3 ms·

You just need to take a cue from wifi and use beam forming to send separately synced audio to each person.