Fouriers Transform inconsistency with intuition

21 Views Asked by At

lets consider Fourier Transform of function $ f(t) = \begin{cases} e^{-t}, & \text{for $t\ge 0$ } \\ -e^t, & \text{for $t\lt0$ } \end{cases}$ with Fourier Transform $F\{\omega\}= \frac{-1}{i\omega-1}+\frac{1}{i\omega+1}$

Magnitude and phase plots below:

Mag and phase

With all phases equal to zero this means that signal can be recreated with infinite amount of pure cosine waves each with diffrent frequency and amplitude but same property at t=0, which is positive value at that point. There is like infinite number of cosine waves only beetween $\omega=0$ and $\omega=0,00000001$, so how is that possible all of these sum to 1 at t=0?

2

There are 2 best solutions below

0
On

They'd need to be sine waves, not cosines, as the function $f$ is odd. The function is discontinuous at the origin, with 'average value' zero ($+1$ to the right, $-1$ to the left), which is what the Fourier series would give you at $t=0$. When $f$ is discontinuous, agreement of the function and its Fourier series becomes difficult (I realise that's a vague statement but I suspect you're not after a fancy justification).

0
On

I am so sorry, Ive made such a trivial error in equation. It should be $$ f(t) = \begin{cases} e^{-t}, & \text{for $t \ge 0$} \\ e^t, & \text{for $t\lt$ 0} \end{cases}$$

With chart like below Function plot

Why did you refer to Fourier Series as function isnt periodic?