Home > matlab code, youtube demo code > illustrate_xcorr – code for cross correlation demos

illustrate_xcorr – code for cross correlation demos

% This function illustrates the cross correlation process in action
% Usage:
%           fs = 1000;
%             T = 1/fs;
%             N = 500; % desired length of signal
%             t = [0:N-1]*T; %time vector 
%             f1 = 8; f2=f1*2; 
%             x = sin(2*pi*f1*t) + sin(2*pi*f2*t);
%           % To step though each sample use the following:
%           illustrate_xcorr(x,x)
%           % to step through using 50 steps use:
%           illustrate_xcorr(x,x, 50)
function illustrate_xcorr(x, y, varargin)
if(length(x) > length(y))
    y(end+1:length(x)) = 0; %zero pad so the signals are the same length
    x(end+1:length(y)) = 0; %zero pad so the signals are the same length

    num_steps = 2*length(x)-1;
if(nargin ==3)
    arg = varargin{1};
        num_steps = ceil(abs(arg));
if(nargin > 3)
    error('See help on this function to see how to use it properly')

[rxy lags] = xcorr(x,y); %cross correlate signals

disp('The signal being autocorrelated is shown in blue (two instances)')
disp('As you hit the space bar the lower plot will move into different lag positions')
disp('The correlation function shown in red is updated for each lag position');
disp('keep pressing the space bar to step through the illustration ...');

plot_width = 0.3; plot_height = 0.25;

top_ax_h = subplot(3,1,1);
axis tight
set(top_ax_h, 'visible','off', 'units', 'normalized')
set(top_ax_h,'position', [0.5-plot_width/2 5/6-plot_height/2 plot_width plot_height])

mid_ax_h = subplot(3,1,2);
axis tight
set(mid_ax_h, 'visible','off', 'units', 'normalized')
set(mid_ax_h,'position', [0.5-plot_width/2-plot_width 5/6-3*plot_height/2-0.01 plot_width plot_height])

bottom_ax_h = subplot(3,1,3);
corr_h = plot(lags,rxy,'r');
axis tight
set(bottom_ax_h,'units', 'normalized','Ytick',[])
set(bottom_ax_h,'position', [0.5-plot_width*3/2 0.2-plot_height/2 plot_width*3 plot_height])
set(corr_h, 'Ydata', ones(1, length(rxy))*NaN); %clear the correlation funciton plot once its set up

normalised_shift_size = 2*plot_width/(num_steps-1);
corr_seg_len = length(rxy)/num_steps;
for k = 1 : num_steps
    if(k > 1)
        new_pos = get(mid_ax_h,'position') + [normalised_shift_size 0 0 0];
        set(mid_ax_h,'position', new_pos);
    set(corr_h, 'Ydata', [rxy(1:round(corr_seg_len*k)) ones(1,length(rxy)-round(corr_seg_len*k))*NaN])

  1. September 16, 2020 at 7:07 am

    Hey! I saw these videos on Youtube while searching for correlation of signals and this was beautifully and fundamentally explained in them. Just wanted to drop a thanks, keep making more videos. Also wish the codes were python instead of matlab. Still, Thanks for sharing these and best wishes!

  1. No trackbacks yet.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: