File:Circular convolution example.svg

Summary

Description
English: Circular convolution can be expedited by the FFT algorithm, so it is often used with an FIR filter to efficiently compute linear convolutions. These graphs illustrate how that is possible. Note that a larger FFT size (N) would prevent the overlap that causes graph #6 to not quite match all of #3.
Date
Source Own work
Author Bob K
Permission
(Reusing this file)
I, the copyright holder of this work, hereby publish it under the following license:
Creative Commons CC-Zero This file is made available under the Creative Commons CC0 1.0 Universal Public Domain Dedication.
The person who associated a work with this deed has dedicated the work to the public domain by waiving all of their rights to the work worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law. You can copy, modify, distribute and perform the work, even for commercial purposes, all without asking permission.

Category:CC-Zero#Circular%20convolution%20example.svg
Category:Self-published work
Other versions This file was derived from: Circular convolution example.png
Category:Derivative versions
SVG development
InfoField
 
The source code of this SVG is invalid due to an error.
 
This W3C-invalid vector image was created with LibreOffice.
Category:Invalid SVG created with LibreOffice#1000Circular%20convolution%20example.svg
Gnu Octave source
InfoField
click to expand

This graphic was created with the help of the following Octave script:

% Options
  frame_background_gray = true;

  if frame_background_gray
   graphics_toolkit("qt")         % has "insert text" option
%  graphics_toolkit("fltk")       % has cursor coordinate readout
   frame_background = .94*[1 1 1];
   d = 2;                         % amount to add to text sizes
  else
   graphics_toolkit("gnuplot")    % background will be white regardless of value below
   frame_background = .94*[1 1 1];
   d=0;
  endif

% (https://octave.org/doc/v4.2.1/Graphics-Object-Properties.html#Graphics-Object-Properties)
% Speed things up when using Gnuplot
  set(0, "DefaultFigureColor",frame_background)
  set(0, "DefaultAxesFontsize",10+d)   % size of numeric tick labels
  set(0, "DefaultTextFontsize",12+d)
  set(0, "DefaultAxesXtick",[])
  set(0, "DefaultAxesYtick",[])
  set(0, "DefaultLineLinewidth",1)
 
xmax = 3000;
%=======================================================
hfig = figure("position",[100 100 488 512], "color",frame_background);

x1 = .02;               % left margin
x2 = .02;               % right margin
y1 = .08;               % bottom margin for annotation
y2 = .08;               % top margin for title
dy = .04;               % vertical space between rows

width = 1-x1-x2;
height= (1-y1-y2-5*dy)/6; % space allocated for each of 6 rows

x_origin = x1;
y_origin = 1;           % start at top of graph area
%=======================================================
y_origin = y_origin -y2 -height;        % position of top row
% subplot() undoes all the "color" attempts above.  (gnuplot bug)
subplot("position",[x_origin y_origin width height])
L = 100;
f = ones(1,L)/L;
plot(-100:200-1, [zeros(1,100) f*L zeros(1,100)], "linewidth",2, "color","magenta")
xlim([-100 xmax]); ylim([0 2])
title("Circular convolution example", "fontsize",16)
text(100, 1.6, "h[n]")
%text(xmax/2, 0.4, '\leftarrow n \rightarrow')
 text(2500, 0.330, '\leftarrow n \rightarrow')

y_origin = y_origin -dy -height;
subplot("position",[x_origin y_origin width height])
a = [zeros(1,20) ones(1,L) zeros(1,300) 0.5*ones(1,100) zeros(1,1000-L-20-400)];
b = [zeros(1,1000-L-20) ones(1,L) zeros(1,20)];
a1 = [zeros(1,1000) a zeros(1,1000)];
b1 = [zeros(1,1000) b zeros(1,1000)];
plot(1:length(a1), a1, "color","blue", 1:length(a1), b1, "color","red")
xlim([0 xmax]); ylim([0 2])
text(200, 1.6, "X[n]")

y_origin = y_origin -dy -height;
subplot("position",[x_origin y_origin width height])
a1 = conv(a1,f);
b1 = conv(b1,f);
plot(1:length(a1), a1+b1, "color","green", "linewidth",2)
xlim([0 xmax]); ylim([0 2*max(a1)])
 text(200, 1.6, "X[n] * h[n]")
%text(200, 1.6, "X[n] ∗ h[n]", "interpreter","none") % requires PERL post-processor

y_origin = y_origin -dy -height;
subplot("position",[x_origin y_origin width height])
a = [a a a];
b = [b b b];
L = 1:length(a);
plot(L, a, "color","blue", L, b, "color","red")
xlim([0 xmax]); ylim([0 2.5])
set(gca,"xtick", [1000 2000]);
%set(gca,"xticklabel",["N" "2N"])
set(gca,"xticklabel",[]); text(981,-.5, "N"); text(1955,-.5, "2N")
text(200, 2.0, 'X_N[n]')

y_origin = y_origin -dy -height;
subplot("position",[x_origin y_origin width height])
a1 = conv(a,f);
b1 = conv(b,f);
b1(1:90) = b1(3000+[1:90]);
L = 1:length(a1);
plot(L,a1,"color","blue", L,b1, "color","red")
xlim([0 xmax]); ylim([0 2*max(a1)])
text(200, 1.6, 'components of X_N[n] * h[n]') % can't use "interpreter","none" here

y_origin = y_origin -dy -height;
subplot("position",[x_origin y_origin width height])
c = a1+b1;
L = length(c);
k=1100;
plot(1:k, c(1:k), "color","red", k+(1:900), c(k+(1:900)), "color","green",...
        "linewidth",2, (k+900+1):xmax, c((k+900+1):xmax), "color","red")
xlim([0 xmax]); ylim([0 2*max(a1+b1)])
text(200, 1.6, 'X_N[n] * h[n]') % can't use "interpreter","none" here

 text(1263, -.6, "X[n] * h[n]", "fontsize",16)
%text(1274, -.6, "X[n] ∗ h[n]", "interpreter","none", "fontsize",16) % requires PERL post-processor

% After a call to annotation(), the cursor coordinates change to the units used below.
annotation("line", [.367 .367], [.113 .022])
annotation("line", [.664 .664], [.113 .022])
Category:Convolution Category:Created with GNU Octave Category:Images with Octave source code Category:Images with Gnuplot source code
Category:CC-Zero Category:Convolution Category:Created with GNU Octave Category:Derivative versions Category:Images with Gnuplot source code Category:Images with Octave source code Category:Invalid SVG created with LibreOffice Category:Pages using deprecated source tags Category:Self-published work