Research And Implementation Of AC-3Audio Coding Algorithm Based On DSP TMS320C6727

Digital audio coding technology is an important part of digital multimediatechnology. There are a variety of digital audio coding standards today, includingDolby AC-3(Dolby Digital AC-3) audio coding technology which is widely used inDVD and ATSC (American National Standard Digital Television). The technology ismarked by transmitting multiply channels but requiring little space. Therefore, it canachieve lower bitrate under the premise of the sound quality. Currently, China is inthe process of developing and promoting of digital TV system. It is believed that theefficient and economical AC-3digital audio coding system will have a good prospect.In recent years, software audio encoders based on DSP platform are greatlydemanded by the market. DSP TMS320C6727is a high performance floating-pointprocessor produced by TI, specifically for audio processing. It has rich peripheralinterface, ample storage space and high clock frequency, which can fully meet therequirements of real-time audio encoding.The AC-3audio coding algorithm especially transient signal detection algorithmand transplanting and optimizing those algorithms based on C6727have beenresearched in the thesis. The importance of detecting transient signals in the timedomain is derived by discussing pre-echo phenomenon processing algorithms ofseveral audio coding standards. Several shortcomings have been pointed out byanalysis of transient signal detection algorithm in AC-3standard. According to theshortcomings, the thsis proposes an improved algorithm of fully using of soundmasking effects and human auditory characteristics in the time domain. Experimentalresults show that the algorithm improves the accuracy of the transient signal detectionwhile avoiding excessive short window switch. Finally, the improved AC-3audiocoding algorithm is transplanted and optimized according to C6727of its owncharacteristics to achieve real-time audio encoding in the frenquecy of300MHz. Atthe end of the thesis, an audio encoding system has beed discussed with its designframework and interface definations. The system has been sold in the market.