Are you rendering the audio of the hint system in real time using some kind of text-to-speech plugin or is it pre-rendered?