I have been doing DTFT practice problems for my DSP course, and I encountered this problem in the textbook that completely stumped me.

enter image description here

The question asks to find the DTFT of the shown signal and to plot its magnitude and phase. I am very comfortable with doing this with real-valued signals, but have no idea where to start when it comes to complex signals. Especially how the plotting would be done. I would appreciate any advice on how to tackle the problem.

  • 1
    $\begingroup$ HINT: Complex numbers are numbers too. In exactly the same way that you say $u=2$ and $v = 4 \cdot u$, you can also say $u = 4 + 5j$ and $v = 8j \cdot u$. All operations are also defined over complex numbers. So, all that you have to do here is run your numbers through the sums. HINT2: This waveform has a particularly useful characteristic when it comes to the DFT. $\endgroup$ – A_A Feb 24 '18 at 7:48
  • $\begingroup$ it's simpler than a "complex" number. it's purely imaginary. just factor out the $j$. $\endgroup$ – robert bristow-johnson Feb 24 '18 at 7:54
  • $\begingroup$ Look up the definition of a linear operator. The DTFT is one. $\endgroup$ – hotpaw2 Feb 24 '18 at 12:52
  • $\begingroup$ So my representation of the signal is the following: jdel[n] -jdel[n-4]. Is this the correct representation? From here, I will factor the j and apply the DTFT as it is a linear operator. $\endgroup$ – DSPer Feb 24 '18 at 18:03

I think you almost have the solution. As you suggest in your comment, if you express your signal as : $$ x(n) = j\cdot \delta(n) - j\cdot \delta(n-4) $$ As DFT is linear, you will get : $$ \begin{align} X(k) &= \sum_{n=0}^{N-1}x(n)\cdot e^{-2j\pi \frac{kn}{N}}\\ &= j\cdot \sum_{n=0}^{N-1}\delta (n)\cdot e^{-2j\pi \frac{kn}{N}} - j\cdot \sum_{n=0}^{N-1}\delta (n-4)\cdot e^{-2j\pi \frac{kn}{N}}\\ &=j-j\cdot e^{-8j\pi \frac{k}{N}}\\ &=\text{ }... \end{align} $$

| improve this answer | |
  • $\begingroup$ Thanks for your answer Frank, I think you misread my question about DTFT as being DFT. I was able to use the same logic to get to the final answer, thanks. $\endgroup$ – DSPer Feb 25 '18 at 0:15

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.