(New page: Image:Week8_Q3_FFT.jpg a) We decimate 3 times to compute DFT. The gain of one path equals to the product of gain of each decimate. According to the figure. <math>\text{ Gain }=(W_N^0...)
 
Line 38: Line 38:
  
 
-----------------------
 
-----------------------
[[Week8_Quiz_Questions|Back to Quiz Pool]]
+
[[ECE438_Week8_Quiz|Back to Quiz Pool]]

Revision as of 17:25, 12 October 2010

Week8 Q3 FFT.jpg

a) We decimate 3 times to compute DFT. The gain of one path equals to the product of gain of each decimate.

According to the figure. $ \text{ Gain }=(W_N^0)*(-1)*(W_N^2)=-W_N^2 $

b) Only one. In general, there is only one path between each input sample and each output sample.

c) $ x[0]\text{ to }X[2]\text{ : The gain is }1 $

$ x[1]\text{ to }X[2]\text{ : The gain is }W_N^2 $

$ x[2]\text{ to }X[2]\text{ : The gain is }-W_N^0=-1 $

$ x[3]\text{ to }X[2]\text{ : The gain is }-W_N^0 W_N^2=-W_N^2 $

$ x[4]\text{ to }X[2]\text{ : The gain is }W_N^0=1 $

$ x[5]\text{ to }X[2]\text{ : The gain is }W_N^0 W_N^2=W_N^2 $

$ x[6]\text{ to }X[2]\text{ : The gain is }-W_N^0 W_N^0=-1 $

$ x[7]\text{ to }X[2]\text{ : The gain is }-W_N^0 W_N^0 W_N^2=-W_N^2\text{, as in Part (a)} $

$ \text{Since }X[k]=\sum_{n=0}^{N-1}x[n]W_N^{nk} \text{, k=0,1,...,N-1} $

Now

$ \begin{align} X[2] &= \sum_{n=0}^7 x[n]W_8^{2n} \\ &= x[0]+x[1]W_8^2+x[2]W_8^4+x[3]W_8^6+x[4]W_8^8+x[5]W_8^10+x[6]W_8^12+x[7]W_8^14 \\ &= x[0]+x[1]W_8^2+x[2](-1)+x[3](-W_8^2)+x[4](1)+x[5]W_8^2+x[6](-1)+x[7](-W_8^2) \end{align} $

Thus, each input sample contributes the proper amount to the output DFT sample.


Back to Quiz Pool

Alumni Liaison

Ph.D. on Applied Mathematics in Aug 2007. Involved on applications of image super-resolution to electron microscopy

Francisco Blanco-Silva