So I have the following equation $$ y[n]=0.9y[n-1]+0.1x[n] $$

We can find easily find the transfer function, which is

$$ H(z)=\frac{0.1}{1-0.9z^{-1}} $$

and from that, the frequency response, which is

$$ H(e^{j \omega})=\frac{0.1}{1-0.9 e^{-j \omega}} $$

To find the cutoff frequency we just make the magnitude of the above equal to $\frac{1}{\sqrt{2}}$. Then, $\omega_0 = 0.0675\Rightarrow f_0 = 0.0107$.

Okay, so how can we say that given $f_0 = f_c$ the filter is a LPF or HPF?

  • 1
    $\begingroup$ Does the frequency response increase or decrease as the frequency increases? $\endgroup$ – MBaz Nov 17 '17 at 2:30

Types of filters, such as LPF, HPF, BPF or BSF, are not described by their cutoff frequencies $f_c$ but through their passband or stopband frequency $\omega$ region.

For example, in your case of the filter $H(z) = 1 / (1 - 0.9 z^{-1})$, the passband is in the low frequencies (where the magnitude of the frequency response $H(e^{j\omega})$ is greater than $1/\sqrt{2}$. And by your definition that region corresponds to $\omega \in (-f_c , f_c)$ which indicates a lowpass range of frequencies.

Other filters like HPF and BPF follow in a similar manner. Note that specific filters such as comb or notch types are defined in a different style than this.

| improve this answer | |

It's actually the Z Transform

\begin{equation} H(z) = \frac{0.1}{1-0.9z^{-1}} \end{equation}

Now get the frequency response by evaluating \begin{equation} H(e^{j\omega}) = \frac{0.1}{1-0.9z^{-1}} \Bigr|_{z = e^{j\omega}} \end{equation}

You can now plot magnitude and phase for a given frequency $\omega$

| improve this answer | |
  • $\begingroup$ you're correct mitch. i have enough rep that i can edit his question. the only thing i need to do to fix yours (which is not the same as completing it) is to change $w$ to $\omega$. $\endgroup$ – robert bristow-johnson Nov 17 '17 at 4:55

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.