A learning application needs to read generated instructions aloud using an appropriate voice. Which Speech capability fits?

1. Document field extraction
2. Speech-to-text
3. Key phrase extraction
4. Text-to-speech with a supported voice