Greetings!
To associate client-side live transcription sessions with logged audio files in Azure AI services, you need to understand how the logging mechanism works. When using the Azure Speech SDK for transcriptions, the audio and transcriptions are automatically logged on the TextToSpeech endpoint, which is associated with log IDs. To identify which log ID corresponds to a particular transcription session (e.g., session A/user A or session B/user B), you can follow these steps:
- Use Log IDs: Each transcription session's logs come with unique log IDs. You need to capture these log IDs during the transcription process.
- Client-Side Logging: Implement client-side logging to capture these log IDs as they are generated. This ensures you can map each log ID to the specific transcription session.
- Metadata Association: Associate the log IDs with user-specific metadata. This metadata can include user identifiers, session details, or any other relevant information that helps in identifying the session.
Unfortunately, the current documentation does not specify a direct method to retrieve the log ID at the client side during the transcription process. However, you can implement a workaround by capturing relevant metadata at the start and end of each transcription session, which can then be used to match with the logged audio files.
For more detailed information, please refer to the Azure documentation on logging audio and transcriptions for speech recognition: How to log audio and transcriptions for speech recognition - Azure AI services.
Hope this helps. If you have any follow-up questions, please let me know. I would be happy to help.
Please do not forget to "up-vote" wherever the information provided helps you, as this can be beneficial to other community members.