AI-102 Question 246
Select 3You are developing an application that transcribes audio files into text using Azure AI Speech service. The application must handle multiple languages, support audio files stored in Azure Blob Storage, and provide real-time transcription feedback to users. Which feature or configuration should you prioritize in the Azure AI Speech service to meet these requirements?
- A
Enable the 'Real-Time Transcription' feature in the Speech SDK.
- B
Use the 'Batch Transcription' feature to process audio files asynchronously from Azure Blob Storage.
- C
Configure the Speech service to use a Custom Speech model tailored for your domain-specific language.
- D
Enable the 'Language Auto-Detection' feature to support multiple languages in audio input.
- E
Use the Speech-to-Text REST API to provide real-time transcription feedback to users.
Show answer and explanation
Correct answers: A, B, D
Explanation
The scenario requires real-time transcription, support for audio files in Azure Blob Storage, and multi-language handling. 'Real-Time Transcription' enables live feedback, 'Batch Transcription' supports processing audio in Blob Storage, and 'Language Auto-Detection' handles multiple languages. Custom Speech models and the REST API are not strictly required for the described functionality.
- A. Correct.
This option is correct because the 'Real-Time Transcription' feature allows the application to process audio streams and provide instantaneous transcription feedback, which is a key requirement for real-time operations.
- B. Correct.
This option is correct because 'Batch Transcription' is specifically designed for processing large files stored in Azure Blob Storage asynchronously, meeting the storage and processing requirement.
- C. Incorrect.
This option is incorrect because while a Custom Speech model can improve transcription accuracy for specific domains, it is not a mandatory feature for the stated requirements in the scenario.
- D. Correct.
This option is correct because 'Language Auto-Detection' enables the Speech service to recognize and transcribe audio in multiple languages, fulfilling the multilingual support requirement.
- E. Incorrect.
This option is incorrect because while the Speech-to-Text REST API can be used for transcription, it isn't explicitly necessary to provide real-time feedback since the same can be achieved using the Speech SDK's real-time features.