Voice Activity Detection

[shibinkumaric]

ello Experts ,

I am trying to use the record voice element feature in CVP as part of a project . The objective is to collect the audio captures from customer. Preferred approach is to detect the silence from customer end instead of customer manually entering the DTMF (like # key) to indicate the end of a dialog. I have enabled VAD on the gateway dial-peers (VRU leg on VXML gateway) however system is not able to detect the silence. max_record_time is set as 15 seconds and final_silence value is set as 4 seconds. When I make an inbound call I am getting prompted for audio however call is getting disconnected only after 15 seconds even though there were no audio after the first 4 seconds.

[bill.king1]

Are you perhaps running into this defect, which looks unfortunately like there isn't a workaround that is realistic?

https://bst.cloudapps.cisco.com/bugsearch/bug/CSCvp93556

[shibinkumaric]

Hi Bill , Thank you so much , I think you are right .. We follow the unsupported call flow .. PSTN --> CUBE-->CVP-->VXML GW Is that a limitation on the CUBE software? do you know if any future CUBE release has got a fix for this ?

[bill.king1]

It shows as a status of Fixed but it doesn't say when/where, so you may need to open a TAC case and refer to this defect and try and find out what they define as fixed. Since combo VXML gateway aren't supported anymore, and people are moving to CUBE, this would seem to be something that they'd have to resolve.

If you find a fix from them (i.e. certain version of CUBE IOS, certain version of CVP, etc.), please let us know.

[shibinkumaric]

Sure Bill , we are in touch with Cisco presently . Will update here if I get any positive response from Cisco .