Role of Multi-Pulse Excitation for Synthesis of Natural-Sounding Voiced Speech

Invalid Date

LPC synthesis using multi-pulse excitation is known to produce natural-sounding speech. The reasons for the high quality of speech achieved in multi-pulse synthesis are not exactly clear but it is generally agreed that the improvements in voice quality result from the use of a single model for the excitation in producing all classes of speech material. An important advantage of multi-pulse excitation model is that it completely eliminates pitch analysis and voiced/unvoiced decision and thus makes it easier to synthesize voiced fricatives, plosives, and other rapidly changing sounds.