The Analysis Of Recommendation G.729 PinYin Synthesis Algorithm And Its Implement Based On DSP

Speech synthesis is important part of speech communication between computer and people, it is also the important part of speech communication system, with the development of the digital signal processing, technology of communication, semiconductor and computer, many varieties of speech synthesis are developed one after the other, but the quality of speech synthesis is not ideal. In this paper, a speech synthesis's method is given. The method realizes speech synthesis based on recommendation G.729, with the characteristics of high synthesis quality and small speech database.Paper first studied the recommendation G.729 offered by ITU-T, realized the algorithm of speech spell synthesis based on this principle, then simulated the main body of the algorithm of recommendation G.729, including coding algorithm used to compress the sound library; decoding algorithm realized the voice synthesis and completed the realization of underwater speech communication. The transfers matter of this system involved in most of the command, reports, etc, do not need to retain the speaker's voice features too much, such as speech timbre and tone information, only require to receive a higher voice clarity and intelligibility, it can veritably reflect the speaker's meaning. Therefore, the transmitter of underwater voice communication system translate the voice information into text information through voice recognition technology; then the receiver of underwater voice communication system realize text to speech translation used speech synthesis technology. In addition, the paper also studied Regulating Law of a voice of speed and tone, finally, realized the small vocabulary speech synthesis System.Through the system's subjective perception test, the results showed that the reconstructed speech was clear and natural, besides, the algorithm had moderate complexity, which was suitable for embedded system of underwater speech communication.