The convolution theorem - basic problem with the formula

63 Views Asked by At

I have a formular for the convolution theorem, and read several chapters in several scripts about it. This is the formula:

$(f*g)(x)=\int_{\mathbb{R}^d}f(x-y)g(y)dy$

However much I read, I cannot figure out where exactly the y comes from.

So, if I had an excercise where I have to convolute to functions, or one with itself, where do I put the y? I have found examples, but mostly they just fill in the formula and in the end the y just miraculously vanishes through some clever crossing out.

But I want to convolute to conditional functions (like, though not quite so simple, the Kronecker $\delta$. And I am just confused.

What exactly is the y, and is there anything its value depends on? if so, what?

Thanks

To the comments:

Say I have a function which gives f(x) = 1 if x<0 and f(x) = 0 in every other case.

I now want to convolute f with itself, so f*f.

I know there are different cases but which?

I just do not know what to put into the formula.

Say for instance the case x<0.

Then I put:

$ =\int_{\mathbb{R}^d}f(x-y)f(y)dy$ =

$ =\int_{\mathbb{R}^d}(f(x)-f(y))f(y)dy$

$ =\int_{\mathbb{R}^d}(1-f(y)f(y))dy$ and that is it?

Is there really no way to figure out whether x-y is greater zero, because the y does not really exist?