This is a static archive of our old Q&A Site. Please post any new questions and answers at

Converting RTP Streams into audiofiles


Hi all,

I use a router called AVM Fritzbox 7270 with Linux Ubuntu 10.10. The router has a hidden feature to capture network traffic. It allows me to produces .eth files that can be analyzed via wireshark.

In order to generate an audio file I use the following procedure:

  1. generate a .eth-file and save it to hard drive.

  2. Wireshark->Telephony->RTP->Show all Streams

  3. Select Stream -> Analyze -> save payload

  4. Choose a name of the audio file, set "Format" to ".au", set "Channels" to "both"

In particular the last step does not work to full satisfaction. I get an error message:

can't save reversed direction in a file: File I/O problem!

All I can do is to save audio files using "Channel" as "forward". Both, "reversed" and "both" do not work. From a test call I have 2 audio streams available. Both streams have been converted into audio files using "Channel" as "forward". The first audio file covers only what I said, whereas the second audio file covers only what the other person replied. Of course, I would like to have an audio file that contains the telephone call completely.

I use 2 different DECT telephones, which are only 3 months old:

Siemens Gigaset E36, Codec G.726. Siemens Gigaset SL400H, Codec G.722, G.726.

Is it due to the codecs used by the phones? What can I do?

When looking at Telephony -> VoIPCalls -> Graph there are fat arrows in both directions tagged as "RTP (g711A)" if this helps. Nevertheless, using the "Player" button I can only listen to what the other person said.

Thank you very much indeed for any help.



asked 27 Dec '10, 06:02

erichm's gravatar image

accept rate: 0%

edited 28 Dec '10, 03:36

You never told which version of Wireshark you use, and on what platform.

(27 Dec '10, 10:39) Jaap ♦