I'm asked to compute the DTFT of the following signal but i'm quite stuck

$$ \begin{cases} (-1)^{\frac{n}{2} + 1} & \text{ if } n \text{ is even} \\ 0 & \text{ if } n \text{ is odd} \end{cases} $$

Here where I am, i'm starting by using the definition of the DTFT

$$ \begin{align} X(e^{j\omega}) &= \sum_{n=-\infty}^{+\infty} x[n]e^{-j\omega n}\\ &= \sum_{n \text{ even}} (-1)^{\frac{n}{2} + 1} e^{-j\omega n} \\ &= \sum_{k=-\infty}^{+\infty} (-1)^{\frac{2k}{2} + 1} e^{-j\omega (2k)} \quad (n = 2k, k \in \mathbb{Z}) \\ &= \sum_{k=-\infty}^{+\infty} (-1)^{k + 1} e^{-j\omega (2k)} \\ &= - \sum_{k=-\infty}^{+\infty} (-1)^{k} e^{-j\omega (2k)} \\ &= - \sum_{k=-\infty}^{+\infty} (-e^{-2j\omega})^k \end{align} $$

and here i'm stuck because I don't see how to compute the last sum... can someone help me ?

Thanks !

  • $\begingroup$ Is your signal causal? i.e, is x[n] = 0 for n < 0? Then the answer is simply an infinite geometric series sum (en.wikipedia.org/wiki/…) 1/(1-exp(-2j\omega)) $\endgroup$
    – orchi_d
    Jun 11, 2022 at 14:16
  • $\begingroup$ @orchi_d: This won't work, even if all terms were zero for $n<0$, since the magnitude of the summands is not less than unity. $\endgroup$
    – Matt L.
    Jun 11, 2022 at 15:13
  • $\begingroup$ That is true, it would only work if if there was a term less than 1 multiplying the exponent. This sum does not converge. $\endgroup$
    – orchi_d
    Jun 11, 2022 at 16:28

2 Answers 2



Try to rewrite the given signal as


with the correct sign and some appropriate $\omega_0$. Then you can just use the well-known identity for the DTFT of a cosine.

Note that the sum you're trying to compute doesn't converge in the conventional sense.

  • $\begingroup$ Thanks ! It works well :) $\endgroup$
    – Bozu
    Jun 12, 2022 at 8:06

This periodic function has a Fourier series.

$$ \sum\limits_{n=-\infty}^{+\infty} \delta(t-n) = \sum\limits_{k=-\infty}^{+\infty} e^{j 2 \pi k t} $$

where $\delta(\cdot)$ is the Dirac delta function and we're playing a little fast-and-loose with the expression of $\delta(t)$ outside of an integral.

variations in scaling:

$$ T\sum\limits_{n=-\infty}^{+\infty} \delta(t-nT) = \sum\limits_{k=-\infty}^{+\infty} e^{j 2 \pi k t/T} $$

$$ \sum\limits_{n=-\infty}^{+\infty} \delta(f_0t-n) = f_0\sum\limits_{k=-\infty}^{+\infty} e^{j 2 \pi k f_0 t} $$


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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