Create a Notch filter with low and high pass filters

26 vues (au cours des 30 derniers jours)
Luke McDevitt
Luke McDevitt le 4 Déc 2022
Commenté : Paul le 5 Déc 2022
I have been attempting to create a notch filter where I can specifiy the range of frequncies that will be filtered out. This is my best attempt:
[y,Fs] = audioread('africa-toto.wav');
notchfilt = highpass(y,1000,Fs) .* lowpass(y,100,Fs);
figure
freqz(notchfilt, 1, 2^16, Fs)
sound(notchfilt,Fs)
I'm not sure why but the resulting audio playback is very very grainy and doesnt remove the frequencies specified correctly.
  1 commentaire
Paul
Paul le 5 Déc 2022
Hi Luke,
Not sure why the outputs of highpass() and lowpass() would be multiplied. If anything, it looks like they should be added
% create some data
Fs = 5000;
rng(100);
y = randn(1024,1);
% DTFT of the input
[hy,wy] = freqz(y,1,2^16);
% the output and its DTFT
notchfilt = highpass(y,1000,Fs) + lowpass(y,100,Fs);
[hn,wn] = freqz(notchfilt,1,2^16);
figure
plot(wy/2/pi*Fs,[abs(hy) , abs(hn)]),grid
Having said, designing the bandstop filter directly is the way to go.

Connectez-vous pour commenter.

Réponse acceptée

Star Strider
Star Strider le 4 Déc 2022
Use the bandstop function, specifying 'ImpulseResponse','iir' for best results.
  3 commentaires
Luke McDevitt
Luke McDevitt le 4 Déc 2022
For anyone wondering this would be the new code:
(where 350 is the low cutoff and 6000 is the high cutoff)
[y,Fs] = audioread('africa-toto.wav');
bandstop(y,[350 6000],Fs,ImpulseResponse="iir",Steepness=0.95)
Star Strider
Star Strider le 4 Déc 2022
As always, my pleasure!

Connectez-vous pour commenter.

Plus de réponses (0)

Produits


Version

R2022b

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by