286 CHAPTER 11. FUNDAMENTAL TRANSFORMS

Proof: An application of Minkowski’s inequality to Borel representatives yields(∫Rn

(∫Rn|h(x−y)| | f (y)|dy

)2

dx

)1/2

≤ ∥ f∥1 ∥h∥2 . (11.16)

Hence∫|h(x−y)| | f (y)|dy < ∞ a.e. x and x→

∫h(x−y) f (y)dy is in L2 (Rn). Let Er ↑

Rn, mn (Er)< ∞. Thus,hr ≡XEr h ∈ L2 (Rn)∩L1 (Rn),

and letting φ ∈ G ,∫

F (hr ∗ f )(φ)dx

≡∫

(hr ∗ f )(Fφ)dx

= (2π)−n/2∫ ∫ ∫

hr (x−y) f (y)e−ix·tφ (t)dtdydx

= (2π)−n/2∫ ∫ (∫

hr (x−y)e−i(x−y)·tdx)

f (y)e−iy·tdyφ (t)dt

=∫

(2π)n/2 Fhr (t)F f (t)φ (t)dt.

Since φ is arbitrary and G is dense in L2 (Rn),

F (hr ∗ f ) = (2π)n/2 FhrF f .

Now by Minkowski’s Inequality, hr ∗ f → h ∗ f in L2 (Rn) and also it is clear that hr → hin L2 (Rn) ; so, by Plancherel’s theorem, you may take the limit in the above and concludeF (h∗ f )= (2π)n/2 FhF f . The assertion for F−1 is similar and 11.15 follows from 11.16. ■

11.6.5 The Schwartz ClassThe problem with G is that it does not contain C∞

c (Rn). I have used it in presenting theFourier transform because the functions in G have a very specific form which made sometechnical details work out easier than in any other approach I have seen. The Schwartzclass is a larger class of functions which does contain C∞

c (Rn) and also has the same niceproperties as G . The functions in the Schwartz class are infinitely differentiable and theyvanish very rapidly as |x|→∞ along with all their partial derivatives. This is the descriptionof these functions, not a specific form involving polynomials times e−α|x|2 . To describe thisprecisely requires some notation.

Definition 11.6.21 f ∈S, the Schwartz class, if f ∈C∞(Rn) and for all positiveintegers N, ρN( f )< ∞ where

ρN( f ) = sup{(1+ |x|2)N |Dα f (x)| : x ∈ Rn , |α| ≤ N}.

Thus f ∈S if and only if f ∈C∞(Rn) and

sup{|xβ Dα f (x)| : x ∈ Rn}< ∞ (11.17)

for all multi indices α and β .

286 CHAPTER 11. FUNDAMENTAL TRANSFORMSProof: An application of Minkowski’s inequality to Borel representatives yields> 1/2(/,.(/.imssiirie) i) < [Illi lll. (11.16)Hence f{ |h(x—y)||f (y)|dy < 0 ae. x and x > [h(x—y) f (y) dy is in L? (R"). Let E, tR", my (E,) <0. Thus,hy = 2z,h € VL? (R") OL! (R"),and letting @ € Y, [ F (h, * f) (@) dx[Gon(royax= aay? ff fn, (x-y)f(y)e™0 (0) atdyas= canyn? [ [ ( [ice yet Max) rye ayo (ar_ / (27)"/? Fh, (t) Ff (t) 6 (t)dr.Since @ is arbitrary and Y is dense in L? (R”),F (h,* f) = (20)"? Fh, Ff.Now by Minkowski’s Inequality, h, * f + h» f in L? (R”") and also it is clear that h, > hin L? (IR"); so, by Plancherel’s theorem, you may take the limit in the above and concludeF (hx f)= (2) FhF f. The assertion for F~! is similar and 11.15 follows from 11.16.11.6.5 The Schwartz ClassThe problem with Y is that it does not contain C? (IR”). I have used it in presenting theFourier transform because the functions in Y have a very specific form which made sometechnical details work out easier than in any other approach I have seen. The Schwartzclass is a larger class of functions which does contain C? (IR") and also has the same niceproperties as Y. The functions in the Schwartz class are infinitely differentiable and theyvanish very rapidly as |x| > ce along with all their partial derivatives. This is the descriptionof these functions, not a specific form involving polynomials times e~“!*!. To describe thisprecisely requires some notation.Definition 11.6.21 f €G, the Schwartz class, if f € C*(R") and for all positiveintegers N, Py(f) < % wherePy(f) =sup{(1+|x|")"|D%f(x)|:x €R", |a| <N}.Thus f € © if and only if f € C*(R") andsup{|x? D® f(x)|:x € R"} <0 (11.17)for all multi indices a and B.