Research On Speech Coding Model Of VBR-CELP

Speech communication is a basic and important way to get information. In the development of communication technology, spectrum resource is more valuable, and speech coding technology can compress transmission bandwidth of speech signal, and increase the system capacity. With the development of the third generation mobile communication, it is trend of digital cellular and micro cellular network to improve utilization extent of band using variable rate speech coding.In this paper, basic theory of wavelet transform and teager energy operator is studied, according to voice activity detection algorithm based on discrete wavelet transform and teager energy operator leads to error decision of unvoiced speech, an improved activity detection algorithm based on discrete wavelet transform and teager energy operator is presented. The algorithm can be adaptive threshold, and achieve good performance and stability. Basic theory of rate decision algorithm is studied, and a new source control rate decision algorithm is presented using speech classification theory, which can provide four rate and is reduction of the average bit-rate. In this paper, the new source control RDA and CELP encoder are combined, and VBR-CELP model is presented. Simulations in MATLAB demonstrates VBR-CELP model. The results prove that it is low average rate and good rebuild speech.VBR-CELP model is high-quality variable bit rate speech coding algorithm. Its rate and quality reach require of mobile communication, which can be applied mobile communication, and increase the system capacity in contrast to fixed speech coding .