Analysing G729 RTP Stream where there are RTPEVENT frames for DTMF signalling. After extracting RTP using Analyse > Save Payload and converting to .au with normal method, found that with the start of the first DTMF digit, the converted audio is "wonkey". Sounds like cross-talk from many, many conversations. Can hear pulses corresponding to DTFM entry. If I manually strip the RTPEVENT frames from the pcap before processing as above - all is well. Using another tool (Voipmonitor) to convert original pcap to .raw there is no problem with te extracted audio. Seems that presense of the RTPEVENT packet is confusing the stream analysis process. Am I doing something wrong? Thanks, John asked 10 Feb '11, 06:57  johnrcampbell |
One Answer:
There's little you're doing wrong. Wireshark is really tailored to continues PCM A/u-law stream. All the 'fancy' Video, RTP Event, VAD/CN packets disrupt this process. Is this feature open for improvement? Yes. There's just no one willing/able to invest the time to improve this difficult peace of code. answered 10 Feb '11, 09:02  Jaap ♦ |
