计算机工程与应用 ›› 2016, Vol. 52 ›› Issue (18): 92-97.

• 网络、通信与安全 • 上一篇    下一篇

AVS-P10音频解码器算法分析与定点化

王  实1,2,涂卫平1,2   

  1. 1.武汉大学 国家多媒体软件工程技术研究中心,武汉大学 计算机学院,武汉 430072
    2.武汉大学 深圳研究院,广东 深圳 518057
  • 出版日期:2016-09-15 发布日期:2016-09-14

Algorithm analysis and fixed-point implementation of AVS-P10 audio decoder

WANG Shi1,2,TU Weiping1,2   

  1. 1.National Engineering Research Center for Multimedia Software, Computer School of Wuhan University, Wuhan 430072, China
    2.Research Institute of Wuhan University in Shenzhen, Shenzhen, Guangdong 518057, China
  • Online:2016-09-15 Published:2016-09-14

摘要: AVS-P10是我国第一部应用于移动环境的音频编解码国家标准。在分析AVS-P10解码算法的基础上,对参考代码进行了精简和封装。针对定点处理器应用需求,对精简后的代码进行了定点化实现和优化。分别采用CMOS评分与SNR指标,对定点解码器的解码质量进行了主、客观测试,并对优化前后的定点解码器的运算效率进行了比对测试。结果表明,提出的AVS-P10定点解码器的解码音质达到与浮点解码信号的音质相当,且运算复杂度明显下降。

关键词: 移动音频, 先进音视频编码标准-第10部分(AVS-P10), 定点化, 算法优化

Abstract: AVS-P10 is the first national audio coding standard for mobile applications. This paper reduces and packages the reference source codes of AVS-P10 decoder based on the analysis of algorithm. To meet the demands of applications on fixed-point processor, the fixed-point algorithm of AVS-P10 decoder is implemented and optimized. CMOS as well as SNR is taken to assess the quality of audio signal decoded by the fixed-point decoder. Also, the complexity of the fixed-point decoder is tested. The results indicate that there is no palpable decline of the audio quality between the floating-point and the fixed-point decoder, and the computational complexity significantly decreases.

Key words: mobile audio, Advanced Audio and Video coding Standard-Part 10(AVS-P10), fixed-point, algorithm optimization