Research Of The Computer Aided Speech Processing Based On GA

With the development of the computer technology applied in every aspect in the society, computer has been playing an indispensible role in people's life. The ability of information processing has been improving so rapidly that it meets people's needs of listening, speaking, reading and writing. Since the advantages of being simple, direct and easy to handle, Computer Speech is getting into different fields of the human society. Many researchers are now paying more attention to the field of The Computer Aided Speech Processing, and especially to speech recognition, speech synthesis and text-to-speech, which have been already achieved a great affect in practice.This thesis discusses how the new approach improves the computer-aided speech processing with GA, and especially focuses on the effect of GA being used in speaker recognition. Basically the thesis has been divided into three parts:1.The discussion on the status of the computer aided speech processing research, and the analysis and conclusion on the development of computer-aided speech processing. The importance of the new technology that has been used widely in today's society is emphasized, then a summary review of the concept of voice processing is given in this part.2.The solution that based on GA technology to improve the computer-aided speech processing is proposed in this part. One outcome is that, a new GA-LBG algorithm has been worked out by combining GA with LBG, and solving the problem of local optimum, which is caused by LBG algorithm used for codebook design by vector quantization in speech processing. The other outcome is the Optimization of the algorithm in network training, which is a combination of the improved adaptive genetic algorithm and the RBF neural network, avoids the network-clustering problem in computer aided speech processing.3.The verification of the effect has been made based on the experiment of the speaker recognition, by the application of GA in improving the traditional method. The outcome of experiment shows that using the GA produces a good simulation effect.