Removing the noise/filter

So I am trying to remove the noise or fillter the data to reduce/remove the unnecessary noise.

This is the code that I am using. I tried different method of filltering but it doesn't work.

% Load the data 
data = importdata('default.txt'); % Load data from the file

% Time vs Channel
time = data(:, 1); % Extract the time column
channel = data(:, 2); % Extract the channel column

% Plot the data
figure; 
plot(time, channel); 
xlabel('Time (s)');
ylabel('Channel (V)'); 
title('Data Plot'); 
grid on;
legend('Channel Data');

Answer 1

[Star Strider]

The signal has broadband noise, so ffrequency-selective filters will not work well to limit that noise. The best approach fro broadband noise is to use the Savitzky-Golay filter (sgolayfilt) or wavelket denoising.

Using sgolayfilt —

T1 = readtable('default.txt')

T1 = 16384×2 table
       Var1           Var2   
    ___________    __________

    -0.00011384     0.0042297
    -0.00011382     0.0053817
     -0.0001138     0.0064339
    -0.00011378     0.0064667
    -0.00011376     0.0056993
    -0.00011374     0.0075507
    -0.00011372     0.0086688
     -0.0001137     0.0057182
    -0.00011368     0.0031314
    -0.00011366     0.0021281
    -0.00011364    0.00099381
    -0.00011362    -0.0027928
     -0.0001136    -0.0059469
    -0.00011358    -0.0033796
    -0.00011356    0.00092691
    -0.00011354     0.0045465

t = T1{:,1};
s = T1{:,2};

figure
plot(t, s)
xlabel('Time')
ylabel('Amplitude')
title('Original')

Ts = mean(diff(t))

Ts = 2.0000e-08

Fs = 1/Ts;
Fn = Fs/2

Fn = 2.5000e+07

% Tstd = std(diff(t))

L = size(T1,1);
NFFT = 2^nextpow2(L);
FTs = fft((s - mean(s)).*hann(L), NFFT)/L;
Fv = linspace(0, 1, NFFT/2+1)*Fn;
Iv = 1:numel(Fv);

figure
plot(Fv, abs(FTs(Iv))*2)
grid
xlabel('Frequency')
ylabel('Magnitude')
title('Original')

sf = sgolayfilt(s, 3, 101);

figure
plot(t, sf)
xlabel('Time')
ylabel('Amplitude')
title('Filtered')

FTsf = fft((sf - mean(sf)).*hann(L), NFFT)/L;

figure
plot(Fv, abs(FTsf(Iv))*2)
grid
xlabel('Frequency')
ylabel('Magnitude')
title('Filtered')

Experiment to get the result you want.

.

[Abdullah Alsayegh]

Is there possibility of making the noise almost a stright line and having the presented rised Amplitude be presented at the same level.

[Star Strider]

Probably not, because no filter is that selective. The best you could hope for is to use the envelope function to identify the ‘peaks’ (or whatever you want to call them), and then force everything else between them to zero. This is artificial, however it is possible —

T1 = readtable('default.txt')

T1 = 16384×2 table
       Var1           Var2   
    ___________    __________

    -0.00011384     0.0042297
    -0.00011382     0.0053817
     -0.0001138     0.0064339
    -0.00011378     0.0064667
    -0.00011376     0.0056993
    -0.00011374     0.0075507
    -0.00011372     0.0086688
     -0.0001137     0.0057182
    -0.00011368     0.0031314
    -0.00011366     0.0021281
    -0.00011364    0.00099381
    -0.00011362    -0.0027928
     -0.0001136    -0.0059469
    -0.00011358    -0.0033796
    -0.00011356    0.00092691
    -0.00011354     0.0045465

t = T1{:,1};
s = T1{:,2};

[et,eb] = envelope(s, 35, 'peaks');

threshold = 0.01;                                   % Set 'threshold' To Produce The Desired Result
Lv = (et >= threshold) & (eb <= -threshold);
% Q = nnz(Lv)

figure
plot(t, s, 'DisplayName','Signal')
hold on
plot(t, et, 'DisplayName','Upper Envelope')
plot(t, eb, 'DisplayName','Lower Envelope')
hold off
grid
xl = xlim;
yl = ylim;
% xlim([-1.25 -0.5]*1E-4)
% ylim([-0.05 0.05])
xlabel('Time')
ylabel('Amplitude')
title('Original')

sf = s;                                             % Duplicate Signal Vector
sf(~Lv) = 0;                                        % Imposae 'threshold' Filter

figure
plot(t, sf)
grid
xlim(xl)
ylim(yl)
xlabel('Time')
ylabel('Amplitude')
title('Threshold Filtered')

This is not the sort of filtering I generally recommend.

EDIT — Cleaned code.

.