摘要
数字音频编解码系统采用MDCT/IMDCT实现时间-频率域的互换,以消除音频分帧引起的时间域混叠效应.本文提出一种新的MDCT/IMDCT快速实现方案,基于N/8点FFT变换核,采用奇偶双路并行和蝶型单元技术,与现有快速算法相比,运算速度和吞吐能力均提高一倍,并且该方案既可以实现MDCT正变换,也可以实现相应的反变换.为了验证方案的正确性,在Altera FPGA开发板上完成了N=256点MDCT的实验.结果表明,该实现方案在运算速度和数据吞吐率等方面取得很大的改进.
MDCT/IMDCT performs the conversion between the time-domain signal and the frequency representation in the high quality audio codec system, which is based on the concept of time-domain aliasing cancellation. Using the N/8-point FFT kernel, a new MDCT/IMDCT general structure is presented in this paper. The computation rate and the data throughput per transformation of the system are twice as much as the latest algorithm, utilizing both the paralleling procedure and the butterfly unit. In addition, the analyzed results of FPGA implementation on Altera Stratix Device show the proposed structure provides a superior performance in terms of computation speed and data throughput.
出处
《中国科学院研究生院学报》
CAS
CSCD
2008年第6期810-815,共6页
Journal of the Graduate School of the Chinese Academy of Sciences
基金
国家自然科学基金(60606005)资助