Which feature allows users to generate live captions in real-time with Azure AI Speech?

The ability to generate live captions in real-time using Azure AI Speech is facilitated by live transcription. This feature is designed to process spoken language and create a text representation almost instantaneously as the speech occurs. Live transcription is particularly useful in scenarios such as meetings or conferences, where participants benefit from having the spoken content displayed in text format for clarity or accessibility.

In contrast, batch transcription is aimed at processing pre-recorded audio files rather than providing real-time capabilities. Proactive speech recognition involves techniques that anticipate speech patterns but does not specifically focus on generating live captions. Audio prompts are utilized in systems to guide users through interactions, often in a question-and-answer format, rather than generating captions. Thus, live transcription is the feature best suited for the real-time captioning requirement.