AI-102 Question 247
Single answerYou are developing an application that transcribes live speech into text using the Azure AI Speech service. The application must support multiple speakers in a conversation and identify each speaker in the transcription. Which feature or configuration should you use to achieve this?
- A
Enable the 'Profanity Filtering' feature in the Speech service.
- B
Use the 'Custom Speech Model' to train the service for speaker recognition.
- C
Enable the 'Speaker Diarization' feature in the Speech service.
- D
Use the 'Real-time Endpoint' without additional configurations.
Show answer and explanation
Correct answer: C
Explanation
The Speaker Diarization feature in the Azure AI Speech service is specifically designed to identify and differentiate between multiple speakers in a conversation. While other features like Custom Speech Models or Real-time Endpoints enhance transcription accuracy or provide real-time capabilities, they do not address speaker identification. Enabling Speaker Diarization fulfills the scenario's requirement of identifying individual speakers in the transcription.
- A. Incorrect.
This option is incorrect. Profanity Filtering helps remove or mask offensive language in the transcription but does not identify speakers.
- B. Incorrect.
This option is incorrect. Custom Speech Models are for improving the transcription accuracy for specific vocabularies or accents, not for speaker identification.
- C. Correct.
This option is correct. The Speaker Diarization feature allows the Speech service to identify and differentiate between multiple speakers in an audio stream.
- D. Incorrect.
This option is incorrect. Using the Real-time Endpoint provides basic transcription capabilities but does not include speaker identification without enabling additional features.