I'm having trouble visualizing the Fourier series of $x \sin(x)$, say if it's graphed over $[0,4 \pi]$ and then repeated.

This is an integrable function, so I should be able take the Fourier series right? I'm just having trouble with the intuition here.

I know the Fourier transform gives a picture of 'how much' of each frequency of sin functions is used in the construction of the function as a Fourier series. Each 'loop' could be thought of as a sum of sines with different frequencies, and so this would certainly not be a Dirac impulse like $sin(x)$.

I was initially worried about the lack of differentiability at each integer multiple of $4 \pi$ but realized that's nonsense, as a saw tooth function has a Fourier series.

Can someone clear up my confusion? Sorry if I'm casually throwing around incorrect terminology; I haven't had to think about Fourier series for over a year.

  • $\begingroup$ You mean the Fourier series of $f(x) = x \sin(x) 1_{x \in [0,4\pi]}$ periodized, not the Fourier transform. And $f(x)$ integrable means its Fourier coefficients are well-defined, but not that its Fourier series converges. Here its Fourier series converges, because except at $x=0$, it is $C^1$, and its discontinuity at $x=0$ is of the same kind as the sawtooth $s(x) = x 1_{x \in [0,4\pi]}$ whose Fourier series converges (you have to prove it). Note that from the Fourier series of $s(x)$ you can deduce the Fourier series of $(x)$ directly $\endgroup$ – reuns Sep 14 '16 at 9:57
  • $\begingroup$ The Fourier transform of a periodic function is well-defined only as a distribution, and is nothing more than a different representation of the Fourier series. $\endgroup$ – reuns Sep 14 '16 at 9:59

A smooth function has a Fourier series with fast decaying coefficients. The best example is the sinusoid (or more generally any trigonometric polynomial), with all coefficients zero but a finite number of them. Another is $e^{\cos(x)}\sin(\sin x)$, with coefficients $1/n!$

On the opposite, discontinuities correspond to small decays because smooth functions have trouble recreating them.

In the given case the function is continous,


but its derviative isn't,


Zero-th order discontinuities give decays in $1/n$, while first order discontinuities (derivative) are in $1/n^2$.

| cite | improve this answer | |

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.