0
$\begingroup$

I found some multiple choices in a well known book . The problem is that I don't get the answers in some and I want to do so.

Question 1:

The signal x(t) with Fourier transform $X(j\omega) = u(\omega)- u(\omega -\omega_0 )$ can undergo impulse-train sampling without aliasing, provided that the sampling period $T <\frac{2\pi}{\omega_0}$.

Answer: True

Okay here's what I think: $x(t)$ should be sampled with at least:

\begin{align} ωs &> 2B \\ &> 2(B_1-B_2)\\ &>2(\omega_0 - 0) \\ &> 2\omega_0 \\ \frac{2\pi}{T_s} &> 2\omega_0 \\ T_s&<\frac{\pi}{\omega_0} \end{align}

What do I miss?

Question 2: enter image description here

Answer: $50\pi$

For $x(t)$, $\omega_{\text{max}} = 50π + 50π = 100π$ and so $ω_s > 2\omega_{\text{max}} = 200π$. So , if we actually sample with $\omega_s<\omega_{\text{Nyquist}} = 200π$ , then there is no chance to avoid aliasing for all $X(j\omega)$ spectrum but we can gain a proper copy of it as long as $\omega_0 < \frac{\omega_s}{2} = 75π$. Again what do I miss?

$\endgroup$
1
  • 2
    $\begingroup$ Have you considered accepting any answers to your previous questions? $\endgroup$
    – Matt L.
    Dec 19 '20 at 12:45
1
$\begingroup$

Question 1

Perhaps the easiest way to understand this is to draw $X(j\omega)$: Plot of X(j\omega)

With the corresponding aliases after sampling in blue: Plot of X(j\omega) with aliases \frac{\omega_s}{2\pi} \sum_{k=-\infty}^{\infty} X(j(\omega-k\omega_s))

So long as the blue rectangle do no overlap you would not have aliasing. At this point it should be more clear that you need $$ \omega_s \gt \omega_0 \\ T_s \lt \frac{2\pi}{\omega_0} \\ $$

Question 2

Again, plotting $G(j\omega)$ (red), together with $75X(j\omega)$ and an alias image after sampling by $\omega_s$ (dashed blue) can help visualize the situation: enter image description here

Indeed, because $\omega_{\max} > \frac{1}{2}\omega_s$ you have some aliasing. However the question is not whether you can avoid aliasing, but rather to find the value $\omega_0$ below which the aliasing does not affect the spectrum. In the above graph you may notice that the red and dashed blue curves overlap up to $50\pi$. So $\omega_0 = 50\pi$.

$\endgroup$

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.