Research And Implementation Of Anti Packet Loss Wideband Audio Coding Algorithms

With the development of wired and wireless packet based networks and the increase of the channel bandwidths, it is expected that high-quality audio communications will replace the conventional speech communications. However real-time audio communications in operations are facing the packet-loss problems resulting from network congestion, channel disturbance and noise, and so on. These problems have been making severe audio quality degradations and are far away from being solved; therefore the reliable transmissions of compressed audio over the wired and wireless channels have gained much attention.Anti Packet Loss Wideband Audio Coding Algorithm - LDX is a novel and excellent audio coding algorithm for real-time communications. The Multiple Description Coding (MDC) used in LDX is a source coding technology with which information can be transmitted in unreliable networks. Without additional delay, MDC is able to provide robust source coding algorithms against packet losses.The principle of the anti packet loss wideband audio coding algorithm is introduced first, then the key technologies of LDX, especially multi-description algorithm, fixed-point implementation, platform porting and optimization are discussed. To fully exploit the advantage of TMS320C64x fixed-point DSP and improve the efficiency of the program, the originally designed float-point LDX algorithms have to be transformed to fixed-point. In the same time, it is necessary to simplify the algorithm while maintaining the audio quality because of the limited memory of DSP and the demand for real-time application. The complexity of LDX has been reduced, the operative speed of the procedure has been enhanced, and the LDX algorithm is implemented on one TMS320DM642 DSP. The results of analysis and evaluation of fixed-point implementation, the DSP porting and the demonstrates have shown, theoretically and practically, that the proposed fixed-point algorithm can assure both high quality and high compress ratio in the same time, and have proved that the algorithm does have application values, and will be put into real applications.