O'Reilly logo

Learning Microsoft Cognitive Services - Second Edition by Leif Larsen

Stay ahead with the world's most comprehensive technology and business learning platform.

With Safari, you learn the way you learn best. Get unlimited access to videos, live online training, learning paths, books, tutorials, and more.

Start Free Trial

No credit card required

Enrolling a profile

With a speaker profile in place, we need to associate spoken audio with the profile. We do this through a process called enrolling. For speaker identification, enrolling is text-independent. This means that you can use whatever sentence you want for enrollment. Once the voice is recorded, a number of features will be extracted to form a unique voice-print.

When enrolling, the audio file you are using must be at least 5 seconds, and 5 minutes at most. Best practice states that you should accumulate at least 30 seconds of speech. This is 30 seconds after silence has been removed, so several audio files may be required. This recommendation can be avoided by specifying an extra parameter, as we will see in a bit.

How you choose ...

With Safari, you learn the way you learn best. Get unlimited access to videos, live online training, learning paths, books, interactive tutorials, and more.

Start Free Trial

No credit card required