0

So a buddy of mine and I are sitting in our signals class, and he asks me a question: if the coefficients of a discrete time Fourier series are period, what are their Fourier coefficients? I understand that the Fourier transform of a Fourier transform returns a translated version of the original function, is this true with series as well?

For a DT signal $x[n]$ with period $N$ and frequency $\omega_0 = \frac{2\pi}{N}$, the Fourier coefficient $a_k$ is

$$a_k = \frac{1}{N}\sum_{n=\langle N \rangle} x[n]e^{jnk\omega_0}$$

If we treat $a_k$ as a periodic signal $a[n] = a_n$ (we know it will be periodic with the same period), then its Fourier coefficients $b_k$ are

$$b_k = \frac{1}{N}\sum_{m=\langle N \rangle} a[m]e^{jmk\omega_0} = \frac{1}{N}\sum_{m=\langle N \rangle}\left( \frac{1}{N}\sum_{n=\langle N \rangle} x[n]e^{jnm\omega_0}\right)e^{jmk\omega_0}.$$

Is there any noticeable comparison between $x[n]$ and $b[n] = b_n$?

  • I'm not familiar with your notation, but for this post, the Fourier series is the Fourier transform on the torus. As such the Fourier series has all the characteristic properties of the Fourier transform.

    The involutivity is one of those.

    – Marco Feb 29 '24 at 18:21

0 Answers0