter is 0 because I’ve captured the call and I can see in wireshark there is jitter in both directions. At the same time all the presentations related to qos I found over the internet show rxjitter 0 too… I think there is something wrong. Any idea?
I am using default provided by ubuntu 16.04 lts Asterisk 13.1.0
Anything that “calculates” MOS based on jitter is only estimating it, not measuring it, as it can only be measured using opinions from a panel of humans. Also, based on rather a short bit of research, it looks liike jitter is actually used to to estimate packet loss and it is the packet loss that actually predicts the MOS degradation. The TX jitter figures you quote shouldn’t be causing packet loss, so there would be no degradation estimated.
I don’t know why RX Jitter is being reported as zero.
Got the same result:
– Executing [s@macro-rtcp3:2] NoOp(“SIP/PS1-00000000”, “-- QoS stats ssrc=1387192683;themssrc=137373955;lp=0;rxjitter=0.000000;rxcount=487;txjitter=0.000147;txcount=487;rlp=1;rtt=0.019027”) in new stack
– Executing [s@macro-rtcp3:3] NoOp(“SIP/PS1-00000000”, “-- QoS stats RTPAUDIOQOS: ssrc=1387192683;themssrc=137373955;lp=0;rxjitter=0.000000;rxcount=487;txjitter=0.000147;txcount=487;rlp=1;rtt=0.019027”) in new stack