Why do we need special code to make it sound right, humans can read text with a neutral voice, why can't text-to-speech do the same? The main problem I notice with these is the speed and tone it reads different words at.