Research Of AVS-M Audio Codec And Implementation On DSP

In our country, although the industrial foundation on audio is already very strong, related corporations are still chained to foreign corporations or organizations due to the lack of important techniques and standards. And many patent disputes are faced by most correlative enterprises, which restrict their development greatly. So it is more necessary and stringent to draw up related standards with proprietary intellectual property rights. AVS-M audio standard is now on draw by our country independently based on the framework of AMR-WB+ audio standard of 3GPP. And it targets at wireless network and mobile equipment, such as cellphone. If it could be drawn up successfully, many patent disputes that audio industries at home are confronted with, such as high patent cost, will be settled down, and the development of audio industries in our country will be faster.At first, several particular techniques of AVS-M audio are analyzed, based on the reseach of the technical framework and the core algorithms of AMR-WB+. Then the algrothm innovation of 3 techniques is presented, including sampling rate conversion filter, vector quantization codebook generation and gain quantization techniques which have applied for China Patent, and the sampling rate conversion technique and gain quantization technique are accepted by AVS-M standard. Then, the fixed-point Codec of AVS-M is programmed (The Reference Codec is AVS-M float-point Codec Version9.2), and the complexity analyse function is integrated. Then, the fixed-point Codec of AVS-M is implemented on TMS320C6416 platform. Finally, the objective and subjective tests of the fixed-point Codec of AVS-M is given. This thesis is have significant impact on the spread and application of AVS-M audio.