Hello,
I have some motion capture data that I need to filter before differentiating to get velocity and acceleration. I have been told to filter at a range of frequencies and then calculate the RMSD between the original and filtered data. I want to used a forth order low pass butterworth filter. I have managed to do so but do not understand how I then interpret the plot to decide on my optimal filtering frequency. I have seen some previous posts mention ‘fft’ but do not understand what this does and if I should be using it.

This is my function:
function plot_rmsd_vs_cutoff(data, fs, f_min, f_max, num_points)
% Plot RMSD against cut-off frequency for given data

% Data and sampling frequency
% data: input data
% fs: sampling frequency

% Frequency range
% f_min: minimum cut-off frequency
% f_max: maximum cut-off frequency
% num_points: number of points in the frequency range

% Create a vector of cut-off frequencies
cutoff_freqs = linspace(f_min, f_max, num_points);

% Initialize RMSD vector
rmsd_values = zeros(size(cutoff_freqs));

for i = 1:length(cutoff_freqs)
% Filter data
[b, a] = butter(4, cutoff_freqs(i)/(fs/2), ‘low’); % Example: 4th order lowpass Butterworth
filtered_data = filtfilt(b, a, data);

% Calculate RMSD
rmsd_values(i) = rms(data – filtered_data);
end

% Plot RMSD against cut-off frequency
plot(cutoff_freqs, rmsd_values);
xlabel(‘Cut-off Frequency (Hz)’);
ylabel(‘RMSD’);
title(‘RMSD vs Cut-off Frequency’);
end

This is how I have used it in my code:
% Call the function
plot_rmsd_vs_cutoff(data, fs, 1, 50, 50);

I have attached my data and the output I have got.
Many thanks for any help.Hello,
I have some motion capture data that I need to filter before differentiating to get velocity and acceleration. I have been told to filter at a range of frequencies and then calculate the RMSD between the original and filtered data. I want to used a forth order low pass butterworth filter. I have managed to do so but do not understand how I then interpret the plot to decide on my optimal filtering frequency. I have seen some previous posts mention ‘fft’ but do not understand what this does and if I should be using it.

This is my function:
function plot_rmsd_vs_cutoff(data, fs, f_min, f_max, num_points)
% Plot RMSD against cut-off frequency for given data

% Data and sampling frequency
% data: input data
% fs: sampling frequency

% Frequency range
% f_min: minimum cut-off frequency
% f_max: maximum cut-off frequency
% num_points: number of points in the frequency range

% Create a vector of cut-off frequencies
cutoff_freqs = linspace(f_min, f_max, num_points);

% Initialize RMSD vector
rmsd_values = zeros(size(cutoff_freqs));

for i = 1:length(cutoff_freqs)
% Filter data
[b, a] = butter(4, cutoff_freqs(i)/(fs/2), ‘low’); % Example: 4th order lowpass Butterworth
filtered_data = filtfilt(b, a, data);

% Calculate RMSD
rmsd_values(i) = rms(data – filtered_data);
end

% Plot RMSD against cut-off frequency
plot(cutoff_freqs, rmsd_values);
xlabel(‘Cut-off Frequency (Hz)’);
ylabel(‘RMSD’);
title(‘RMSD vs Cut-off Frequency’);
end

This is how I have used it in my code:
% Call the function
plot_rmsd_vs_cutoff(data, fs, 1, 50, 50);

I have attached my data and the output I have got.
Many thanks for any help. Hello,
I have some motion capture data that I need to filter before differentiating to get velocity and acceleration. I have been told to filter at a range of frequencies and then calculate the RMSD between the original and filtered data. I want to used a forth order low pass butterworth filter. I have managed to do so but do not understand how I then interpret the plot to decide on my optimal filtering frequency. I have seen some previous posts mention ‘fft’ but do not understand what this does and if I should be using it.

This is my function:
function plot_rmsd_vs_cutoff(data, fs, f_min, f_max, num_points)
% Plot RMSD against cut-off frequency for given data

% Data and sampling frequency
% data: input data
% fs: sampling frequency

% Frequency range
% f_min: minimum cut-off frequency
% f_max: maximum cut-off frequency
% num_points: number of points in the frequency range

% Create a vector of cut-off frequencies
cutoff_freqs = linspace(f_min, f_max, num_points);

% Initialize RMSD vector
rmsd_values = zeros(size(cutoff_freqs));

for i = 1:length(cutoff_freqs)
% Filter data
[b, a] = butter(4, cutoff_freqs(i)/(fs/2), ‘low’); % Example: 4th order lowpass Butterworth
filtered_data = filtfilt(b, a, data);

% Calculate RMSD
rmsd_values(i) = rms(data – filtered_data);
end

% Plot RMSD against cut-off frequency
plot(cutoff_freqs, rmsd_values);
xlabel(‘Cut-off Frequency (Hz)’);
ylabel(‘RMSD’);
title(‘RMSD vs Cut-off Frequency’);
end

This is how I have used it in my code:
% Call the function
plot_rmsd_vs_cutoff(data, fs, 1, 50, 50);

I have attached my data and the output I have got.
Many thanks for any help. cut off frequency, butterworth filter, filter, rmsd MATLAB Answers — New Questions

​