Assume a bandlimited signal $X(t)$. Given that the output for this signal is $X(t/2)$, what will be the impulse response $h(t)$ of such a system?

\begin{array}{l} X( \omega ) \ =\ \int ^{\infty }_{-\infty } x( t) e ^{-j\omega t} dt \\ Y( \omega ) \ =2X( 2\omega )\\ \end{array}

As the system won't be LTI the output won't be the convolution of the input and the system response. How do I proceed with finding the response of such a system? Any guidance in the direction of the solution will be very helpful.

Edit: Thanks for all the answers, i had reframed this question which was on a similar line. The question goes like this -> Given an input pulse of duration T, the system scales it to a pulse of 2T, design such a system? It would be of great help if someone could point out the right direction to proceed.

  • 1
    $\begingroup$ Note that this system is not time-invariant. You may well calculate the response to an impulse signal, but calling it H (as if it were a transfer function for an LTI system) is not correct. $\endgroup$ – Juancho Jun 2 '20 at 12:59
  • $\begingroup$ Yes i agree, but is there any way to calculate this impulse response, as convolution also won't be defined between the input and the impulse response. $\endgroup$ – user8206055 Jun 2 '20 at 13:10
  • $\begingroup$ I have edited the question, the earlier version was how I had reframed the question from what I understood, I have posted the original question in the Edit. If anybody could give any direction, it would be of great help. $\endgroup$ – user8206055 Jun 3 '20 at 2:40
  • $\begingroup$ You could convolve the impulse with itself, this will give you a new impulse twice as long. $\endgroup$ – Matt L. Jun 3 '20 at 7:33
  • $\begingroup$ Maintaining the shape of pulse is also needed. $\endgroup$ – user8206055 Jun 3 '20 at 7:38

The input-output relation of such a system can be written as


Note that the system is linear but time-varying, and such systems can generally be described by a two-dimensional impulse response $h(t,\tau)$:


In the given example we simply have


For linear time-invariant (LTI) systems, $h(t,\tau)$ only depends on the difference $t-\tau$.


Because you have this information about the fourier transform of the output that $Y(\omega) = 2X(2\omega)$, so I will suggest you to use the time-frequency scaling property of Fourier Transform itself to compute response of the system to any input $x(t)$.

Time-Frequency scaling of the Fourier Transform says that: $$x(at) \leftrightarrow \frac{1}{|a|}X(\frac{\omega}{a})$$

Notice that the arrow is double sided and so we know that the system's output for any input $x(t)$ in time domain will be inverse fourier transform of $Y(\omega)$. Here $a=\frac{1}{2}$ and so the time-domain output $y(t)$ will be given by: $$y(t) = \mathcal F^{-1}\{Y(\omega)\} = x(\frac{t}{2})$$.

You can use this relation between input and output to find the response of this system.

You can not call it impulse response of the system. It is just input-output relationship of the given system. But if you could give an impulse as input to the system then you would have gotten $y(t) = \delta(\frac{t}{2}) = 2\delta(t)$ as output.


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.