How inline emotion tags change text-to-speech output

Most text-to-speech integrations treat delivery as a settings problem. You pick a voice, then you set...