d*****u 发帖数: 17243 | 1 【 以下文字转载自 EE 讨论区 】
发信人: daigaku (๑۩۞۩๑), 信区: EE
标 题: 离散傅立叶变换,这么说对吗
发信站: BBS 未名空间站 (Tue Sep 7 02:19:19 2010, 美东)
N久以前学的了,忘光了。。。
离散傅立叶变换相当于对信号采样后进行傅立叶变换,然后再在频域采样? | k**********g 发帖数: 989 | 2 Yes.
Fixed length, fixed interval, finite time (input N time samples, output N
frequency samples)
Time domain: (1) multiply by impulse train, (2) multiply by boxcar (finite
time)
<-->
Frequency domain: (1) convolve by impulse train (hence periodic 2*pi), (2)
convolve by sinc (hence it is allowed to interpolate the N frequency
responses by a sinc function)
有错请指正,谢
Edit: not sinc, should be "periodic sinc" (a ratio between two sinc)
http://www.mathworks.com/help/toolbox/signal/diric.html;jsessioni |
|