Thanks for all your time

Payload Type is ITU-T G.711 PCMU ( From EtherReal )

Thanks

Will,

That is the RTP payload type. Unsniff can save those to WAV format without problem. I was interested in the signalling protocol that sets up the voice channels.

DO you have a test file you can send via email to
v i v e k [at] unleashnetworks.com ? (remove the spaces).

If you cant send the file, can you check the following ?

Do you see any \"SIP\" packets (in Ethereal) ? If yes, what UDP port number is it using ?

Regards,
Vivek Rajan

Hi Will,

It turns out that there is no signalling information present in the capture file. The capture file contains just voice packets. So, you have to tell Unsniff explictly what ports are used by RTP. In your capture file, ports 19450 -> 29710 are used.

Use the following steps to correctly deal with this.

• Open Unsniff
• Select Plugins->Manage Access Points
• Click on UDP, then click on \"New UDP Access Point\"
• Enter 19450 for the port, Type \"RTP\" for the protocol.
• Click OK
• Click OK to close the Access Point manager

Now import your capture file.

Switch over to the \"User Objects\" tab.
You can see the calls, Click on \"Play\" to hear the call, or \"Save\" to save as WAV file.

I was able to play back the test file without problems. I also sent you the saved WAV file.

-

Next, we can see how to script this whole thing.



Regards,
Vivek Rajan

Thanks Vivek - much appreciated that works great through the GUI.

Is it possible to save both channels(in and out) into one wav file as when I selected both and clicked save, alothough it saved a wav file there was no sound. but when i saved each indivdually there was sound ? I presume this is because the example I gave had no outgoing noises ( i.e. i wasnt speaking but had incoming noise from the call.

Does that make sense !

Will

Will,

The bad news first.
The current version of Unsniff (bld 1127) only tracks each leg of a call separately. We know it is restrictive and some other testers have pointed it out as well. One is using external mixers, which is quite painful and sometimes inaccurate.


The good news.
We have working on it and it will be there in the next build (in a week or so). However, it has the following restriction.

* It will only mix a two-way call. Multiway (conference calls) will still be exported as individual legs.

You could probably start off the scripting exercise by saving each file as <filename>_Fwd.wav and <filename>_Rev.wav. You could also wait until we have the mixing (2-way only) support before starting the entire exercise.

Best Regards,
Vivek Rajan

Vivek - Thanks again for all your time.

Is it possible you could push me in the correct direction re scripting the conversion of the cap to wav, as per your previous post.

Thanks

Will