【作者】 苏祖辉；

【导师】 陈钟鸣；

【作者基本信息】 合肥工业大学 ， 微电子学与固体电子学， 2005， 硕士

【摘要】 MDCT和IMDCT是MPEG/Audio编译码中两个运算量比较大的模块,它们的功能在于把编译码中的样本进行从时域到频域、从频域到时域的转换,以便于编码压缩处理或者是可以使信号能够实时的播放。在整个编译码系统中,它们都具有很重要的地位,其运算量占30%以上。随着各种数码产品的不断出现,编译码技术也在不断提高,MDCT和IMDCT的研究也越来越深入,不断有新的实现方法出现,但是要寻找一种既快速又适合于硬件VLSI上实现的并不是很容易。本文在研究传统的各种实现方法的基础之上,依据硬件设计的规律,提出一种新型的实现方法--递归循环实现的方法,并利用现有的新的工具--DSP Builder对这种实现方法进行仿真、综合并进行验证,为以后的算法实现提供了一种新的思路。更多还原

【Abstract】 The MDCT and IMDCT are two of the most computational intensive operations in he MPEG/Audio. 1 LayerIII standard .They can transform the data from time field to the frequency field or from frequency field to the time field , thus we can easily encode 、 compression and to play the file in realtime . In the whole encode and decode system , they are both very important .With the emerge of all kinds of digital products, the technology of encode and decode are improving too . The study on the MDCT and IMDCT is deeper and deeper ,and there were many new fast implementations , but it is difficult to find a fast and fit to the VLSI in hardware.In this paper ,on the study of many traditional implementations ,we derive a new recursive way , it obeys the design rule of hardware .and we simulations 、synthesis and verify it by the new tool- DSP Builder . This gives the otheralgorithms a new solutions to implementation .更多还原