Writing a loop to compute Indices

Please help me to write a loop to compute the indices (starting and stopping) – . Assume that your time history has 5,000 points. Use a record length of 2^10 = 1,024 points. List both sets of indices, preferably in a matrix

a. For the case of 0% overlap

b. For the case of 50% overlap

Answer 1

[Mathieu NOE]

hello

see basic demo below

select which option you want

overlap = 0*buffer_size; % overlap expressed in samples (a. For the case of 0% overlap)

% overlap = 0.5*buffer_size; % overlap expressed in samples (b. For the case of 50% overlap)

% dummy data (a noisy sinusoid)
% NB : data must be column oriented 
n=5000;
t= (1:n)';
y = sin(2*pi*t./max(t))+0.1*rand(size(t));


% choose the amount of overlap
buffer_size = 1024; 
overlap = 0*buffer_size; % overlap expressed in samples (a. For the case of 0% overlap)
% overlap = 0.5*buffer_size; % overlap expressed in samples (b. For the case of 50% overlap)

%%%% main loop %%%%
[tt,yy,start_index,stop_index] = my_movmean(t,y,buffer_size,overlap);

plot(t,y,tt,yy,'*-r');
title('data');
legend('raw data','1024 samples mean');
xlabel('Time(s)');

% display start_index,stop_index in command window 
start_index

start_index = 1×4

1 1025 2049 3073

<mw-icon class=""></mw-icon>

<mw-icon class=""></mw-icon>

stop_index

stop_index = 1×4

1024 2048 3072 4096

<mw-icon class=""></mw-icon>

<mw-icon class=""></mw-icon>

%%%%%%%%%% my functions %%%%%%%%%%%%%%
function [new_time,data_out,start_index,stop_index] = my_movmean(t,data_in,buffer_size,overlap)

% NB : buffer size and overlap are integer numbers (samples)
% data (in , out) are 1D arrays (vectors)

shift = buffer_size-overlap;    % nb of samples between 2 contiguous buffers  
[samples,~] = size(data_in);
nb_of_loops = fix((samples-buffer_size)/shift +1);

    for k=1:nb_of_loops
        start_index(k) = 1+(k-1)*shift;
        stop_index(k) = min(start_index(k)+ buffer_size-1,samples);
        x_index(k) = round((start_index(k)+stop_index(k))/2);  
        data_out(k,:) = mean(data_in(start_index(k):stop_index(k),:),1,'omitnan');  % 
        
    end

new_time = t(x_index); % time values are computed at the center of the buffer

end

Answer 2

[Maneet Kaur Bagga]

Hi,

As per my understanding, you have given a signal of length 5000 and a window length of 1024. You want to compute the start and stop indices of segments for 0% and 50% overlap.

You can refer to the following MathWorks documentation to segment time series using "pwelch" function.

https://in.mathworks.com/help/signal/ref/pwelch.html

The segmenting data involves defining a hop size for which you can refer to the following code snippet:

% Parameters
N = 5000;        % Total number of points
L = 1024;        % Record length
overlap_0 = 0;   % Overlap for case (a)
overlap_50 = 512; % 50% overlap for case (b)

% Function to calculate indices
function idx_matrix = compute_indices(N, L, overlap)
    hop = L - overlap;
    start_idx = 1:hop:(N - L + 1);  % Start indices of each segment
    stop_idx = start_idx + L - 1;  % End indices of each segment
    idx_matrix = [start_idx.', stop_idx.'];  % Nx2 matrix
end

Hope this helps!