Hi, I have an RTP packet capture 273 seconds long. There are 13654 packets in the forward direction and 13646 in the reverse. There are no lost packets and virtually no jitter. The codec is G.729. When I use the "Save payload" feature I get files of 285260 and 272920 bytes respectively for forward and reverse. When I then convert these into PCM the audio file in the forward direction is ~15 seconds too long! I would have expected the file sizes to be 20 bytes x the number of RTP packets but it is bigger for the forward. The behavior is the same in Wireshark 1.6.10 and 1.8.2. Why is the forward direction file bigger than expected and converted audio too long? P.S. Unfortunately I can't figure out what is happening by listening to the audio because it is in Greek. Thanks, David asked 20 Sep '12, 07:43  David Sorkin |
One Answer:
If you load the two audio files in an audio editor (Audacity for instance) you can see what it takes to align these streams, even though they are intelligible. Usually one party speaks at a time, so the speech and silence periods should line up somehow. That may show you where the 'extra' data is relative to the audio. 15 seconds sounds like call pickup delay for instance. answered 21 Sep '12, 05:00  Jaap ♦ |
