Fourier Graphs of my lowpass and high pass filtered signals are looking weird

Hello everyone,

I am currently doing a project in MATLAB and I am using a lowpass filter in a signal with a cutoff frequency of 30Hz and then I find the fourier transform of it and I graph it. I want to know if the actual graph is right or am I doing something wrong? I am going to attach the part of the code for the filters and plots.

This is the plot that I get for the low pass filter:

It shouldn't go to 0 after 30 Hz? or around 30 Hz?

And for the high pass filter:

I get it to be 0 around until like 190 Hz, but why is it not 270Hz or closer? I did use smaller frequencies for the HPF and they were more accurate until when they stayed 0.

%LPF 30Hz
y_LPF_30 = lowpass(y,30,Fs);
yf_LPF_30 = fft(y_LPF_30);
% 30Hz
%set up
%% Compute the two-sided spectrum P2. Then compute the single-sided spectrum P1 based on P2 and the even-valued signal length L.
P2_LPF30 = abs(yf_LPF_30/L);
P1_LPF30 = P2_LPF30(1:L/2+1);
P1_LPF30(2:end-1) = 2*P1_LPF30(2:end-1);
figure (5)
%plot
subplot(3,2,1);
plot(f,P1_LPF30) 
title('Single-Sided Amplitude Spectrum of LPF 30Hz')
xlabel('f (Hz)')
ylabel('|P1(f)|')

%HPF 270Hz
y_HPF_270 = highpass(y,270,Fs);
yf_HPF_270 = fft(y_HPF_270);
%270Hz
%set up
P2_HPF270 = abs(yf_HPF_270/L);
P1_HPF270 = P2_HPF270(1:L/2+1);
P1_HPF270(2:end-1) = 2*P1_HPF270(2:end-1);
%plot
subplot(3,2,6);
plot(f,P1_HPF270) 
title('Single-Sided Amplitude Spectrum of HPF 270Hz')
xlabel('f (Hz)')
ylabel('|P1(f)|')

[Nicolas Picon]

That typo explains the high pass filter question, oops.

But I don't know what happens with the LPF.

The code that I used for f is:

[y,Fs] = audioread('myvoice.m4a');
L = length(y);
f = Fs*(0:(L/2))/L;

[Nicolas Picon]

Also, the original sound signal graph is this:

And the Fourier Transform graph of the original signal is this:

I don't know if this helps

Answer 1

[Ridwan Alam]

Last update:

Running your Lowpass code:

%LPF 30Hz
y_LPF_30 = lowpass(y,30,Fs);
yf_LPF_30 = fft(y_LPF_30);
% 30Hz
%set up
%% Compute the two-sided spectrum P2. Then compute the single-sided spectrum P1 based on P2 and the even-valued signal length L.
P2_LPF30 = abs(yf_LPF_30/L);
P1_LPF30 = P2_LPF30(1:L/2+1);
P1_LPF30(2:end-1) = 2*P1_LPF30(2:end-1);
figure
%plot
plot(f,P1_LPF30) 
title('Single-Sided Amplitude Spectrum of LPF 30Hz')
xlabel('f (Hz)')
ylabel('|P1(f)|')

gives this output:

Update:

For the low-pass try adding some properties such as:

y_LPF_30 = lowpass(y,30,Fs,'StopbandAttenuation',100);

Also, please show us the plots generated by this:

lowpass(y,30,Fs,'StopbandAttenuation',100);

lowpass(___) with no output arguments plots the input signal and overlays the filtered signal.

Hope it helps.

Earlier:

The filters look ok. some example data would be helpful to debug.

Also, it might just be a typo, but you wrote: P2_HPF270 = abs(yf_HPF_220/L);

[Nicolas Picon]

So, I tried adding that and I get this for

lowpass(y,30,Fs,'StopbandAttenuation',100);

And for the Fourier Transform graph, I get:

[Ridwan Alam]

Can you please share the y and Fs as a .mat file?

[Nicolas Picon]

Here are the files.

[Ridwan Alam]

After running

lowpass(y,30,Fs,'Steepness',0.95,'StopbandAttenuation',100)

the result looks like this:

[Ridwan Alam]

Looking at the fft: