Monday, August 17, 2009


I wrote a post about convolution in my other blog, but I'll write here how to use the convolution in Scilab.

The convolution is a operation with two functions defined as:

The function in Scilab that implements the convolution is convol(.).

Let's do the test: I'll convolve a cosine (five periods) with itself (one period):

N1 = 100;
N2 = 20;

n1 = 1:N1;
n2 = 1:N2;

w = %pi/10;

f1 = cos(w*n1);
f2 = cos(w*n2);

y = convol(f1, f2);

scf(); plot(f2);
scf(); plot(y);

The result is:

The top left graph is f1, the top right is f2 and the bottom graph is y.

Again, if anyone wants, I can write about discrete convolution properties.


naza sundae said...

Yes, i am want it ;p
Discrete Convolution rawks ~~

Younes said...


You seem to be quite good using Scilab. If you have a few minutes, i'd like to ask you some questions about the convolution via Fourier transform.

Thank you!