Learning DSP on my own time. Can't figure out the proof for DFT shift theorem which states the following:

Given, $x[n]$ to be a periodic with period $N$, $\text{DFT}\{x[n]\} = X[k]$, then $$ DFT\{x[n-a]\} = e^{-j\frac{2\pi}{N}a}X[k] $$

I found a proof here, but I can't figure out how did they leap from $$\sum_{m = -\Delta}^{N-1-\Delta}{e^{-j\frac{2\pi}{N}k(m+\Delta)}x[m]}$$ to $$ \sum_{m = 0}^{N-1}{e^{-j\frac{2\pi}{N}km} e^{-j\frac{2\pi}{N}k\Delta} x[m]} $$

Specifically how did they change the limits of the summation without changing the variable?

  • 3
    $\begingroup$ By the way, you haven't accepted a single answer to any of the questions you've asked so far on this site. Why is that? $\endgroup$ – Matt L. Jan 26 '18 at 8:27
  • 1
    $\begingroup$ @MattL. For some questions I can't accept an answer because it doesn't the answer doesn't fully satisfy me however when I add answers from other people I do get a full picture. However, I did upvote the answers that I liked $\endgroup$ – flashburn Jan 26 '18 at 20:18

For an $N$-periodic sequence $y[n]$, the sum over one period is always the same, no matter how you choose the first index:


for any integer $\Delta$. Try to verify that the sequence you sum over is indeed $N$-periodic.

| improve this answer | |
  • $\begingroup$ i'm glad someone else here seems to agree with me that the DFT and the Discrete Fourier Series are the same damn thing. $\endgroup$ – robert bristow-johnson Jan 27 '18 at 17:52

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.