logo

Convolution Support 📂Fourier Analysis

Convolution Support

Theorem

Given two sets of real numbers A,BA, B, we define A+BA + B as follows.

A+B:={a+b:aA,bsuppB}A + B := \left\{ a + b : \forall a \in A, \forall b \in \supp B \right\}

For two functions f,gf, g, the following holds.

suppfgsuppf+suppg\supp f \ast g \subset \supp f + \supp g

Here, supp\supp is the function’s support, and \ast is the convolution.

Proof1

Assume xsuppf+suppgx \notin \supp f + \supp g. Then, for any yy chosen, f(y)g(xy)=0f(y)g(x-y)=0 occurs.

  • Case 1 ysuppfy \in \supp f

    In this case, xysuppgx - y \notin \supp g holds. If we assume xysuppgx - y \in \supp g,

    suppf+suppg(xy)+y=xsuppf+suppg \supp f + \supp g \ni (x - y) + y = x \notin \supp f + \supp g

    this leads to a contradiction. Therefore, xysuppgx - y \notin \supp g and g(xy)=0g(x-y) = 0 hold.

  • Case 2 ysuppfy \notin \supp f

    In this case, f(y)=0f(y) = 0 holds.

Therefore, if xsuppf+suppgx \notin \supp f + \supp g, then f(y)g(xy)dy=fg(x)=0\displaystyle \int f(y)g(x - y) dy = f \ast g(x) = 0, and so xsuppfgx \notin \supp f \ast g. Thus,

suppfgsuppf+suppg \supp f \ast g \subset \supp f + \supp g