Reputation: 958
as we know that another function convolve with delta function, the output will same as the function itself, for example
f(n) * δ(n) = f(n)
For example
n=-10:10;
h=[n>=0];
i=[(n-4)>=0];
f=h-i;
I get the function f(n)
, but how should I code the δ(n)
in order to get a stem plot of f(n) * δ(n)
?
One more question, if we convolve one function itself repeatedly, what actually is happening? I try doing some chain convolution in Matlab like conv(conv(f,f,'same'),f,'same');
, the stem plot are varying but i cant figure out the reason.
Please assist, thank you.
Upvotes: 1
Views: 3236
Reputation: 46
Convolution with delta shifts input by k-1 places:
delta=zeros(n); delta(k)=1 %k'th canonical basis vector
In higher dimensions, delta(ind(1),ind(2),..)=1
delta=zeros(dims); delta(num2cell(inds){:})=1 %canonical basis of matrices
Repeated convolution of the probability distribution of a random variable X results in the probability distribution of the sum of 'n' indpendent identically distributed copies X1+X2+..+Xn.
Upvotes: 0
Reputation: 2080
In discrete math the impulse is δ(0) = 1
, δ(n) = 0
for n != 0
.
If you convolve a function with itself, you just do exactly that. There is nothing special about it.
Convolve it with the mirrored(at the y-axis) function of itself would be an auto-correlation.
Upvotes: 1