A client needs to turn a user's microphone speech into text for further processing. Which Speech capability should it use?

1. Text entity recognition without an audio recognizer
2. Speech-to-text
3. Image synthesis
4. Text-to-speech