Research And Implementation Of G.726Speech Codec System

With the development of mobile communication and the Internet, the growth ofspeech coding is becoming increasingly fast. The International TelecommunicationUnion also updates constantly its coding scheme. G.726is a waveform coding moderecommended by ITU, which is characterized by high quality of encoded speech, shortdelay, and good stability. G.726is based on the ADPCM (Adaptive Differential PulseCode Modulation), which recompresses the output signal of standard G.711. G.726is avery effective speech waveform coding scheme. This paper aims to design a G.726coding system in order to get voice capture, encoding and decoding, playing, storing,transmission and other functions.This article focuses on the implementation of various algorithms on a DSP module.From the whole of coding algorithm, this design is divided into adaptive quantizationand adaptive prediction module respectively. In adaptive quantization module, it canachieve the six sub-adaptive algorithms: input PCM format conversion, calculation ofthe differential signal, quantization scale factor adaptation, adaptive speed control, tonesignal and transfer (transient) detection. In adaptive prediction module, it can implementthe reverse adaptive quantization, adaptive prediction and reconstruction signalsub-algorithms. This article will describe in detail about the hardware implementationmethods of algorithm and take the reliability into consideration from a cost and strivetowards the development and commercialization of product. This article follows strictlyITUG.726standards, and uses MATLAB verified each encoding module according tothe standard recommended, and optimizes algorithms to meet the coding algorithm ofMOS score. When this algorithm will verify conformity to the requirements, it willtransplant to DSP system. It will achieve multi-rate coding in one system, which willaim to get a great G.726coding system...