The Research On Segmentation And Recognition Algorithm Of Syllables And Phonemes For Mandarin

Language is the most important means of exchanging information among the mankind. With the continuous developing and application of electronic computer and the increasing improvement of artificial intelligent, people expect to make the computer recognizing human language. This requirement makes the technique of speech recognition have immense space to develop. Up to now, speech recognition of English and other sign languages have matured, while the Mandarin’s has developed. For a long time, most study on speech recognition is based on conventional linear system theory, such as Vector Quantization(VQ), Dynamic Time Warping(DTW), and Hidden Markov Model(HMM). With the deep study of speech recognition, it is found that the human vocal system is a very complex nonlinear system. Speech recognition that is based on traditional linear model has been limited. To bring break through, nonlinear system theory method must be introduced to it. Recently, with the development of nonlinear-system theories such as artificial neural networks(ANN), it is becoming the hot spot for today’s speech recognition today.This paper has studied on the theory and method for speech signal processing and speech recognition. It has proposed researching technical route of segmentation of syllables and phonemes and recognition algorithm of voice for Mandarin, which starts from the preprocessing of speech signal, then extracts feature of speech signal and estimates glottal closure instants, after that, segments the voice as voiced/unvoiced/silent and recognizes the pitch of speech, at last, introduces the recognition method of English syllables and applies BP artificial neural network to recognise syllables and phonemes for Mandarin. This paper has completed the design and development of system of segmentation of syllables and phonemes and recognition algorithm of voice for Mandarin based on the Matlab platform.By doing analysis of MFCC, LPC, and LPCC parameters on the Mandarin syllable phoneme recognition system, the results show that the LPCC parameter is better than the other else in the syllable phoneme segmentation and recognition system. LPCC parameters of BP artificial neural network model has a good effect for the recognition of Mandarin syllables based on phoneme.