How can I apply a lowpass filter samplewise in my code?

I have a real time plugin that does a little bit of reverberation. After each delay line in v(n) I want to apply a lowpass filter to cut out the high frequencies. How can I do that?

My code below:

function out = process(plugin, in) 
            out = zeros(size(in));
            
            for i = 1:size(in,1)
                % Summieren der L/R - Kan�le
                inL = in(i,1);
                inR = in(i,2);
                inSum = (inL + inR)/2;
                plugin.buffInput(plugin.pBuffInput + 1) = inSum;
                
                % loop over delay lines
                for n=1:plugin.N
                    % d_n = gain * delayed v_n
                    for k=1:plugin.N
                        plugin.d(k) = plugin.g(k) * plugin.buffDelayLines(k, mod(plugin.pBuffDelayLines + plugin.m(k), plugin.maxDelay +1) + 1);
                    end
                    % f_n = A(n,:) * d'
                    plugin.f(n) = plugin.A(n,:) * plugin.d(:);
                    % v_n with pre delay
                    plugin.v(n) = plugin.b(n) * plugin.buffInput(mod(plugin.pBuffInput + plugin.preDelayS, (plugin.maxPreDelay * plugin.fs + 1)) + 1) ...
                        + plugin.f(n);
                    plugin.buffDelayLines(n, plugin.pBuffDelayLines + 1) = plugin.v(n);

                    % output lines
                    plugin.s(n) = plugin.c(n) * plugin.d(n);
                
                    out(i,:) = out(i,:) + real(plugin.s(n));
                    
                end

                % Assign to output
                out(i,1) = plugin.mix/100 * out(i,1) + (1.0 - plugin.mix/100) * in(i,1);
                out(i,2) = plugin.mix/100 * out(i,2) + (1.0 - plugin.mix/100) * in(i,2);

                calculatePointer(plugin);
            end 
        end

Answer 1

[Drishti]

Hi Muhsin,

I understand that you are trying to implement a low pass filter to cut out the high frequencies.

To include the low-pass filter, refer to the implemented code:

% v_n with pre delay
rawVn = plugin.b(n) * plugin.buffInput(mod(plugin.pBuffInput + plugin.preDelayS, (plugin.maxPreDelay * plugin.fs + 1)) + 1) ...
                + plugin.f(n);
            
% Apply low-pass filter
plugin.v(n) = alpha * rawVn + (1 - alpha) * prevY(n);
prevY(n) = plugin.v(n);
plugin.buffDelayLines(n, plugin.pBuffDelayLines + 1) = plugin.v(n);

To achieve this, I have made certain assumptions which includes ‘cuttoffFreq’ and ‘alpha’ parameters as mentioned below:

% Define the cutoff frequency and calculate alpha
 cutoffFreq = 100; % Example cutoff frequency in Hz
 alpha = (2 * pi * cutoffFreq) / (plugin.fs + 2 * pi * cutoffFreq);
% Initialize the previous output for the filter
 prevY = zeros(plugin.N, 1);

I hope this helps in applying the low pass filter in the provided code.

[Muhsin Zerey]

Hi Thanks very much for your help! Do you know how I can apply a butterworth first order as a lowpass filter?

Answer 2

[Image Analyst]

One way is movmean.

[Muhsin Zerey]

how is that possible?

[Image Analyst]

You just pass your signal to it. The wider the window, the more samples are included in your average and the smoother the signal will be. Smoothing a signal (replacing elements by the local average) is a low pass filter operation. It's the same thing as convolution or Fourier filtering.

[Muhsin Zerey]

How can I apply it to my code tho? I will need the previous and future next sample to maake an average.