Hi,
I am looking for a tool to test a audio conference application.

I really like the Kitcat _frame_work but I'm block in the testing of vocal activity.

I want to check that 2 SIPagents connected to the same audio conference can hear each other.
I can't use the playAudio(...) / has(incomingMedia()) mechanism because there is no silence suppression on the system (RTP packects are always sent even when nobody talk).
I can't use the playDtmf()/getDtmfReceived() mechanism because the system don't let the DTMF go the whole conference.

Did you have any idea on how to manage that ?

Thanks.

Laurent

Hi Laurent

You are right that has(incomingMedia()) only check if there are incoming RTP paclets. We have a similar test but we can enable silence suppression on the mixing media server so we use that as a workaround.

However, KitCAT saves the received RTP packets as a raw audio file in the out/ directory. So this can be analyzed after the KitCAT test completes. Here are two suggestions:

1) Check that the received RTP packets contains some audio, and not just all silence. e.g. you can use SoX to trim off silence below a certain threshold.

2) Check that the received audio has the same fingerprint as the sent audio by the other agent. This is probably a better test than above, but more complex. This is a page with software implementation that computes a fingerprint of an audio file. http://en.wikipedia.org/wiki/Acoustic_fingerprint

Hope this helps. I think it would be ideal if we can enhance KitCAT to perform (1) or (2), but these are some workarounds in the meantime.

Hi,

Thank you Eric for your answer.

I found a workaround to manage the check of communication via the DTMF send and get mechanism.

I will let you know if I'm blocked by something else