That's the rate that's important. Human speech is sent and received by humans, if you're going to send computer generated speech faster than that (faster than a human can comprehend) and decode on the other end by computer, why use speech at all?