Research And Implementation Of Speech Coding At Low Rating Based On CELP

With the rapid development of modern mobile communication, the bandwidth utilization and speech quality have been the crucial problems. The valid approach to solve this problem is low rate, low cost and high quality speech codec.Syncretize method of coding based upon CELP (Code Excited Linear Prediction) is extensive researched in the thesis. The CELP algorithm includes many important module such as LP (Linear Predictive) analysis, adaptive codebook, random codebook and gain quantization of code etc. On the base of the G.729 algorithm, the procession-objection is 20ms-time voice-frame. The 4.2kb/s algorithm of multiple pulse-dispersion algebraic codebook-excited linear prediction is developed. Multiple pulse-dispersion technique is applied to disperse the vectors of fixed codebook to reduce the algorithm complexity.The DSP hardware platform is TMS320VC5509A DSP, while software platform is CCS2.0. The software system design based on hardware platform includes the interrupt request service and speech coding. It also adopted some optimized methods according to the characteristic of TMS320VC5509A DSP chip.The test of the optimal code programming shows that the encoding and decoding results are correct and the output speech has a highly quality.