先进音频编码中MDCT的新型递归结构及其FPGA设计  

A New Recursive Architectures for MDCT in Advanced Audio Coding and its FPGA Design

作　　者：孙剑[1] 郭立[1] 林海涛[1] 李果[1] 

机构地区：[1]中国科学技术大学,安徽合肥230027

出　　处：《微电子学与计算机》2006年第4期61-63,共3页Microelectronics & Computer

基　　金：安徽自然科学基金项目(050420102)

摘　　要：文章以新型的MDCT递归算法为基础,首先给出了一种主要模块分时共享、数据流同步多路处理的改进型并行处理结构,其可以高效地应用于国际音频标准MPEG-2以及MPEG-4的先进音频编码(AAC)当中。另外,针对递归实现结构高度模块化和较强规则性的特点,进行了FPGA的设计和仿真,结果表明:改进后的处理结构在增加少量硬件模块的条件下,大大提高了运算速度。Based on the new recursive algorithm for MDCT, this paper presents a modified parallel processing architecture which has the features that the main modules are shared and the data sequence is processed synchronously in parallel. It can be applied more efficiently in the Advanced Audio Coding （AAC） of MPEG-2 and MPEG-4 international audio standards. In addition, because of the high module and the better regulation of this modified architecture we have also made a FPGA design and simulation for it. The result indicates that the modified architecture can make the time consumed in computation fewer than before at the condition that only few hardware modules are added in.

关 键 词：改进的离散余弦变换(MDCT) MPEG-2/4 先进音频编码(AAC) 递归结构 

分 类 号：TP393.11[自动化与计算机技术—计算机应用技术]