Microsoft Cognitive-Speaker Recognition Enrollment

Question

I am trying to make use of Microsoft Cognitive for Voice recognition https://github.com/Microsoft/Cognitive-SpeakerRecognition-Android

I have uploaded a voice sample for enrolling and got a OperationLocation object with URL in return.(Didnt see any errors or exceptions)

OperationLocation ops = speakerClient.enroll(audioStream, profile.identificationProfileId);

But when I check for status of Enrollment it always shows its "Enrolling"

Profile p = speakerIdClient.getProfile(UUid);

How much time does it take for Enrollment to be completed or is there any other issue.

Answer 1

The documentation for the enroll operation for speaker verification states that you need to submit at least 3 audios speaking the same text for the enrollment to succeed.

If you are using speaker verification, the documentation says that you need to submit at least 30 seconds of audio.