Research And Realization Of Embedded Speech Synthesis System

Speech synthesis offers the way to incorporate speech output in a machine, which means computer can represent information with speech. The history of speech synthesis research dates from 200 ago. With the developing of technology of computer and digital signal processing, speech synthesis has had application in many areas. Speech synthesis is generally known as text to speech (TTS).TTS synthesis system which can realize the process of a speech signal output from input text is a system that consists of a range of components, from abstract linguistic analysis of discourse structure to speech coding. To apply TTS synthesis to embedded system has became one trend of development of speech synthesis.The embedded English-Chinese translation system is composed of three blocks: speech recognition blocks, machine translation block and speech synthesis blocks. The speech synthesis blocks implements the function of transform the text information from the machine translation block to speech output. Base on the requirement of implementation of embedded English-Chinese translation system, the research of this paper is focus on how to realize TTS synthesis in translation system. Following is the main contents of the paper:(1) At first the paper gives a historic retrospect of speech synthesis and the present status of research at home and abroad. And then the paper induces the main theory of speech synthesis including of the main components of TTS synthesis system and all the steps to realize TD-PSOLA synthesis algorithm.(2) Introducing the present technology of speech coding and theory of speech coding on Recommendation G729a. For the purpose of applying this algorithm of coding, the paper has analyzed the implementation of G. 729a in C language.(3) Designing and implementation a Chinese TTS system based on SPCE061A microcontroller. Having collected and compressed syllables which is basic data used to synthesis speech including pronunciations of whole Chinese word by tools, the system can be used to synthesis speech output from text input by using multiplextable scanning and speech coding API functions.(4) For the purpose of actual application and need of, it is necessary to reduce the algorithm in complexity and time in development of embedded English-Chinese translation system. Using the present technology of speech synthesis, the paper designs and realizes the function of speech synthesis block in translation system basing on TTS chip.A conclusion of the paper is made in the end. After pointing out the flaw of the system in the conclusion, the paper makes same suggestion and prospect for further research.