I'm not sure if I can discuss some less math-majored problems here(i'm not a math-majored students), I'm studying convolution in the Fourier transform, and I see a sentence that convolution can be understood as using one function to smooth and average another function. I don't understand why it's smoother (I can probably feel some of it through some cases and through some thoughts, but I'd like to get a more robust explanation), and as I think about it deeply, I come across another question: how to define smoothness?
I have some rough explanations of g*f, such as using a sharp centered normalized g(x), and then I can find that the highest point of f(x) will fall, and the lowest point will rise, so the slope becomes smaller, so it is more smoother. However, I always feel the explanations are not strong enough, and I hope to have a better explanation.
or may be you have some excellent materials about convolutions with smooth and average can recommend to me?