| With the rapid development of computer science and digital signal processing, speech synthesis technology, which is a method of pronunciation for information transportation by machine, is applied for the field of underwater acoustic communication more and more. Speech synthesis is also called TTS (Text to Speech). At the moment, the apply of embedded system to TTS technology will be developed current in the future.Under this background, this paper discusses the method of combining both TTS technology and embedded technology and applies it to underwater acoustic speech communication. The speech synthesis device in this paper is designed based on microprocessor chips S3C2440A of ARM9 series and audio decoder chip UDA1341TS and change text to speech on embedded Linux system. Since the platform is controlled by embedded Linux, the function expandability is very good. As well as, because of the uniformity of ARM9 architecture and generality of buffer circuit, this system can be applied for other multiple utility without modification.In this paper, the feature and principle is introduced firstly, and the merits and faults of three conventional speech synthesis technologies are compared. Next, hardware architectures, design principles and embedded development theory of speech synthesis device are researched in depth. And then the software part of embedded Linux is introduced, the start-up process, cross-compile environment, configuring and compiling of Linux kernel and produce of a root file system of embedded Linux are studied, and the transplant of embedded Linux is successfully completed. In the last part of the paper, the driver and the working process of audio chip on embedded Linux, the manufacture process of voice bank and the generation method of WAV sound file are explained in detail. The audio driver is added in embedded Linux and drive static compilation to a part of Linux kernel. Finally, speech synthesis is implemented by experiment. |
PDF Full Text Request