FAQs (Voice Verify)

A: The deployment into a development environment, along with the testing of the integration and getting the feeling of internal use can take 1-2 weeks.

From our experience, the integration into existing infrastructure is what takes the most of the time. The whole process starting from defining processes, deployment of Voice Verify, configuration, calibration and pulling results to Call Center SW is estimated to take several weeks.

The customer’s internal processes can affect the timeframe considerably. Additional info including timelines could be found here.

A: No, Voice Verify does not extract these characteristics from the user’s voice.

You can obtain them by using Speech Platform, which is another Phonexia product. More details about Speech Platform can be found here.

A: It is possible to enroll users via recordings. In this case however only mono files are accepted.

There is currently no mechanism to detect which channel in stereo or multi-channel recording contains the voice of the desired speaker.

For that reason, the admin of Voice Verify must ensure that recordings used for voiceprint creation are mono and contain the voice of the desired speaker only.

A: Please note that audio recordings can only be used for enrollment, not for verification. Voice Verify expects the audio coming from the user’s device to originate from a relatively calm environment.

The user should avoid extensive background noise such as loud music, a street with heavy traffic, etc. The agent should warn the user during the enrolment call if speech quality is not good (audible) enough. The speech quality can be optionally indicated by Audio Quality Estimation score. More info about this feature is available here.

During the verification part of process, if the user cannot be authenticated, the agent should ask the user to move to a quieter environment if possible. In case verification is not possible, the agent should switch to a different authentication method.

In general, audio coming to Voice Verify should follow standard telephony encoding without alteration (i.e. not to be compressed to lossy format and then decompressed back to telephony standard before reaching Voice verify).

A: Voice Verify virtual machine is currently compatible with VMware virtual machine hardware version 16. If you plan on running Voice Verify in ESXi hypervisor, please inform us about your virtual hardware version according to the table available here.