I have added queueing support so that when we read from the softsynth we can put the text and commands we have read into a queue. This will allow us to retry calling espeak_synth() if we fill espeak's internal buffer.
