[asterisk-dev] Interpreting RTPAUDIOQOS statistics

[Don Morrison]

Thanks Olle - the RTPQUDIOQOS code works perfectly as far as I can tell. It
looks like I'll have to do some digging to figure out what those statistics
mean and what part of the call they apply to. I'm still a bit lost on
how/if  the one set of statistics apply to the 2 call legs. A README
clarification from the original developers or somebody more familiar with
RTCP would be great.

Don Morrison



Don,
I just arrived back from holiday, so I could not answer quickly.

The RTPAUDIOQOS variable is a product of my coding, but the RTCP
support is not, so I can't
answer your questions. I added the variable to get something out of
the RTCP support code
and did not get any other suggestions on what to do with the data. I
think this desperately
needs to be clarified in a README file.

In order to get answers, I guess we have to consult the RFC on RTP/
RTCP and the code.
The RTCP reports is sent two ways. The tx data is propably from the
reports from the other end,
sent to us. rxjitter/rxcount is what we receive and send in reports
to the other side.

The other end is teh far side of the asterisk call leg. If SIP device
1 calls SIP device 2 through
Asterisk, you have two calls. One from SIP1 to ASterisk, one from
Asterisk to SIP2. Asterisk
is "our end" and the device is the "other end" of each call.

Frustrating, yes. But I hope my answer gives you some hints.

Regards,
/Olle

26 feb 2007 kl. 23.47 skrev Don Morrison:

>
> I've starting logging RTPAUDIOQOS statistics to CDR records in
> Master.csv
> and I wrote a program to parse and extract out those statistics
> (running
> 1.4). I think it will be very useful, assuming the data is
> accurate. But
> I'm not sure I fully understand the data. Here is an example:
>
> "ssrc=358281685;themssrc=1432487772;lp=0;rxjitter=0.000482;rxcount=386
> ;txjitter=12593.274200;txcount=0;rlp=6424412;rtt=6683.786029"
>
> Questions:
>
> 1) rxjitter is "our calculated jitter" and txjitter is "reported
> jitter of
> the other end". I assume RTP has some way to do this reporting back an
> forth. But which end is the other end? If I have a call that
> originates
> from Sip device 1 and calls Sip device 2, which one is it?
>
> 2) My guess is that these statistics only reflect the connection to
> the
> call originator. Is that correct? If so, are there statistics on the
> connection to the callee? Can I get at them?
>
> 3) Are there standard thresholds for jitter and lost packets that
> indicate
> that the audio is degrading or is unacceptable?
>
> 4) A couple of the values in the example above look bad (eg. rlp and
> txjitter). Can I really trust this?
>
> Thanks.
>
> Don Morrison
>
> _______________________________________________
> --Bandwidth and Colocation provided by Easynews.com --
>
> asterisk-dev mailing list
> To UNSUBSCRIBE or update options visit:
>    http://lists.digium.com/mailman/listinfo/asterisk-dev

---
Olle E. Johansson * Asterisk Evangelist, developer * VOOP A/S
olle at voop.com