If a machine is not fast enough to handle text-to-speech conversion, or if it becomes overburdened, the text-to-speech engine may fail to produce digital-audio data quickly enough. The following techniques can be used to resolve this problem:
· Maintain an internal cache of up to 1 second of digital audio, making notes where a word or a phrase breaks. Do not send digital-audio buffers until enough cached data exists for a whole word or phrase. If the engine falls behind, the pauses will occur at word or phrase breaks where they will be less noticeable.
· If the engine is running on a machine that is too slow, incorporate operations that reduce computation time at the expense of quality.