Loading [MathJax]/jax/output/SVG/jax.js
 
+0  
 
+1
1294
3
avatar

How can I solve the convolution for this example in steps? 


 Jun 17, 2017
 #1
avatar+9675 
+1

Any convolution is defined as
(fg)(t)=f(x)g(tx)dx=g(x)f(tx)dx

Trying to do it... Never done this type of question before.

(fh)(3)=f(x)h(3x)dx=h(x)f(3x)dx

But f(x) is only defined in {-1<=x<=1|xZ}

h(x) is only defined in {1<=x<=5|xZ}

So Idk what does it mean by integrating f(x)h(3-x) from infinity to negative infinity... How on Mars could I do that?

 Jun 18, 2017
 #2
avatar+169 
+1

The definition of the discrete convolution is:

 

fh[n]=m=f[m]g[nm].

 

We require the solution for the case n=3:

 

fh[3]=1m=1f[m]g[3m],

 

since f[m]=0,for: m1,0,1.

 

Our solution is:

 

fh[3]=f[1]g[4]+f[0]g[3]+f[1]g[2]=aI4+bI3+cI2.

.
 Jun 18, 2017
 #3
avatar+169 
+1

An intuitive approach of solving discrete convolution is sliding an inverted version of f over h. For n=0 we have:

 

m1012345h00I1I2I3I4I5fcba0000fh[0]00aI10000

 

the convolution product in this case is the product of each column.Now we slide f three places:

 

m1012345h00I1I2I3I4I5f000cba0fh[3]000cI2bI3aI40

 

and we have the same solution.

.
 Jun 18, 2017
edited by Honga  Jun 18, 2017
edited by Honga  Jun 18, 2017

1 Online Users

avatar