MediaSense - Behavior during transfers/conferences

My understanding is that when CUBE is forking the audio then the recordings will contain the entire interaction even when the caller was transferred. For example if a caller is transferred from one device to another, both segments will be included in the recording.

I am wondering if there are any options to change this behavior so that it is similar to when the devices are forking the audio. This might be either a setting in the CUBE to start a new session automatically when a transfer takes place, or possibly our application could send a request when we wanted a recording to be broken into separate recordings.

There is no way to enforce CUBE/MediaSense to start a new session after transfer.

However, if transfer causes a mid-call codec change, a new session would be created, because MediaSense does not supports multiple codec for the same stream.

I don't suppose you know of a good way to force a codec change when a transfer takes place?

You can not enforce CUBE to do mid-call codec change. Usually, it happens in the scenario, like this:

- Phone A supports G.711 & G.722, Phone B supports only G.711, so the call between A & B use G.711, and it is recorded as session 1.

- Phone B transfer phone A to phone C that supports G.711 & G.722, so after the transfer has been completed the call between A & C use G.722.

- MediaSense close session 1 (with g.711) and open session 2 (with g.722)

Sorry, there's no way to do that, unless you really are changing codecs. 

Is this a UCCE contact center environment, using CVP?  Are you specifically trying to separate the IVR and music-on-hold portion of the call from the agent portion?  If so, then please see the discussion titled "Omitting the VRU Segment from a Recording" in the SRND for MediaSense 10.0.

The main driver here is desktop recording. If the call gets broken up every time the caller is transferred to a new agent then each of those recording would show the desktop of the associated agent. It gets more complicated when a single recording contains multiple segments where the caller is speaking to different agents. I realized that I was probably grasping at straws so I am not disappointed that the answer was essentially no, you can't do that.

Thanks for all the information though!