37.3. FOURIER TRANSFORM 755

Then for f ∈ L1, let fR (t)≡X[−R,R] (t) f (t). Then for R large,∫∞

−∞

| f (t)− fR (t)|dt =∫

R| f (t)|dt +

∫ −R

−∞

| f (t)|dt < ε

Now fR is Riemann integrable and so there is a step function s(t) = ∑ni=1 aiXIi (t) such that

|s(t)| ≤ | fR (t)| and ∫ R

−R| fR (t)− s(t)|dt =

∫∞

−∞

| fR (t)− s(t)|dt < ε

This follows from the definition of the Riemann integral as a limit of integrals of stepfunctions, details are left for you. Therefore,∫

−∞

|s(t)− f (t)|dt < 2ε

Now ∣∣∣∣∫ ∞

−∞

f (t)sin(rt)dt∣∣∣∣ ≤ ∫

−∞

|( f (t)− s(t))sin(rt)|dt +∣∣∣∣∫ ∞

−∞

s(t)sin(rt)dt∣∣∣∣

≤ 2ε +

∣∣∣∣∫ ∞

−∞

s(t)sin(rt)dt∣∣∣∣ (37.1)

It remains to verify that limr→∞

∫∞

−∞s(t)sin(rt)dt = 0. Since s(t) is a sum of scalars times

XI for I an interval, it suffices to verify that limr→∞

∫∞

−∞X[a,b] (t)sin(rt)dt = 0. However,

this integral is just ∫ b

asin(rt)dt =

−1r

cos(rb)+1r

cos(ra)

which clearly converges to 0 as r→ ∞. Therefore, for r large enough, 37.1 implies∣∣∣∣∫ ∞

−∞

f (t)sin(rt)dt∣∣∣∣< 3ε

Since ε is arbitrary, this shows that 3. holds. ■

Definition 37.3.3 The following notation will be used assuming the limits exist.

limr→0+

g(x+ r)≡ g(x+) , limr→0+

g(x− r)≡ g(x−)

Theorem 37.3.4 Suppose that g∈ L1 (R) and that at some x, g is locally Holder continuousfrom the right and from the left. This means there exist constants K,δ > 0 and r ∈ (0,1]such that for |x− y|< δ ,

|g(x+)−g(y)|< K |x− y|r (37.2)

for y > x and|g(x−)−g(y)|< K |x− y|r (37.3)

for y < x. Then

limr→∞

∫∞

0

sin(ur)u

(g(x−u)+g(x+u)

2

)du

=g(x+)+g(x−)

2.

37.3. FOURIER TRANSFORM 755Then for f €L!, let fr(t) = 2\_rp (t) f(t). Then for R large,[iro-molar=[irolas fo rolar<eNow fr is Riemann integrable and so there is a step function s(t) = °"_, a; 27, (t) such thatIs(t)| < [fr ()| andR co[lin -sOlar= [be -s@lat<e—R —ooThis follows from the definition of the Riemann integral as a limit of integrals of stepfunctions, details are left for you. Therefore,J b@-r@lar <2Now[reosin (rt) dtoo}[ire -scaysincro|ar| | s(t) sin (rt) dtIA—ooIA2€+ [ s(t) sin (rt) dt (37.1)It remains to verify that lim,_,.. [™,, s (t) sin (rt) dt = 0. Since s(t) is a sum of scalars times2 for J an interval, it suffices to verify that lim,.. f~., jas) (t) sin (rt) dt = 0. However,this integral is justb 1| sin (rt) dt = — cos (rb) + — cos (ra)a r rwhich clearly converges to 0 as r — oo. Therefore, for r large enough, 37.1 implies< 3€[fo sin (rt) dtSince € is arbitrary, this shows that 3. holds. MfDefinition 37.3.3 The following notation will be used assuming the limits exist.li = li —rj= —jim g(x+r) = g(xt), lim g(x—r) = g(x)Theorem 37.3.4 Suppose that g € L' (IR) and that at some x, g is locally Holder continuousfrom the right and from the left. This means there exist constants K,6 > 0 and r € (0, 1]such that for |x —y| < 6,Ig (a+) —g(y)|<K|x—yI' (37.2)for y >x andIg (x-) —g(y)| <K|x—y’ (37.3)for y <x. Thenlim —roo JC Jo u2 sin (ur) (oe au