Dsp-based Voice Processing System Design

This thesis realized and described speech code based on the issue of ITU-T pronunciation code standard G723.1 (signal source) and the process method on C54xDSP.In all kinds of modern communications system, the pronunciation correspondence always is an item of essential service. How to reduce the voice signal's transmission bit speed on the condition that do not sacrifice the communication quality is an important topic in front of researcher. At the same time, because the voice signal processing carried on with the digital method is at present developed one of most rapid information research areas, many kinds of compress processing method is borning.ITU-T proposed a series of speech code solution above the foundation about the low bit rate, The G723.1 standard is one of them, this code standard is one of the solution based on the CELP model, it contains 5.3kb/s and 6.3kb/s two code rate, this article mainly studied the lower code rate (ACELP) 5.3kb/s.DSP is one kind of special-purpose digital signal processor, the high integrated DSP digital signal processor has many merits that small volume ,low power loss and the operating speed is quick and so on while the ultra large scale integrated circuit technology develop in deep. Therefore DSP is extremely suitable for voice signal compression processing. This topic uses TI Corporation's DSP chip which is series of TMS320C54x realized the coding and the decoding function.This article expounded the low bit rate speech code's basic principle from the discrete mathematics model which produces from the voice signal embarks, and then analyzed the G723.1 standard structure, Then the article has carried on the detailed elaboration to some essential technologies. In the realization process of G.723.1 standard, I have first carried on the algorithm realization and the debugging work in the VC++ environment, Then final solution does in the DSP software simulation platform CCS. Considered algorithm standard itself will be unceasing improved, and the key question that how to reduce the algorithm complex needs to considerate in the concrete realization. Finally in the paper, I also further discussed the question that how to reduce coding algorithm speed and complex.