Thanks for the update!
I can think of feature enhancements, such as the calculation of the minimum
jitter buffer size to accommodate a call, as well MOS/R-value calculations.
Regards,
Frank
-----Original Message-----
From: Jacques, Olivier (OpenCall Test Infra) [mailto:olivier.jacques@xxxxxx]
Sent: Friday, October 06, 2006 1:43 AM
To: frnkblk@xxxxxxxxx; Community support list for Wireshark
Subject: RE: [Wireshark-users] VoIP analysis and assessment
> Yes, Wireshark can re-construct the audio, but it's without the
> jitter-buffer of the client device in mind. It merely strings the RTP
> packets together and makes a WAV file. I learned this the hard way.
This is not true anymore. The "VoIP Calls/RTP Player" feature (as
available in latest development releases of Wireshark 0.99.4) allows to
reconstruct the audio _with_ jitter buffer in mind.
It works this way:
- You specify the jitter buffer size (in ms)
- You press "Decode" button: Wireshark re-construct the audio.
- RTP packets with an excessive jitter are dropped
- The number of RTP packets dropped are counted and displayed
- You can listen to resulting audio from within Wireshark
See picture attached.
Of course, this doesn't take into account other client-side parameters
like adaptive jitter buffer, bad clocking, bad RTP implementation, ...
Last warning, RTP player supports G711 A/u law codecs at the moment. It
is possible to add your own codecs, the RTP player feature being well
designed for that, but codecs licensing issues will certainly prevent
many codecs to be included in Wireshark.
Olivier.
