This action might not be possible to undo. Are you sure you want to continue?

BooksAudiobooksComicsSheet Music### Categories

### Categories

### Categories

Editors' Picks Books

Hand-picked favorites from

our editors

our editors

Editors' Picks Audiobooks

Hand-picked favorites from

our editors

our editors

Editors' Picks Comics

Hand-picked favorites from

our editors

our editors

Editors' Picks Sheet Music

Hand-picked favorites from

our editors

our editors

Top Books

What's trending, bestsellers,

award-winners & more

award-winners & more

Top Audiobooks

What's trending, bestsellers,

award-winners & more

award-winners & more

Top Comics

What's trending, bestsellers,

award-winners & more

award-winners & more

Top Sheet Music

What's trending, bestsellers,

award-winners & more

award-winners & more

Welcome to Scribd! Start your free trial and access books, documents and more.Find out more

**Richard S. Laugesen January 9, 2009
**

*

c 2009, Richard S. Laugesen (Laugesen@illinois.edu). This work is licensed under the Creative Commons Attribution–Noncommercial–Share Alike 3.0 Unported License. To view a copy of this license, visit http://creativecommons. org/licenses/by-nc-sa/3.0/.

* Copyright

2

Preface

A textbook presents more than any professor can cover in class. In contrast, these lecture notes present exactly* what I covered in Harmonic Analysis (Math 545) at the University of Illinois, Urbana–Champaign, in Fall 2008. The ﬁrst part of the course emphasizes Fourier series, since so many aspects of harmonic analysis arise already in that classical context. The Hilbert transform is treated on the circle, for example, where it is used to prove Lp convergence of Fourier series. Maximal functions and Calder´n– o Zygmund decompositions are treated in Rd , so that they can be applied again in the second part of the course, where the Fourier transform is studied. Real methods are used throughout. In particular, complex methods such as Poisson integrals and conjugate functions are not used to prove boundedness of the Hilbert transform. Distribution functions and interpolation are covered in the Appendices. I inserted these topics at the appropriate places in my lectures (after Chapters 4 and 12, respectively). The references at the beginning of each chapter provide guidance to students who wish to delve more deeply, or roam more widely, in the subject. Those references do not necessarily contain all the material in the chapter. Finally, a word on personal taste: while I appreciate a good counterexample, I prefer spending class time on positive results. Thus I do not supply proofs of some prominent counterexamples (such as Kolmogorov’s integrable function whose Fourier series diverges at every point). I am grateful to Noel DeJarnette, Eunmi Kim, Aleksandra Kwiatkowska, Kostya Slutsky, Khang Tran and Ping Xu for TEXing parts of the document. Please email me with corrections, and with suggested improvements of any kind. Richard S. Laugesen Department of Mathematics University of Illinois Urbana, IL 61801 U.S.A. Email: Laugesen@illinois.edu

*

modulo some improvements after the fact

3

Introduction

Harmonic analysis began with Fourier’s eﬀort to analyze (extract information from) and synthesize (reconstruct) the solutions of the heat and wave equations, in terms of harmonics. Speciﬁcally, the computation of Fourier coeﬃcients is analysis, while writing down the Fourier series is synthesis, and the harmonics in one dimension are sin(nt) and cos(nt). Immediately one asks: does the Fourier series converge? to the original function? In what sense does it converge: pointwise? mean-square? Lp ? Do analogous results hold on Rd for the Fourier transform? We will answer these classical qualitative questions (and more!) using modern quantitative estimates, involving tools such as summability methods (convolution), maximal operators, singular integrals and interpolation. These topics, which we address for both Fourier series and transforms, constitute the theoretical core of the course. We further cover the sampling theorem, Poisson summation formula and uncertainty principles. This graduate course is theoretical in nature. Students who are intrigued by the fascinating applications of Fourier series and transforms are advised to browse [Dym and McKean], [K¨rner] and [Stein and Shakarchi], which are o all wonderfully engaging books. If more time (or a second semester) were available, I might cover additional topics such as: Littlewood–Paley theory for Fourier series and integrals, Fourier analysis on locally compact abelian groups [Rudin] (especially Bochner’s theorem on Fourier transforms of nonnegative functions), short-time Fourier transforms [Gr¨chenig], discrete Fourier transforms, the o Schwartz class and tempered distributions and applications in Fourier analysis [Strichartz], Fourier integral operators (including solutions of the wave and Schr¨dinger equations), Radon transforms, and some topics related to o signal processing, such as maximum entropy, spectral estimation and prediction [Benedetto]. I might also cover multiplier theorems, ergodic theorems, and almost periodic functions.

4 .

Contents I Fourier series 7 9 15 25 27 31 35 43 47 53 57 61 67 71 5 1 Fourier coeﬃcients: basic properties 2 Fourier series: summability in norm 3 Fourier series: summability at a point 4 Fourier coeﬃcients in 5 Fourier coeﬃcients in 6 Maximal functions 7 Fourier summability pointwise a.e. 8 Fourier series: convergence at a point 9 Fourier series: norm convergence 10 Hilbert transform on L2 (T) 11 Calder´n–Zygmund decompositions o 12 Hilbert transform on Lp (T) 13 Applications of interpolation 1 (Z) (or. f ∈ L2 (T)) 2 . f ∈ A(T)) (Z) (or.

19 Fourier integrals: norm convergence 20 Hilbert and Riesz transforms on L2 (Rd ) 21 Hilbert and Riesz transforms on Lp (Rd ) III Fourier series and integrals 127 129 135 141 22 Band limited functions 23 Periodization and Poisson summation 24 Uncertainty principles IV V Problems Appendices 147 159 161 163 165 A Minkowski’s integral inequality B Lp norms and the distribution function C Interpolation .e.6 CONTENTS II Fourier integrals 75 79 87 95 97 101 107 113 123 14 Fourier transforms: basic properties 15 Fourier integrals: summability in norm 16 Fourier inversion when f ∈ L1 (Rd ) 17 Fourier transforms in L2 (Rd ) 18 Fourier integrals: summability a.

Part I Fourier series 7 .

.

2π-periodic functions}. continuous. Banach space with norm · L∞ (T) 1/p Trigonometric polynomial P (t) = Translation fτ (t) = f (t − τ ) N n=−N an eint 9 . 2π-periodic functions} 1 f Lp (T) = 2π T |f (t)|p dt where T can be taken over any interval of length 2π Nesting of Lp -spaces: L∞ (T) ⊂ L2 (T) ⊂ L1 (T) C(T) = {complex-valued. p-th power integrable.1 Notation T = R/2πZ is the one dimensional torus Lp (T) = {complex-valued.Chapter 1 Fourier coeﬃcients: basic properties Goal Derive basic properties of Fourier coeﬃcients Reference [Katznelson] Section I.

deﬁne f (n) = n-th Fourier coeﬃcient of f 1 = f (t)e−int dt. See Chapter 5.10 CHAPTER 1. g ∈ L1 (T). By (1. Thus f (n) =amplitude of f in direction of eint . T since e−iπ = −1 (1. Linearity (f + g)(n) = f (n) + g(n) and (cf )(n) = cf (n) Conjugation f (n) = f (−n) int Trigonometric polynomial P (t) = N has P (n) = an for |n| ≤ N n=−N an e and P (n) = 0 for |n| > N takes translation to modulation. Lemma 1. T . τ ∈ T. f (n) = − 1 2π 1 =− 2π f (t)e−in(t+π/n) dt T 1 4π [f (t) − f (t − π/n)] e−int dt.2 (Basic properties). 2π T The formal series S[f ] = f (n)eint is the Fourier series of f .1). [f (t)eijt ] (n) = f (n − j) : L1 (T) → ∞ (Z) is bounded. Proof. Let f. fτ (n) = e−inτ f (n) takes modulation to translation. Theorem 1. eint where f. f (n) = Proof.2) and the deﬁnition (1. 1 1 1 f (n) = f (n) + f (n) = 2 2 4π [f (t) − f (t − π/n)] e−int dt. (1. For f ∈ L2 (T).1. For f ∈ L1 (T) and n ∈ Z. c ∈ C. For n = 0.2) f (t − π/n)e−int dt T by t → t − π/n and periodicity.3 (Diﬀerence formula).1) 1 Aside. g = 2π T f (t)g(t) dt is that L2 inner product. note f (n) = f. FOURIER COEFFICIENTS: BASIC PROPERTIES Deﬁnition 1. j. n ∈ Z. Exercise. with |f (n)| ≤ f L1 (T) Hence if fm → f in L1 (T) then fm (n) → f (n) (uniformly in n) as m → ∞.

1 ≤ p < ∞. the faster its Fourier coeﬃcients decay. since f = f0 .11 Lemma 1. Lemma 1. 1 ≤ p < ∞. The map φ : T → Lp (T) τ → fτ is continuous. with f (n) = O(1) explicitly by Theorem 1. Hence lim supτ →τ0 fτ − fτ0 Lp (T) = 0. Corollary 1. Take g ∈ C(T) and observe f τ − f τ0 Lp (T) ≤ fτ − gτ Lp (T) + gτ − gτ0 = 2 f − g Lp (T) + gτ − gτ0 → 2 f − g Lp (T) Lp (T) Lp (T) + gτ0 − fτ0 Lp (T) as τ → τ0 . Smoothness and decay The Riemann–Lebesgue lemma says f (n) = o(1). f (n) → 0 as |n| → ∞. Here C α (T) denotes the H¨lder continuous functions: f ∈ C α (T) if f ∈ o C(T) and there exists A > 0 such that |f (t) − f (τ )| ≤ A|t − τ |α whenever |t − τ | ≤ 2π. Fix f ∈ Lp (T). Theorem 1. as desired. then f (n) = O(1/|n|α ). by uniform continuity of g. 0 < α ≤ 1. If f ∈ C α (T). Proof.4. We show the smoother f is. the diﬀerence f − g can be made arbitrarily small.2. Proof.4 (Continuity of translation). Let τ0 ∈ T. which tends to zero as |n| → ∞ by the L1 -continuity of translation in Lemma 1.6 (Less than one derivative).5 (Riemann–Lebesgue lemma).3 implies |f (n)| ≤ 1 f − fπ/n 2 L1 (T) . By density of continuous functions in Lp (T). .

7 (One derivative). f (n) = with |f (n)| ≤ 1 1 1 f (n) = o(1) = o( ). Remark 1. Theorem 1.8 (Higher derivatives). Proof. in By Riemann-Lebesgue applied to f . Similar decay results hold for functions of bounded variation.3.12 Proof. where f ∈ L1 (T). CHAPTER 1. Integrating by parts gives f (n) = 1 2π f (t)e−int dt = T 1 2π T e−int f (t) dt. . If f is 2π-periodic and k times diﬀerentiable (f ∈ W k. provided one integrates by parts using the Lebesgue–Stieltjes measure df (t) instead of f (t) dt. Absolute continuity of f says t f (t) = f (0) + 0 f (τ ) dτ. in in n 1 1 |f (n)| ≤ f |n| |n| L1 (T) . |n|α Theorem 1.1 (T)) then f (n) = o(1/|n|k ) and |f (n)| ≤ f (k) L1 (T) /|n|k . Integrate by parts k times. Therefore |f (n)| ≤ 1 π A 4π n α 2π = const. Proof. If f is 2π-periodic and absolutely continuous (f ∈ W 1.9. . FOURIER COEFFICIENTS: BASIC PROPERTIES f (n) = 1 4π [f (t) − f (t − π/n)]e−int dt T by the Diﬀerence Formula in Lemma 1.1 (T)) then f (n) = o(1/n) and |f (n)| ≤ f L1 (T) /|n|.

8] 1. If f ∈ Lp (T). Theorem 1. n ∈ Z. Convolution facts [Katznelson.3) T f (t − τ )e−in(t−τ ) dt g(τ )e−inτ dτ T T = f (n)g(n). and linear with respect to f and g. if f ∈ C(T) and g ∈ L1 (T) then f ∗ g ∈ C(T). That f ∗ g ∈ Lp (T) satisﬁes (1. 1 ≤ p ≤ ∞. Further. if f ∈ C(T) and g ∈ L1 (T) then f ∗ g is continuous because (f ∗ g)(t + δ) → (f ∗ g)(t) as δ → 0 by uniform continuity of f .3.11 (Convolution and Fourier coeﬃcients). T t ∈ T.13 Convolution Deﬁnition 1. then f ∗ g ∈ Lp (T) with f ∗g and (f ∗ g)(n) = f (n)g(n). deﬁne their convolution (f ∗ g)(t) = 1 2π f (t − τ )g(τ ) dτ.1. where τ = t − θ. Given f. g ∈ L1 (T). Finally. Then by Fubini’s theorem. . Thus takes convolution to multiplication. Convolution is commutative: (f ∗ g)(t) = 1 f (t − τ )g(τ ) dτ 2π T 1 = f (θ)g(t − θ) dθ 2π T = (g ∗ f )(t). (f ∗ g)(n) = 1 2π 1 = 2π 1 2π 1 2π f (t − τ )g(τ ) dτ e−int dt T Lp (T) ≤ f Lp (T) g L1 (T) (1. Section I.10. Proof. dτ = −dθ Convolution is also associative.3) is exactly Young’s Theorem A. and g ∈ L1 (T).

(1. Proof.] More generally. Use linearity and (1. 3. (1. Convolution is continuous on Lp (T): if fm → f ∈ Lp (T). to prove fm ∗ g → f ∗ g in Lp (T).14 CHAPTER 1. FOURIER COEFFICIENTS: BASIC PROPERTIES 2. n (P ∗ f )(t) = j=−n n aj 1 2π eij(t−τ ) f (τ ) dτ T = j=−n aj eijt f (j).4) holds for P (t) = ∞ j=−∞ aj eijt provided {aj } ∈ 1 (Z). and g ∈ L1 (T) then fm ∗ g → f ∗ g in Lp (T). . [Sanity check: (P ∗ f )(j) = aj f (j) = P (j)f (j) as expected. 1 ≤ p ≤ ∞.4) Proof. Convolution with a trigonometric polynomial gives a trigonometric polyijt nomial: if f ∈ L1 (T) and P (t) = n then j=−n aj e n (P ∗ f )(t) = j=−n aj f (j)eijt .3).

Indeed. Aside. Norm convergence is stronger than summability. meaning σn (f ) → f in Lp (T) when 1 ≤ p < ∞. 15 n f (j) . then the arithmetic means (s0 + · · · + sn )/(n + 1) also converge to s (Exercise).2 Write n (Sn f )(t) = j=−n f (j)eijt = n-th partial sum of Fourier series of f . when 1 < p < ∞. In this chapter we prove summability of Fourier series. In Chapter 9 we prove norm convergence of Fourier series: Sn (f ) → f in Lp (T). if a sequence {sn } in a Banach space converges to s.Chapter 2 Fourier series: summability in norm Goal Prove summability (averaged convergence) in norm of Fourier series Reference [Katznelson] Section I. where S0 (f ) + · · · + Sn (f ) |j| 1− σn (f ) = = n+1 n+1 j=−n = arithmetic mean of partial sums.

1) n+1 j=−n sin n+1 t 2 1 sin 2 t 2 ∼ (2.1) (2.4). so that (S2) fails. (S4) Call the kernel positive if kn ≥ 0 for each n.) log n as n → ∞. Deﬁne the Fej´r kernel e Fn (t) = D0 (t) + · · · + Dn (t) n+1 n |j| = 1− eijt by (2. ∴ {Dn } is not a summability kernel.4) (2. Example 2. π).3. You can show (optional exercise) that Dn (const.1.5) 1 = n+1 by (2.4) and (2.2) (2. (S1) (S2) (S3) lim |kn (t)| dt = 0 {δ<|t|<π} Some kernels satisfy a stronger concentration property: n→∞ δ<|t|<π lim sup |kn (t)| = 0 (L∞ concentration) for each δ ∈ (0.2) and geometric series (2. Deﬁne the Dirichlet kernel n Dn (t) = j=−n eijt by geometric series (2. FOURIER SERIES: SUMMABILITY IN NORM Deﬁnition 2.16 CHAPTER 2.2.3) L1 (T) = ei(n+1)t − e−int eit − 1 1 sin (n + 2 )t = sin 1 t 2 (S1) holds by (2. Example 2. A summability kernel is a sequence {kn } in L1 (T) satisfying: 1 kn (t) dt = 1 2π T 1 sup |kn (t)| dt < ∞ n 2π T n→∞ (Normalization) (L1 bound) (L1 concentration) for each δ ∈ (0. (2. π).1).6) .

1: Dirichlet kernel with n = 10 Π 10 Π Figure 2.17 20 10 Π Figure 2.2: Fej´r kernel with n = 10 e Π .

.7).7) (2.11) 2π 2 =√ e−(t+2πn) /4s 4πs n=−∞ by Example 23.8) and simplifying.9) = j=−∞ 1 − r2 = 1 − 2r cos t + r2 by summing two geometric series (j < 0 and j ≥ 0) in (2.5.7 later in the course. with limiting process r 1.4. →0 ∴ {Pr } is a positive summability kernel. and Pr ≥ 0 by (2. Deﬁne the Gauss kernel ∞ Gs (t) = j=−∞ e−j s eijt ∞ 2 (2. 1).9) so that (S2) holds also. Example 2. sup |Pr (t)| ≤ δ<|t|<π 1 − r2 1 − 2r cos δ + r2 by (2. sup |Fn (t)| ≤ δ<|t|<π 1 1 2 1 n + 1 sin 2 δ by (2. we see (S1) holds by (2.6) as n → ∞. Example 2. Deﬁne the Poisson kernel ∞ Pr (t) = 1 + 2 j=1 ∞ rj cos(jt) r|j| eijt (2. FOURIER SERIES: SUMMABILITY IN NORM (S1) holds by (2.9) as r 1. For (S4). For (S4). The Poisson kernel is indexed by r ∈ (0. →0 ∴ {Fn } is a positive summability kernel.5). and Fn ≥ 0 so that (S2) holds also.18 CHAPTER 2. After suitably modifying the deﬁnition of summability kernel.8) (2.10) (2.

19 20 10 Π Figure 2.3: Poisson kernel with r = 0.4: Gauss kernel with s = 0.01 Π .9 Π 20 10 Π Figure 2.

4). we want Fn ∗ f → f .4) n j=−n (1 − |j| )eijt n+1 n implies |j| f (j)eijt = σn (f ) n+1 (Fn ∗ f )(t) = j=−n 1− by Convolution Fact (1. ∞). use that σn (f ) = [S0 (f ) + · · · + Sn (f )]/(n + 1) and Fn = [D0 + · · · + Dn ]/(n + 1).4). FOURIER SERIES: SUMMABILITY IN NORM The Gauss kernel is indexed by s ∈ (0.11) as s 0. Dn (t) = (2. σn (f ) = Fn ∗ f Proof.1) n j=−n by (2. The analogue of (S1) holds by (2.10). Thus for summability of Fourier series. with limiting process s 0. Alternatively. 1eijt implies n (Dn ∗ f )(t) = j=−n 1f (j)eijt = Sn (f ) by Convolution Fact (1. Fn (t) = (2. For (S4). Abel mean of S[f ] = Pr ∗ f . Connection to Fourier series Sn (f ) = Dn ∗ f Proof.11) so that (S2) holds also.20 CHAPTER 2. and Gs ≥ 0 by (2. 2π 2 2 e−δ /4s + e−(πn) /4s sup |Gs (t)| ≤ √ 4πs δ<|t|<π n=0 →0 ∴ {Gs } is a positive summability kernel.

12) by Convolution Fact (1.13) Then (kn ∗ f )(t) − f (t) = Lp (T) 1 kn (τ )[fτ (t) − f (t)] dτ Lp (T) by (S1) 2π T 1 ≤ |kn (τ )| fτ − f Lp (T) dτ 2π T by Minkowski’s Integral Inequality. If {kn } is a summability kernel and f ∈ Lp (T).8) ∞ j=−∞ r|j| eijt implies ∞ (Pr ∗ f )(t) = j=−∞ r|j| f (j)eijt (2. we can choose 0 < δ < π such that max fτ − f |τ |≤δ Lp (T) · sup kn n L1 (T) < ε. = 1 2π |τ |≤δ δ + −δ {δ<|τ |<π} Lp (T) |kn (τ )| fτ − f 1 2π 1 2π δ Lp (T) dτ ≤ max fτ − f |τ |≤π |kn (τ )| dτ −δ + max fτ − f <ε+ε Lp (T) |kn (τ )| dτ {δ<|τ |<π} . if f ∈ C(T) then kn ∗ f → f in C(T). Similarly. Pr (t) = (2.4) (with the series converging absolutely and uniformly). as n → ∞. By (S2) and continuity of translation on Lp (T) (Lemma 1. Theorem A.6 (Summability in Lp (T) and C(T)). 1 ≤ p < ∞.4). then kn ∗ f → f in Lp (T).21 Proof. Proof. Summability in norm Theorem 2.1. (2. Let ε > 0. and this last expression is the Abel mean of S[f ].

6. 1. Density of trigonometric polynomials in C(T) proves the Weierstrass Trigonometric Approximation Theorem.4).14) In other words.6. σn (f ) is a trigonometric polynomial arbitrarily close to f . 1 2π 1 − r2 f (τ ) dτ 1 − 2r cos(t − τ ) + r2 T . (2. Choose kn = Fn = Fej´r kernel. Fix f ∈ L1 (T). we connect our summability kernels to some important partial diﬀerential equations. If f ∈ C(T) then repeat the argument with p = ∞. Aside. 1 ≤ p < ∞. the map : L1 (T) → ∞ (Z) is injective. 1 ≤ p < ∞: σn (f ) → f in norm. Lp (T). then f = g in L1 (T). Proof. Letting n → ∞ gives f = g. • Uniqueness theorem: if f. The Poisson kernel solves Laplace’s equation in a disk: v(reit ) = (Pr ∗ f )(t) = solves ∆v = vrr + r−1 vr + r−2 vtt = 0 on the unit disk {r < 1}. t) = f (t) in the sense of Theorem 2.13) and (S3). since f = g. • Trigonometric polynomials are dense in C(T).22 CHAPTER 2. g ∈ L1 (T) with f (n) = g(n) for all n. Then σn (f ) = Fn ∗ f → f in norm by e Theorem 2. Proof. for all large n. with boundary value v(1. Lp (T). Connection to PDEs To ﬁnish the section. using uniform continuity of f get that fτ → f in L∞ (T). Fn ∗f = Fn ∗g by Convolution Fact (1. Consequences • Summability of Fourier series in C(T). Proof. FOURIER SERIES: SUMMABILITY IN NORM by (2.

t) = (Gs ∗ f )(t) solves ws = wtt for (s.4). t) ∈ (0. Proof.23 That is. Gs (t) = (2. v is the harmonic extension of f from the boundary circle to the disk. 2 . ∞) × T.12) for Pr ∗ f and note that ∂2 1∂ 1 ∂2 + + 2 2 (r|j| eijt ) = 0.6.10) ∞ −j 2 s ijt e j=−∞ e implies ∞ (Gs ∗ f )(t) = j=−∞ e−j s f (j)eijt 2 by Convolution Fact (1. ∂r2 r ∂r r ∂t 2. Diﬀerentiate through formula (2. The Gauss kernel solves the diﬀusion (heat) equation: w(s. t) = f (t) in the sense of Theorem 2. Proof. with initial value w(0. Now diﬀerentiate through the sum and use that ∂ ∂2 − ∂s ∂t2 (e−j s eijt ) = 0.

24 CHAPTER 2. FOURIER SERIES: SUMMABILITY IN NORM .

then (kn ∗ f )(t0 ) → L 25 as n → ∞. But what if f is merely continuous at a point? Theorem 3. σn (f )(t) → f (t).f ∈ L1 (T) and t0 ∈ T. (b) If in addition the summability kernel is even (kn (−t) = kn (t)) and L = lim exists (or equals ±∞).3 By Chapter 2. That is. f (t0 + h) + f (t0 − h) h→0 2 . or else f ∈ L∞ (T).Chapter 3 Fourier series: summability at a point Goal Prove a suﬃcient condition for summability at a point Reference [Katznelson] Section I.1 (Summability at a point). σn (f ) → f uniformly. Assume {kn } is a summability kernel. (a) If f is continuous at t0 then (kn ∗ f )(t0 ) → f (t0 ) as n → ∞. Suppose either {kn } satisﬁes the L∞ concentration hypothesis (S4). if f is continuous then σn (f ) → f in C(T). In particular. for each t ∈ T.

1).1) using here (S2) and continuity of f at t0 . for each integrable f .1(a).2. for all large n. Remark 3. in Theorem 3. Theorem 3.26 CHAPTER 3. Proof. diﬀer from the proof of summability in norm. |(kn ∗ f )(t0 ) − f (t0 )| = {|τ |<δ} kn (τ )[f (t0 − τ ) − f (t0 )] dτ − {δ<|τ |<π} kn (τ ) dτ · f (t0 ) + {δ<|τ |<π} L1 (T) L∞ (T) kn (τ )f (t0 − τ ) dτ using (S1) < <ε ε + o(1) + o(1) · f ε + o(1) + o(1) · f by (3. (S3) and (S1).6? 2. in Theorem 2. and so Theorem 3. . FOURIER SERIES: SUMMABILITY AT A POINT Note if f has limits from the left and right at t0 .1 treats summability at a single point t0 at which f is continuous. |τ |≤δ (3. but uses symmetry of the kernel. then the quantity L equals the average of those limits. Chapter 7 will prove kn ∗ f → f at almost every point.1). (b) The proof is similar to (a). 1.1 applies e in particular to summability at a point for σn (f ) = Fn ∗ f and for the Abel mean Pr ∗ f . (a) Let ε > 0 and choose 0 < δ < π such that sup kn n L1 (T) · max |f (t0 − τ ) − f (t0 )| < ε. (S3) and (S4). or else by (3. The Fej´r and Poisson kernels satisfy (SR4). How does the proof of summability at a point. Then as n → ∞.

let {cn } ∈ 1 (Z) and deﬁne g(t) = n∈Z cn eint . To prove surjectivity. 27 .14). and so g = {cm } as desired. (Hence g is continuous.) We have g(m) = cm for every m. The series for g converges uniformly since sup t∈T |n|>N cn eint ≤ |n|>N |cn | → 0 as N → ∞.Chapter 4 Fourier coeﬃcients in 1(Z) (or. Proof.6 Deﬁne A(T) = {f ∈ L1 (T) : n∈Z |f (n)| < ∞} 1 = functions with Fourier coeﬃcients in (Z). The map : A(T) → 1 (Z) is a linear bijection. Injectivity follows from the uniqueness result (2. f ∈ A(T)) Goal Establish the algebra structure of A(T) Reference [Katznelson] Section I.

2) b . A(T) is an algebra. F ∈ A(T)) Our proof has shown each f ∈ A(T) is represented by its Fourier series: f (t) = n∈Z f (n)eint a. Clearly a∗b because |(a ∗ b)(n)| ≤ n m 1 (Z) ≤ a 1 (Z) b 1 (Z) (4. f g is continuous. So (f g) 1 (Z) ≤ f 1 (Z) g 1 (Z) by (4. b ∈ (a ∗ b)(n) = m∈Z 1 (Z) is one). . meaning that if f.1. 1 A(T) is a Banach space under this norm (because Deﬁne the convolution of sequences a. and hence integrable. FOURIER COEFFICIENTS IN 1 (Z) (OR.1) = m f (m)g(n − m) = (f ∗ g)(n). This Fourier series converges absolutely and uniformly. (Z) by a(m)b(n − m). Deﬁne a norm on A(T) by f A(T) = f 1 (Z) = n |f (n)|. Deﬁnition 4.2 ( takes multiplication to convolution). Indeed fg = f ∗ g and f g A(T) ≤ f A(T) g A(T) .2).e. g ∈ A(T) then f g ∈ A(T). (4. Proof. with (f g)(n) = = m 1 2π f (t)g(t)e−int dt T f (m) 1 2π g(t)e−i(n−m)t dt T by (4. |a(m)| n |b(n − m)| = a 1 (Z) 1 (Z) Theorem 4.1) so that f is continuous (after redeﬁnition on a set of measure zero).28 CHAPTER 4.

We omit the proof.3 (Wiener’s Inversion Theorem).29 Suﬃcient conditions for membership in A(T) are discussed in [Katznelson. Section I. H¨lder continuity: C α (T) ⊂ A(T) when α > 1 . o 2 Theorem 4. but it is not clear that (1/f ) belongs to 1 (Z).6]. Clearly 1/f is continuous. If f ∈ A(T) and f (t) = 0 for every t ∈ T then 1/f ∈ A(T). for example. .

FOURIER COEFFICIENTS IN 1 (Z) (OR.30 CHAPTER 4. F ∈ A(T)) .

and the series S({cn }) = cn un converges unconditionally. deﬁne the synthesis operator S: 2 u. then so is synthesis.1.Chapter 5 Fourier coeﬃcients in 2(Z) (or. un |2 ≤ (const. u . 31 . Theorem 5. (Z) → H c n un n {cn }n∈Z → and analysis operator T : H → 2 (Z) u → { u.) u 2 for all u ∈ H). un }n∈Z . using analysis and synthesis operators Notation and deﬁnitions Let H be a Hilbert space with inner product u. f ∈ L2(T)) Goal Study the Fourier ONB for L2 (T). v and norm u = Given a sequence {un }n∈Z in H. If analysis is bounded ( n | u.

we have N T ∗ ({cn }N n=−N ). u ∈ H. therefore T ∗ ({cn }) = ∞ n=−∞ cn un .2. Hence S is bounded. The Fourier coeﬃcient (or analysis) operator : L2 (T) → 2 (Z) is an isometry. meaning analysis and synthesis are adjoint operations. g L2 (T) L2 (T) = f 2 (Z) 2 (Z) (Plancherel) (Parseval) = f. and for 2 = n=−N N cn u. Convergence of the synthesis series is unconditional. so that ∗ T = S. First we prove Plancherel’s identity: since Pr ∗ f → f in L2 (T) by . g ∈ L2 (T). F ∈ L2 (T)) (Z) → H is bounded. N ≥ 1. Theorem 5. the adjoint T ∗ : each sequence {cn }. The limit as N → ∞ exists on the left side.3 (Fourier coeﬃcients on L2 (T)). Remark 5. u . T u N 2 2 (Z) (OR. The last proof shows S = T ∗ . and hence on the right side. which tends to 0 as A expands to ﬁll Z. because if A ⊂ Z then S({cn }n∈Z ) − S({cn }n∈A ) = S({cn }n∈Z\A ) ≤ S {cn } 2 (Z\A) . u = {cn }n=−N . n=−N by deﬁnition of T u = N Hence T ∗ ({cn }n=−N ) = N n=−N cn un . un c n un . FOURIER COEFFICIENTS IN Proof. with f f. g for all f. regardless of the order in which A expands.32 CHAPTER 5. Proof. Since T is bounded.

(b) If c = {cn } ∈ 2 (Z) then ( n∈Z cn eint ) (j) = cj . (f )ˇ = f . (ˇ)ˆ = c. That is. Argue similarly for part (b).6. Since the Fourier analysis operator is bounded. Part (a) says Fourier series converge in L2 (T). Parts (a) and (b) together show that Fourier analysis and synthesis are inverse operations. . and analysis followed by synthesis equals the identity ( n f (n)eint = f ) on the class of trigonometric polynomials. Fourier analysis and synthesis are bounded operators.33 Theorem 2. g (and using dominated instead of monotone convergence). using the dense class of ﬁnite sequences in 2 (Z). so is its adjoint. That is. and so by continuity. f changed to f. Proof.12) for Pr ∗ f r |f (n)|2 |f (n)|2 by monotone convergence. (a) If f ∈ L2 (T) then n f (n)eint = f with unconditional convergence in ˆ L2 (T). we have 1 2π |f (t)|2 dt = lim T 1 r→1 2π 1 = lim r→1 2π = lim = n∈Z r→1 n∈Z f (t)(Pr ∗ f )(t) dt T f (t) T |n| n∈Z r|n| f (n)eint dt by (2. c int 2 (c) {e }n∈Z is an orthonormal basis of L (T). That class is dense in L2 (T). or by repeating the argument for Plancherel with f. analysis followed by synthesis equals the identity on L2 (T).4 (Fourier ONB). the Fourier synthesis operator ˇ: 2 (Z) → L2 (T) cn eint n {cn }n∈Z → Theorem 5. Parseval follows from Plancherel by polarization.

observe eint .3. eint Remark 5. p p .34 CHAPTER 5. noting f (n) = f. 0 if n = m. FOURIER COEFFICIENTS IN For orthonormality in part (c). Fourier analysis satisﬁes : L1 (T) → : L2 (T) → ∞ 2 (Z) (Z) by Theorem 1. eimt = 1 2π eint e−imt dt = T 2 (Z) (OR. (isometrically) by Theorem 5. F ∈ L2 (T)) 1 if n = m. The basis property follows from part (a). In Chapter 13 we will interpolate to show : Lp (T) → p (Z). 1 1 + = 1. Further. whenever 1 ≤ p ≤ 2. : L2 (T) → 2 (Z) is a linear bijection by Theorem 5.2.4. L2 (T) .5.

) Call T strong (p. f ∈ Lp (X). (We do not assume T is linear.1 Deﬁnition 6. q ≤ ∞. in next Chapter References [Duoandikoetxea] Section 2. Prove weak and strong bounds on the Hardy–Littlewood maximal function Prepare for sumability pointwise a.e. and 1 ≤ p. we call T weak (p. q) if C > 0 exists such that ν {y ∈ Y : |(T f )(y)| > λ} 1/q ≤ C f 35 Lp (X) λ ∀ λ > 0. . When q < ∞.1 [Stein] Section 1. Suppose T : Lp (X) → {measurable functions on Y }.Chapter 6 Maximal functions Goals Connect abstract maximal functions to convergence a. Let (X. ν) be measure spaces. µ) and (Y.2 [Grafakos] Section 2. meaning a constant C > 0 exists such that Tf Lq (Y ) ≤C f Lp (X) . q) if T is bounded from Lp (X) to Lq (Y ).e.1 (Weak and strong operators). f ∈ Lp (X).

Proof. Then |g(y)|r dν(y) Z ∞ 0 1 = ≤ 0 rλr−1 ν {y ∈ Z : |g(y)| > λ} dλ ∞ 1 by AppendixB Lp (X) q rλr−1 ν(Z) dλ + rλr−1 C f λ dλ by weak (p. Write g = T f . Then λq ν E(λ) = E(λ) λq dν(y) |(T f )(y)|q dν(y) E(λ) q Lq (Y ) ≤ ≤ Tf and so since λ < |T f | on E(λ) ν E(λ) 1/q ≤ ≤ Tf C f Lq (Y ) λ Lp (X) λ if T is strong (p.3. T “almost” maps Lp into Lq . Lemma 6. ∞): Tf L∞ (Y ) ≤C f Lp (X) . Suppose q < ∞. Proof. q).36 CHAPTER 6. f ∈ Lp (X). ∞) if it is strong (p. We will show T f ∈ Lr (Z). Let f ∈ Lp (X) and suppose Z ⊂ Y with ν(Z) < ∞. MAXIMAL FUNCTIONS When q = ∞. Lemma 6. q). q) ⇒ weak (p. Write E(λ) = {y ∈ Y : |(T f )(y)| > λ} for the level set of T f above height λ.2. . we call T weak (p. locally. Strong (p. loc Thus intuitively. If T is weak (p. When q = ∞ the result is immediate by deﬁnition. q) then T f ∈ Lr (Y ) for all 0 < r < q. q) <∞ since ν(Z) < ∞ and ∞ 1 λ−1−q+r dλ < ∞ (using that −q + r < 0).

Note T ∗ is not linear. ≤ µ {x ∈ X : T ∗ (f − fk )(x) + |(f − fk )(x)| > 2λ} ∗ by triangle inequality ≤ µ {x ∈ X : T (f − fk )(x) > λ} + µ {x ∈ X : |(f − fk )(x)| > λ} C f − fk Lp (X) q f − fk Lp (X) p ≤ + by weak (p. Deﬁne T ∗ : Lp (X) → {measurable functions on X} by (T ∗ f )(x) = sup |(Tn f )|(x)|.e} n is closed in Lp (X). ∞]. .e.4 (Maximal functions and convergence a.. Suppose q < ∞. .37 Theorem 6. For any λ > 0. T ∗ is called the maximal operator for the family {Tn }. in general. Taking a countable sequence of λ 0. The case q = ∞ is left to the reader. q)) implies a qualitative conclusion (closure of the collection C where Tn f → f a. Proof. 3. Therefore lim supn |(Tn f )(x) − f (x)| ≤ 2λ a.. In this theorem a quantitative hypothesis (weak (p. Let fk ∈ C with fk → f in Lp (X). Assume Tn : Lp (X) → {measurable functions on X} for n = 1.5.). Therefore limn (Tn f )(x) = f (x) a. then the collection C = {f ∈ Lp (X) : lim(Tn f )(x) = f (x) a.e. we conclude lim supn |(Tn f )(x) − f (x)| = 0 a. .e. q) and each Tn is linear. µ {x ∈ X : lim sup |(Tn f (x) − f (x)| > 2λ} n = µ {x ∈ X : lim sup |Tn (f − fk )(x) − (f − fk )(x)| > 2λ} n by linearity and the pointwise convergence Tn fk → fk a. 2.e.e. Clearly it takes values in [0. so that f ∈ C. q) on T ∗ λ λ →0 as k → ∞. n x ∈ X. We show f ∈ C. Remark 6.e. . If T ∗ is weak (p.).

MAXIMAL FUNCTIONS To apply maximal functions on Rd and T. so that radius (Bi ) ≤ radius(Bij ). . If Bi is not one of the Bij chosen. . Let {Bi }k be a ﬁnite collection of open balls in i=1 Rd . Bij . . Then there exists a pairwise disjoint subcollection {Bij }l of balls such j=1 that k l l Bi ≤ 3 i=1 d j=1 Bij = 3 d j=1 |Bij |. Thus k l Bi ≤ i=1 l (3Bij ) j=1 ≤ j=1 |3Bij | l d j=1 l =3 =3 by disjointness of the Bij . . by construction.38 CHAPTER 6. we will need: Lemma 6. Choose i1 = 1 and employ the following greedy algorithm. k}. Re-label the balls in decreasing order of size: |B1 | ≥ |B2 | ≥ · · · ≥ |Bk |. . . |Bij | Bij j=1 d . The Bij are pairwise disjoint. Proof. Continue until no such ball Bi exists. After choosing ij . then Bi must intersect one of the Bij and be smaller than it. . Let i ∈ {1.6 (Covering). choose ij+1 to be the smallest index i > ij such that Bi is disjoint from Bi1 . Thus the subcollection covers at least 1/3d of the total volume of the balls. . Hence Bi ⊂ 3Bij (where we mean the ball with the same center and three times the radius).

M is weak (1. For weak (1. Bk . .7. F is covered by ﬁnitely many such balls. . Thus E(λ) is open (and measurable).r) = “largest local average” of |f | around x. p) for 1 < p ≤ ∞. If x ∈ E(λ) then 1 |B(x.8 (H-L maximal operator).39 Deﬁnition 6.) ≥ 0 Theorem 6. . .1) where E(λ) = {x ∈ Rd : M f (x) > λ}. The same inequality holds for all x close to x. say B1 . Properties Mf ≥ 0 |f | ≤ |g| ⇒ M f ≤ M g M (f + g) ≤ M f + M g (sub-linearity) M c = c if c = (const. Let F ⊂ E(λ) be compact. r)| |f (y)| dy > λ B(x.r) for some r > 0. . Each x ∈ F is the center of some ball B such that 1 |B| < |f (y)| dy. r)| |f (y)| dy B(x. and M f is lower semicontinuous (and measurable). so that x ∈ E(λ). 1) and strong (p.2) λ B By compactness. The Hardy–Littlewood (H-L) maximal function of a locally integrable function f on Rd is (M f )(x) = sup r>0 1 |B(x. (6. 1) we show |E(λ)| ≤ 3d f L1 (Rd ) λ (6. Proof.

. p) when 1 < p < ∞. Bil . . Hence M f L∞ (Rd ) ≤ f L∞ (Rd ) . Then k |F | ≤ i=1 Bi l d j=1 ≤3 ≤ ≤ |Bij | l by Covering Lemma 6. For strong (p. 1) result (6. ∞). . so that M f ≤ M g + λ/2. λ Taking the supremum over all compact F ⊂ E(λ) gives (6. For strong (∞.3) .6 |f (y)| dy by (6. . let λ > 0 and deﬁne f (x) if |f (x)| > λ/2 = “large” part of f .2) by disjointness 3d λ d j=1 Bij 3 |f (y)| dy λ Rd 3d = f L1 (Rd ) .6 yields a subcollection Bi1 . λ {x:|f (x)|>λ/2} by the above weak (1. 0 otherwise g(x) = h(x) = Then f = g + h and |h| ≤ λ/2. note M f (x) ≤ f L∞ (Rd ) for all x ∈ Rd . by deﬁnition of M f . 0 otherwise f (x) if |f (x)| ≤ λ/2 = “small” part of f .1).40 CHAPTER 6. Hence |E(λ)| = {x : M f (x) > λ} ≤ {x : M g(x) > λ/2} 3d g L1 (Rd ) ≤ λ/2 2 · 3d = |f (x)| dx. MAXIMAL FUNCTIONS The Covering Lemma 6.

We have proved the strong (p. / The maximal function is locally integrable provided f ∈ L log L(Rd ). p) bound.3) ≤ 2 · 3d p = 2p 3d λp−2 {x:|f (x)|>λ/2} p p−1 |f (x)|p dx Rd by Lemma B. p) bound blows up as p 1.1] in 1 dimension has M f (x) ∼ c/|x| when |x| is large.41 Therefore |M f (x)|p dx Rd ∞ 0 = pλp−1 |E(λ)| dλ ∞ 0 by Appendix B |f (x)| dxdλ by (6. α = 2. the indicator function f = 1[−1. For example. Notice the constant in the strong (p. so that M f ∈ L1 (R). the Hardy–Littlewood maximal operator is not strong (1. 1). . see Problem 9.1 with r = 1. As this observation suggests.

MAXIMAL FUNCTIONS .42 CHAPTER 6.

t+h 1 Notice (Lh ∗ f )(t) = 2h t−h f (τ ) dτ is a local average of f around t.e. Dirichlet maximal function (D∗ f )(t) = sup |(Dn ∗ f )(t)| = sup |Sn (f )(t)| n n Fej´r maximal function (F f )(t) = sup |(Fn ∗ f )(t)| = sup |σn (f )(t)| e n n ∗ Poisson maximal function (P f )(t) = sup |(Pr ∗ f )(t)| 0<r<1 ∗ ∗ Gauss maximal function (G f )(t) = sup |(Gs ∗ f )(t)| 0<s<∞ 0<h<π Lebesgue maximal function (L f )(t) = sup |(Lh ∗ f )(t)| 1 where the Lebesgue kernel is Lh (t) = 2π 2h 1[−h. ∗ Lemma 7. then |(k ∗ f )(t)| ≤ k ∗ L1 (T) (L f )(t) for all t ∈ T. using Fej´r and Poisson maximal functions e Deﬁnition 7. Goal Prove summability a. 43 . π].1.Chapter 7 Fourier series: summability pointwise a. If k ∈ L1 (T) is nonnegative and symmetric (k(−t) = k(t)). f ∈ L1 (T).e. extended 2π-periodically.2 (Majorization). and decreasing on [0.h] (t).

. We ﬁrst establish a “layer cake” decomposition of k. so that P ∗ f ≤ L∗ f . e F ∗ f ≤ 2L∗ |f | P ∗ f ≤ L∗ f Proof. π] (exercise).44 CHAPTER 7. Pr (t) is nonnegative. For all f ∈ L1 (T). π 1 2h(Lh ∗ f )(t) − k (h) dh 2π 0 T π 1 2h − k (h) dh (L∗ f )(t) |(k ∗ f )(t)| ≤ k(π)|(Lπ ∗ f )(t)| + 2π 0 using k(π) ≥ 0 and k ≤ 0 π 2 ≤ k(h) dh (L∗ f )(t) by parts 2π 0 = k L1 (T) (L∗ f )(t) (k ∗ f )(t) = k(π) 1 2π f (τ ) dτ + by symmetry of k. FOURIER SUMMABILITY POINTWISE A. symmetric.2. if h ≥ |t|. if h < |t|. Proof. Theorem 7. representing it as a linear combination of kernels Lh : π k(t) = k(|t|) = k(π) − |t| k (h) dh 1 2π π = k(π) − since 1 2hLh (t) = 2π Hence 2hLh (t)k (h) dh.3 (Lebesgue dominates Fej´r and Poisson). 0. Thus convolution with a symmetric decreasing kernel is majorized by the Hardy–Littlewood maximal function.E. 0 1. and decreasing on [0. for simplicity. Hence |Pr ∗ f | ≤ L∗ f by Majorization Lemma 7. Assume k is absolutely continuous. with Pr L1 (T) = 1.

e.e.45 The Dirichlet kernel is not decreasing on [0. and decreasing on [0. P ∗ and L∗ are weak (1. so that G∗ f ≤ L∗ f . 1) proof for the Hardy–Littlewood maximal function. but it is bounded by a symmetric decreasing kernel. Note the kernel k is nonnegative. but we omit the proof. Theorem 7. if π/(n + 1) ≤ |t| ≤ π.e. as follows: 1 Fn (t) = n+1 ≤ k(t) = since sin(n + 1)θ ≤ (n + 1) sin θ. 0 ≤ t ≤ π. as h 0 (Lebesgue diﬀerentiation theorem) .3. symmetric.5 (Summability a.e. π]. so that F ∗ f ≤ 2L∗ |f |. as r 1 (Abel summability) Lh ∗ f → f a. π]. If f ∈ L1 (T) then σn (f ) = Fn ∗ f → f a. These weak (1. Proof.4.2. 1) estimates for L∗ f and L∗ |f | imply weak (1. since if (F ∗ f )(t) > λ then (L∗ |f |)(t) > λ/2 by Theorem 7. 2π n+1 Hence |Fn ∗ f | ≤ k ∗ |f | ≤ 2L∗ |f | by Majorization Lemma 7. 2 def sin n+1 t 2 sin 1 t 2 1 n+1 2 (n + 1)2 π 2 /t2 if |t| ≤ π/(n + 1). Argue similarly for P ∗ f . F ∗ . Corollary 7. The Gauss kernel can be shown to be symmetric decreasing. 2 π 0≤θ≤ π . 1) for F ∗ f . 1) on T. as n → ∞ (Fej´r summability) e Pr ∗ f → f a. {t ∈ T : (L∗ f )(t) > 2πλ} ≤ {t ∈ T : (L∗ |f |)(t) > 2πλ} ≤ 3 λ |f (t)| dt T by repeating the weak (1. with k L1 (T) = 1 2π 4π − < 2. sin 1 t t ≥ .).

Argue similarly for Pr ∗ f and Lh ∗ f .e. Thus C is dense in L1 (T).4 and the abstract convergence result in Theorem 6.e.} n is closed in L1 (T). By the weak (1. means 1 2h t+h f (τ ) dτ → f (t) a.. . Proof. FOURIER SUMMABILITY POINTWISE A. Because C is also closed. 1) estimate in Corollary 7.e. thus proving Fej´r summability a.E. the set C = {f ∈ L1 (T) : lim(Fn ∗ f )(t) = f (t) a. since Fn ∗ f → f uniformly when f is continuous.4. The result that Lh ∗ f → f a. Obviously C contains the continuous functions on T. t−h which is the Lebesgue diﬀerentiation theorem on T.e. for each e 1 f ∈ L (T).46 CHAPTER 7. it must equal L1 (T).

47 . Recall Sn (f ) = Dn ∗ f and D∗ f is the maximal function for the Dirichlet kernel. II. [Katznelson.Chapter 8 Fourier series: convergence at a point Goals State divergence pointwise can occur for L1 (T) Show divergence pointwise can occur for C(T) Prove convergence pointwise for C α (T) and BV (T) References [Katznelson] Section II. That is.1 (Kolmogorov). Section II. so that D∗ f ≡ ∞. There exists f ∈ L1 (T) whose Fourier series diverges unboundedly at every point. Proof.2.1 Fourier series can behave badly for integrable functions.3]. sup |Sn (f )(t)| = ∞ n for all t ∈ T. Theorem 8.3 [Duoandikoetxea] Section 1.

Each Tn is bounded since |Tn (f )| = |Sn (f )(0)| = |(Dn ∗ f )(0)| 1 Dn (τ )f (0 − τ ) dτ = 2π T ≤ Dn L1 (T) f L∞ (T) . Then Tn is linear. sup |Sn (f )(0)| = ∞.48 CHAPTER 8. We show Tn = Dn L1 (T) . π − using deﬁnition (2. Deﬁne Tn : C(T) → C f → Sn (f )(0) = (nth partial sum of f at t = 0). Then |Tn (g)| = 1 2π 1 ≥ 2π 1 = 2π ≥ Dn Dn (τ )g(τ ) dτ T |Dn (τ )| dτ − T\{intervals} 1 2π |Dn (τ )| dτ {intervals} |Dn (τ )| dτ − T L1 (T) 2 2π |Dn (τ )| dτ {intervals} 1 ε (2n + 1) π 2n + 1 ε = Dn L1 (T) − π ε = Dn L1 (T) − g L∞ (T) . FOURIER SERIES: CONVERGENCE AT A POINT Even continuous functions can behave badly. Let ε > 0 and choose g ∈ C(T) with g L∞ (T) = 1 and g even and if Dn (t) > 0. There exists a continuous function whose Fourier series diverges unboundedly at t = 0. Thus Tn ≤ Dn L1 (T) . n g(t) = except for small intervals around the zeros of Dn .1) of Dn . Theorem 8. with total length of those intervals < ε/(2n + 1). That is. n Proof.2. 1 −1 if D (t) < 0.

Let f ∈ L1 (T). by the Riemann–Lebesgue Corollary 1.2] gives an explicit construction of f .1) [f (t − τ ) − f (t)] cos(nτ ) dτ −π by expanding sin (n + 1 )τ with a trigonometric identity. If f ∈ C α (T). and so Tn = Dn L1 (T) . [Katznelson. Corollary 8. 2 Notice the factor {· · · } is integrable with respect to τ .1]) we conclude from the Uniform Bounded Principle (Banach–Steinhaus) that there exists f ∈ C(T) with supn |Tn (f )| = ∞. Proof. Now we prove convergence results. Ex. Another proof. Theorem 8. And τ → [f (t − τ ) − f (t)] is integrable too. for every t ∈ T. then the Fourier series of f converges to f (t).1.1) tend to 0 as n → ∞.4 (Convergence for H¨lder continuous f ). Dn ∼ c log n by [Katznelson. Sec II.49 Thus Tn ≥ Dn L1 (T) − ε/π for all ε > 0. .3 (Dini’s Convergence Test). as desired.5 (after expressing sin(nτ ) and cos(nτ ) in terms of e±inτ ). If π −π f (t − τ ) − f (t) dτ < ∞ τ then the Fourier series of f converges at t to f (t). II. Recalling that Dn L1 (T) → ∞ as n → ∞ (in fact. 1 Sn (f )(t) − f (t) = 2π sin (n + 1 )τ 2 [f (t − τ ) − f (t)] dτ 1 sin 2 τ −π 1 π using that Dn (τ ) dτ = 1 2π −π π π −π = 1 2π 1 τ f (t − τ ) − f (t) cos τ 1 τ 2 sin 2 τ + 1 2π π sin(nτ ) dτ (8. 0 < o α ≤ 1. t ∈ T. Hence both integrals in (8. proving divergence not only at t = 0 but on a dense set of t-values. by the Dini hypothesis.

say f = g − h.)|τ |α dτ < ∞. t + π).50 CHAPTER 8. In particular.3) = 1 δ 1 π G(τ )Hn (τ ) dτ + G(τ )Dn (τ ) dτ 2π 0 2π δ 1 1 = G(δ)Hn (δ) − Hn (τ ) dG(τ ) + o(1) 2π 2π (0. If f vanishes on a neighborhood of t.6 (Convergence for bounded variation f). Theorem 8. Write τ π Hn (τ ) = 0 Dn (σ) dσ so that Hn = Dn . then the Fourier series of the other function converges at t to the same value.3. π). and hence converges to f (t) at every point of continuity. FOURIER SERIES: CONVERGENCE AT A POINT Proof. On the interval (t − π.3) 2π 0 1 since Dn (τ ) is even and 2π 0 Dn (τ ) dτ = 1 . If f ∈ BV (T) then the 1 Fourier series converges everywhere to 2 [f (t+)+f (t−)]. then Sn (f )(t) → 0 as n → ∞. Thus Fourier series depend only on local information. We have 1 1 π g(t − τ ) − g(t−) Dn (τ ) dτ (8. It suﬃces to prove the theorem for g and h individually. Prove the Fourier series in fact converges uniformly. Proof. t ∈ T.3. 2 Let G(τ ) = g(t + τ ) − g(t+) for τ ∈ (0. Let 0 < δ < π. Let f ∈ L1 (T). express f as the diﬀerence of two bounded increasing functions. Let t ∈ T. so that G is increasing with G(0+) = 0.) Corollary 8. Put H¨lder into Dini: o π −π f (t − τ ) − f (t) dτ ≤ τ π −π (const. if two functions agree on a neighborhood of t and the Fourier series of one of them converges at t. Proof.5 (Localization Principle).δ] . Apply Dini’s Theorem 8.2) Sn (g)(t) − [g(t+) + g(t−)] = 2 2π 0 1 π + g(t + τ ) − g(t+) Dn (τ ) dτ (8. Then (8. (Exercise. |τ | Now apply Dini’s Theorem 8.

provided we show sup Hn n L∞ (T) < ∞. A much deeper result is: Theorem 8.) σ by a change of variable <∞ since limρ→∞ ρ sin σ σ 0 dσ exists. and for h. Hence lim sup |(8. since the function G(τ ).3)| ≤ n 1 sup Hn 2π n 1 sup Hn = 2π n →0 L∞ (T) G(δ) + (0. Therefore (8. We have τ |Hn (τ )| ≤ 0 sin (n + 1 )σ 2 dσ + 1 σ 2 1 (n+ 2 )τ τ 0 1 sin (n + )σ 2 1 sin 1 σ 2 − 1 1 σ 2 dσ ≤2 0 ρ 0 ≤ 2 sup ρ>0 π sin σ σ3 dσ + (const.7 (Carleson–Hunt). −π < τ < −δ.51 as n → ∞. Argue similarly for (8.δ] dG(τ ) since G(0+) = 0 L∞ (T) · 2G(δ) as δ → 0. G(−τ ).2). δ < τ < π. . 0. If f ∈ Lp (T). 1 < p < ∞ then the Fourier series of f converges to f (t) for almost every t ∈ T. −δ < τ < δ. The convergence results so far in this chapter rely just on Riemann– Lebsgue and direct estimates.) 2 dσ σ σ 0 sin σ dσ + (const. Thus we are done. by parts in the ﬁrst term and by the Localization Principle in the last term.3) → 0 as n → ∞. vanishes near the origin.

FOURIER SERIES: CONVERGENCE AT A POINT For p = 1. the result is spectacularly false by Kolmogorov’s Theorem 8. Then it is weak (p. Thus one wants sup |Dn ∗ f | n Lp (T) ≤ Cp f Lp (T) for 1 < p < ∞. Omitted.1. p) for 1 < p < ∞. but that is not good enough to prove Carleson–Hunt! . Proof. and so convergence a. follows from Chapter 6. p). The idea is to prove that the Dirichlet maximal operator (D∗ f )(t) = supn |(Dn ∗ f )(t)| is strong (p.52 CHAPTER 8.e. The next Chapters show sup Dn ∗ f n Lp (T) ≤ Cp f Lp (T) .

53 .1 Theorem 9. (b) If supn Sn B→B = ∞ then there exists f ∈ B whose Fourier series diverges unboundedly: supn Sn (f ) B = ∞. Let B be one of the spaces C(T) or Lp (T). (b) This part follows immediately from the Uniform Boundedness Principle in functional analysis.1.Chapter 9 Fourier series: norm convergence Goals Characterize norm convergence in terms of uniform norm bounds Show norm divergence can occur for L1 (T) and C(T) Show norm convergence for Lp (T) follows from boundedness of the Hilbert transform Reference [Katznelson] Section II. Proof. (a) If supn Sn B→B < ∞ then Fourier series converge in B: n→∞ lim Sn (f ) − f B =0 for each f ∈ B. 1 ≤ p < ∞.

Further. . Consider a sequence fm ∈ C with fm → f .2. For B = C(T) we have Sn C(T)→C(T) = Dn L1 (T) . Let B be any Banach space and assume the Tn : B → B are bounded linear operators.3 (Divergence in C(T)). Example 9. Then Tn f − f ≤ Tn f − Tn fm + Tn fm − fm + fm − f ≤ (A + 1) f − fm + Tn fm − fm < ε whenever n > N . Proposition 9. if g is a trigonometric polynomial then Sn (g) = g whenever n exceeds the degree of g. since we know Dn L1 (T) → ∞. by the following proposition. FOURIER SERIES: NORM CONVERGENCE (a) The collection of trigonometric polynomials is dense in B (as remarked after Theorem 2. We must show f ∈ C.6). Choose ε > 0 and ﬁx m such that fm − f < ε/2(A + 1). If supn Tn B→B < ∞ then C = {f ∈ B : lim Tn f = f in B} n→∞ is closed. Norm Estimates Sn B→B ≤ Dn L1 (T) when B is C(T) or Lp (T). Proof. This upper estimate is not useful. Since fm ∈ C there exists N such that Tn fm − fm < ε/2 whenever n > N . and so C = B. Hence the set C = {f ∈ B : lim Sn (f ) = f in B} n→∞ is dense in B. 1 ≤ p < ∞. since Sn (f ) B = Dn ∗ f B ≤ Dn L1 (T) f B. so that C is closed. Let A = supn Tn B→B . The set C is also closed.54 CHAPTER 9. which proves part (a). as desired.

5. see [Grafakos. n=−∞ . Aside. For B = L1 (T) we have Sn L1 (T)→L1 (T) = Dn L1 (T) . Convergence in Lp (T).1(b)) there exists an integrable function f ∈ L1 (T) whose Fourier series diverges unboundedly in the L1 norm: supn Sn (f ) L1 (T) = ∞. Of course. so that (by Theorem 9. (Thus H is a Fourier multiplier operator. 1 < p < ∞ 1. Fix n.9]. 1 < p < ∞.) That is ∞ Hf ∼ (−i) sign(n)f (n)eint . We shall prove (in Chapters 10–12) the existence of a bounded linear operator H : Lp (T) → Lp (T). called the Hilbert transform on T. For an explicit example of L1 divergence. Proof. with the property (Hf )(n) = −i sign(n)f (n).55 Indeed. this result follows already from the pointwise divergence in Chapter 8. Example 9. for each ε > 0 one can construct g ∈ C(T) that approximates sign(Dn ) (like in Chapter 8). Exercise 3. Then Sn (FN ) = FN ∗ Dn → Dn in L1 (T) as N → ∞. so that (by Theorem 9. FN L1 (T) Therefore supn Sn L1 (T)→L1 (T) = ∞. so that Sn (g) C(T) ≥ |Sn (g)(0)| ≥ Dn L1 (T) −ε g C(T) . Therefore supn Sn C(T)→C(T) = ∞. and so Dn L1 (T) = lim Sn (FN ) N →∞ L1 (T) ≤ Sn = Sn L1 (T)→L1 (T) L1 (T)→L1 (T) .4 (Divergence in L1 (T)).1(b)) there exists a continuous function f ∈ C(T) whose Fourier series diverges unboundedly in the uniform norm: supn Sn (f ) C(T) = ∞.

1) must equal Sm (f ). when 1 < p < ∞.1) e imt f∼ n=−∞ f (n)ei(m+n)t f (n)ei(m+n)t n≥−m P (eimt f ) ∼ e e i(m+1)t −imt P (e imt f) ∼ n≥−m f (n)eint f (n)eint n≥m+1 P (e −i(m+1)t f) ∼ Subtracting the last two formulas gives Sm (f ). The following formula expresses the Fourier partial sum operator in terms of the Riesz projection and some modulations: e−imt P (eimt f ) − ei(m+1)t P (e−i(m+1)t f ) = Sm (f ). Proof. on the right side. Hence from Theorem 9. and we conclude that the left side of (9. Then the Riesz projection P : Lp (T) → Lp (T) deﬁned by 1 1 P f = f (0) + (f + iHf ) 2 2 is also bounded. ∞ (9.) o Observe P projects onto the nonnegative frequencies: Pf ∼ n≥0 f (n)eint since i(−i sign(n)) = sign(n). By the uniqueness result (2. (Note the constant term f (0) is bounded by f Lp (T) . It remains to prove Lp boundedness of the Hilbert transform.5 (Fourier series converge in Lp (T)). 3. From (9. .14).1) has the same Fourier coeﬃcients as Sm (f ).1 we conclude: Theorem 9. FOURIER SERIES: NORM CONVERGENCE 2. the left side of (9. 4.1) and boundedness of the Riesz projection it follows that sup Sm m Lp (T)→Lp (T) ≤2 P Lp (T)→Lp (T) <∞ when 1 < p < ∞.56 CHAPTER 9. by H¨lder’s inequality. Then n→∞ lim Sn (f ) − f Lp (T) =0 for each f ∈ Lp (T). Let 1 < p < ∞.

Hence H L2 →L2 = 1. with Hf 2 L2 (T) = n∈Z |(Hf )(n)|2 = n=0 |f (n)|2 ≤ f 2 2 (Z) = f 2 L2 (T) by Plancherel in Chapter 5. The Hilbert transform on L2 (T) is H : L2 (T) → L2 (T) ∞ f→ n=−∞ − i sign(n)f (n) eint .3 Deﬁnition 10.1. the deﬁnition indeed yields Hf ∈ L2 (T).2 (Adjoint of Hilbert transform). H ∗ = −H 57 . Observe also H 2 (f ) = H(Hf ) = − n=0 f (n)eint = −f + f (0).Chapter 10 Hilbert transform on L2(T) Goal Obtain time and frequency representations of the Hilbert transform Reference [Edwards and Gaudry] Section 6. We call {−i sign(n)} the multiplier sequence of H. Since | sign(n)| ≤ 1. Lemma 10.

ε].4. First.58 CHAPTER 10. g L2 (T) = Hf . −Hg = f. geometric series calculations show that N −1 N − i sign(n)einτ = i n=−N n=−N einτ − i n=1 −iτ einτ =i −e ei(N +1)τ − eiτ −i e−iτ − 1 eiτ − 1 e−i(N +1/2)τ − e−iτ /2 + ei(N +1/2)τ − eiτ /2 =i e−iτ /2 − eiτ /2 1 cos τ − cos (N + 2 )τ 2 = . Proof. taking the limit of integrals over T \ [−ε. That is why (10.1) (10. g ∈ L2 (T). −Hg 2 (Z) 2 (Z) L2 (T) .3. HILBERT TRANSFORM ON L2 (T) Proof. If f ∈ L2 (T) is C 1 -smooth on an open interval I ⊂ T. Proposition 10. 2 But the convolution is ill-deﬁned because the Hilbert kernel cot(t/2) is not integrable.3) sin( τ ) 2 e −i(N +1)τ . g(n) = f (n). Hf. 1 ε→0 2π f (t − τ ) cot ε<|τ |<π τ dτ 2 Remark 10. Formally (10. then (Hf )(t) = 1 2π π [f (t − τ ) − f (t + τ )] cot 0 τ dτ 2 (10. g 2 (Z) 2 (Z) = −i sign(n)f (n).2) says that Hf = f ∗ cot t . i sign(n)g(n) = f .2) = lim for almost every t ∈ I. For f. (10.2) evaluates the convolution in the principal valued sense.

1) now follows.2). Formula (10.1) as (Hf )(t) = lim 1 ε→0 2π π [f (t − τ ) − f (t + τ )] cot ε τ 2 dτ and use oddness of cot(τ /2) to obtain (10. the N th partial sum of Hf is N − i sign(n)f (n) einτ n=−N 1 = 2π 1 = 2π = 1 2π 1 = 2π − N f (τ ) T π n=−N (−i) sign(n)ein(t−τ ) dτ N f (t − τ ) −π π n=−N (−i) sign(n)einτ dτ cos τ 2 by τ → t − τ [f (t − τ ) − f (t + τ )] 0 π − cos (N + 1 )τ 2 dτ sin( τ ) 2 dτ by (10. because the partial sum N −i sign(n)f (n) einτ n=−N converges to Hf (t) in L2 (T) and hence some subsequence of the partial sums converges to (Hf )(t) a. by the C 1 -smoothness of f .59 Second. Now write (10. Hence the second integral tends to 0 as N → ∞ by the Riemann-Lebesgue Corollary 1.3) [f (t − τ ) − f (t + τ )] cot 0 τ 2 1 2π π 0 f (t − τ ) − f (t + τ ) 1 cos (N + )τ dτ. τ sin( 2 ) 2 If t ∈ I then the second integrand belongs to L1 (T) since it is bounded for τ near 0.5. .e.

HILBERT TRANSFORM ON L2 (T) .60 CHAPTER 10.

Notice the cubes in Qk are small when k is large.3 Deﬁnition 11. For k ∈ Z. preparing for a weak (1. That is. 61 . 1)d + m : m ∈ Zd }. 2. Given Q ∈ Qk and j < k. Facts (exercise) 1. there exists a unique Q ∈ Qj with Q ⊂ Q. Call ∪k Qk the collection of dyadic cubes. let Qk = {2−k [0. 1)d + m .Chapter 11 Calder´n–Zygmund o decompositions Goal Decompose a function into good and bad parts.1.5 [Grafakos] Section 4. there exists a unique Q ∈ Qk such that x ∈ Q. For all x ∈ Rd and k ∈ Z. 1) estimate on the Hilbert transform References [Duoandikoetxea] Section 2. there exists a unique m ∈ Zd with x ∈ 2−k [0.

Deﬁnition 11. Clearly Ωk ⊂ Ω. 1 (2−j )d Rd j→−∞ f (y) dy .3. Since Md f ≤ Md |f |. either one of them is contained in the other. λ > 0. We employ a “stopping time” argument like in probability theory for martingales. Ωk = {x ∈ Rd : (Ek f )(x) > λ and (Ej f )(x) ≤ λ for all j < k}. Given two dyadic cubes. we can assume f ≥ 0. let loc (Ek f )(x) = Q∈Qk 1 |Q| f (y) dy 1Q (x). and Ek f dx = Ω Ω f dx (11. CALDERON–ZYGMUND DECOMPOSITIONS 3. because j→−∞ lim (Ej f )(x) ≤ lim =0 < λ. 1).1) whenever Ω is a ﬁnite union of cubes in Qk . And if x ∈ Ω then (Ek f )(x) > λ for some k. Q Theorem 11. let f ∈ L1 (Rd ). For part (a). (b) If f ∈ L1 (Rd ) then limk→∞ (Ek f )(x) = f (x) a. Each cube in Qk contains exactly 2d cubes in Qk+1 . Let Ω = {x ∈ Rd : (Md f )(x) > λ}.62 ´ CHAPTER 11.e.2. Deﬁne the dyadic maximal function (Md f )(x) = sup |(Ek f )(x)| k = sup 1 |Q| f (y) dy : Q is a dyadic cube containing x . a smallest such k exists. Q Then Ek f is constant on each cube in Qk (equalling there the average of f over that cube). loc Proof. For f ∈ L1 (Rd ). (a) Md is weak (1. 4. or else the cubes are disjoint.

g L∞ (Rd ) ≤ 2d λ. Then there exists a “good’ function g ∈ L1 ∩ L∞ (Rd ) and a “bad” function b ∈ L1 (Rd ) such that i.4 (Calder´n–Zygmund decomposition at level λ). b = l bl where bl is supported in a dyadic cube Q(l) and the {Q(l)} are disjoint. 1). Rd Therefore Md is weak (1. just Q(l) ∈ Qk for some k. Hence Ω = ∪k Ωk . Part (b) holds if f is continuous.1).63 Choosing the smallest k implies (Ej f )(x) ≤ λ for all j < k. λ > 0. and so x ∈ Ωk . b (x) dx Q(l) l L1 (Rd ) =0 v. l ≤ 2d+1 λ|Q(l)| 1 λ |Q(l)| ≤ f L1 (Rd ) . b L1 (Rd ) ≤2 f L1 (Rd ) iii. Theorem 11. 1). because disjointness of the cubes is built into the construction. g L1 (Rd ) ≤ f L1 (Rd ) . 1). when proving the dyadic maximal function is weak (1. Note we did not need a covering lemma. and hence if f ∈ L1 (Rd ) by Theorem 6. we do not assume Q(l) ∈ Ql . iv. bl vi. since Ωk equals a union of cubes in Qk (recall Ek f is constant on each cube in Qk ) ≤ f dx. Let f ∈ o 1 d L (R ).4 loc (exercise). using that the dyadic maximal operator Md is weak (1. f = g + b ii. so that |Ω| = k |Ωk | by disjointness of the Ωk Ek f dx since Ek f > λ on Ωk 1 ≤ λ 1 = λ 1 λ k Ωk f dx k Ωk by (11.

note g we have 1 |Q(l)| Q(l) f (y) dy on Q(l). so that |g(x)| ≤ λ. so that x ∈ Ωk for some k. Q(l) |g(x)| dx ≤ Q(l) Hence b L1 (Rd ) = f − g L1 (Rd ) ≤ 2 f L1 (Rd ) . (iii).3 to |f |. Next we show g L∞ (Rd ) ≤ 2d λ. on Rd \ ∪l Q(l). since g = f oﬀ ∪l Q(l) and on Q(l) |f (x)| dx. Hence 1 2d |Q(l)| Q(l) |f (y)| dy ≤ λ (11.3(b). Suppose x ∈ Rd \ ∪l Q(l). 1) estimate that we proved. (iv). For (i). Therefore |g(x)| ≤ 2d λ. on Rd \ ∪l Q(l). these cubes form the collection {Q(l)}. Hence / |f (x)| ≤ λ (for almost every such x) by Theorem 11. Property (vi) is just the weak (1.2) since Q(l) ⊂ Q and side(Q) = 2 side(Q(l)). Deﬁne b(x) = l 1 |Q(l)| f (y) dy 1Q(l) (x) Q(l) bl (x) = f (x) − 0 1 |Q(l)| Q(l) f (y) dy on Q(l). which means 1 |Q| |f (y)| dy ≤ λ Q for some cube Q ∈ Qk−1 with x ∈ Q(l) ⊂ Q. CALDERON–ZYGMUND DECOMPOSITIONS Proof. . and decompose the disjoint sets Ωk into dyadic cubes in Qk . Then (Ek−1 |f |)(x) ≤ λ. Then let g =f −b = For (ii). argue as follows. Let bl (x) = f (x) − so that bl integrates to 0. Together. Then g(x) = f (x). Apply the proof of Theorem 11. Since x ∈ Ωk for all k we have (Ek |f |)(x) ≤ λ for all k. Next suppose x ∈ Q(l) for some l. for each l. f (x) L1 (Rd ) ≤ f L1 (Rd ) . for each l.64 ´ CHAPTER 11. by deﬁnition of g.

1) + m where k ≥ 1. Corollary 11. Let d = 1. We will restrict to “large” λ values. b = l bl where bl is supported in some interval I(l) of the form 2π · 2−k [0.5 (Calder´n–Zygmund decomposition on T). Let f ∈ L1 (T). to get f = g + b.65 For (v). since (Ek |f |)(t) ≤ 1 1 |f (τ )| dτ 2−k 0 2π 1 = 2k |f (τ )| dτ 2π 0 ≤ f L1 (T) since k ≤ 0 <λ . 0. Note Ωk is empty for k ≤ 0.4 to o f (t) = f (2πt). and where the {I(l)} are disjoint. iv. b (t) dt I(l) l L1 (T) =0 4 λ|I(l)| 2π 2π λ v. f = g + b ii. b L1 (T) ≤2 f L1 (T) iii. so that the dyadic intervals have length at most 2π and thus ﬁt into T. 0 ≤ t < 1. g L∞ (T) ≤ 2λ. g L1 (T) ≤ f L1 (T) . λ|Q(l)| Now we adapt the theorem to T. l ≤ |I(l)| ≤ f L1 (T) Proof. 0 ≤ m ≤ 2k − 1. Apply the Calder´n–Zygmund Theorem 11. Then there exists a “good’ function g ∈ L∞ (T) and a “bad” function b ∈ L1 (T) such that i.2). λ > o f L1 (T) . just note |bl (x)| dx ≤ 2 Q(l) |f (x)| dx Q(l) d+1 by deﬁnition of bl ≤2 by (11. bl vi. otherwise.

Ωk ⊂ [0.66 ´ CHAPTER 11. since Ek |f | = 0 outside [0. . Further. CALDERON–ZYGMUND DECOMPOSITIONS by assumption on λ. 1] for k ≥ 1. The Corollary now follows from Theorem 11. with f = g + b yielding f = g + b.4. Thus I(l) = 2πQ(l) has the form stated in the Corollary. 1].

So suppose λ > f L1 (T) . Apply the Calder´n–Zygmund Corollary 11. Hence Hb = l Hbl with convergence in L2 (T). we have |{t ∈ T : |(Hf )(t)| > λ}| ≤ |{t ∈ T : |(Hg)(t)| > λ/2}| + |{t ∈ T : |(Hb)(t)| > λ/2}| = γ + β.Chapter 12 Hilbert transform on Lp(T) Goals Prove a weak (1.3 Theorem 12. There exists A > 0 such that |{t ∈ T : |(Hf )(t)| > λ}| ≤ for all λ > 0 and f ∈ L2 (T). hence Hg ∈ L2 (T) by Chapter 10. 1) on L2 (T)).5 to get f = g + b. Proof. If λ ≤ f L1 (T) then A = 2π works. Further. bl ∈ L2 (T) and b = l bl with convergence in L2 (T). using disjointness of the supports of the bl . Note g ∈ L∞ (T) o and so g ∈ L2 (T). And b = f − g ∈ L2 (T) so that Hb ∈ L2 (T). 1) estimate on the Hilbert transform on T Deduce strong (p. Since Hf = Hg + Hb. p) estimates by interpolation and duality Reference [Duoandikoetxea] Section 3. 67 A f λ L1 (T) .1 (weak (1.

68

CHAPTER 12. HILBERT TRANSFORM ON LP (T)

say. First, use the L2 theory on g:

γ≤

T

**|(Hg)(t)|2 dt (λ/2)2 since H since g since g
**

L2 (T)→L2 (T)

4 ≤ 2 |g(t)|2 dt λ T 8 ≤ |g(t)| dt λ T 8 · 2π ≤ f L1 (T) λ

= 1 by Chapter 10

L∞ (T)

≤ 2λ ≤ f

L1 (T) .

L1 (T)

Second, use L1 estimates on b, as follows:

β≤

l

2I(l) + |{t ∈ T \

l L1 (T)

2I(l) : |(Hb)(t)| > λ/2}| |(Hb)(t)| dt λ/2

4π ≤ f λ

+

T\∪l 2I(l)

by the Calder´n–Zygmund Corollary 11.5(vi) o 4π 2 |(Hbl )(t)| dt ≤ f L1 (T) + λ λ l T\2I(l)

since |Hb| ≤

l

|Hbl | a.e.

To ﬁnish the proof, we show

|(Hbl )(t)| dt ≤ (const.) f

l T\2I(l)

L1 (T) .

(12.1)

**69 By Proposition 10.3 on the interval T \ 2I(l), we have |Hbl (t)| dt
**

T\2I(l)

=

T\2I(l)

1 2π

bl (τ ) cot

I(l)

1 (t − τ ) dτ dt 2

noting t − τ is bounded away from 0, since τ ∈ I(l) and t ∈ 2I(l), / 1 1 1 bl (τ ) cot (t − τ ) − cot (t − cl ) dτ dt = 2 2 T\2I(l) 2π I(l) where cl is the center of I(l), using here that

I(l)

bl (τ ) dτ = 0,

**sin 1 (τ − cl ) 2 dτ dt 1 sin 2 (t − τ ) sin 1 (t − cl ) T\2I(l) I(l) 2 |I(l)| ≤ (const.) |bl (τ )| dtdτ. I(l) R\2I(l) |t − τ ||t − cl | = 1 2π bl (τ ) Note that |t − cl | ≤ |t − τ | + |τ − cl | 1 ≤ |t − τ | + |I(l)| 2 ≤ 2|t − τ | Hence
**

R\2I(l)

when τ ∈ I(l) when t ∈ R \ 2I(l).

|I(l)| dt ≤ 2 |t − τ ||t − cl | =4

R\2I(l) ∞ 2r

|I(l)| dt |t − cl |2 where 2r = |I(l)|

2r dt t2

**= 4. Thus the left side of (12.1) ≤ (const.)
**

l I(l) L1 (T) L1 (T)

|bl (τ )| dτ

= (const.) b ≤ (const.) f

70

CHAPTER 12. HILBERT TRANSFORM ON LP (T)

by the Calder´n–Zygmund Corollary 11.5. o We have proved (12.1), and thus the theorem. Corollary 12.2. The Hilbert transform is strong (p, p) for 1 < p < ∞, with (Hf )(n) = −i sign(n)f (n) for all f ∈ Lp (T), n ∈ Z. Proof. H is strong (2, 2) and linear, by deﬁnition in Chapter 10, and H is weak (1, 1) on L2 (T) (and hence on the simple functions on T) by Theorem 12.1. So H is strong (p, p) for 1 < p < 2 by Remark C.4 after Marcinkiewicz Interpolation (in Appendix C). That is, H : Lp (T) → Lp (T) is bounded and linear for 1 < p < 2. For 2 < p < ∞ we will use duality and anti-selfadjointness H ∗ = −H on 1 1 2 L (T) (see Lemma 10.2) to reduce to the case 1 < p < 2. Write p + p = 1. If f ∈ Lp ∩ L2 (T) then Hf

p

= sup

**1 (Hf )g dt : g ∈ Lp (T) with norm 1 2π T 1 = sup{ (Hf )g dt : g ∈ Lp ∩ L2 (T) with norm 1} 2π T by density of Lp ∩ L2 in Lp 1 = sup{ f (Hg) dt : g ∈ Lp ∩ L2 (T) with norm 1} 2π T since H ∗ = −H on L2 (T) ≤ f
**

Lp (T)

sup{ Hg

Lp (T)

Lp (T)

: g ∈ Lp ∩ L2 (T) with norm 1} by Holder

≤ (const.)p f

by the strong (p , p ) bound proved above, noting 1 < p < 2. Thus H is bounded and linear on the dense subset Lp ∩ L2 (T) of Lp (T). Hence H extends to a bounded operator on Lp (T). Finally, for f ∈ Lp (T), 1 < p < ∞, let fm ∈ Lp ∩ L2 (T) with fm → f in Lp (T). Boundedness of H on Lp implies Hfm → Hf in Lp . Hence fm → f and Hfm → Hf in L1 (T). Thus passing to the limit in (Hfm )(n) = −i sign(n)fm (n) yields (Hf )(n) = −i sign(n)f (n), as desired.

maximal operator. p) for 1 < p < ∞ by the Marcinkiewicz Interpolation Theorem C. by the Marcinkiewicz interpolation and duality argument in Corollary 12. 1 ≤ p ≤ 2. in Chapter 6.2. p p It fails for p > 2 [Katznelson. ∞) by Chapter 6. Fourier analysis.) Strong (p. M is weak (1. for 1 < p < ∞. Section IV. already.3]. Some important applications are: Hilbert transform. p) was proved directly.2. (Note M is sublinear. Hardy–Littlewood maximal operator. Fourier analysis and convolution The Marcinkiewicz and Riesz–Thorin interpolation theorems are covered in Appendix C. 1 1 + = 1. 1) and strong (∞. The Hausdorﬀ–Young theorem says : Lp (T) → p (Z).Chapter 13 Applications of interpolation Goal Apply Marcinkiewicz and Riesz–Thorin interpolation to the Hilbert transform. and hence M is strong (p.2. 71 . H : Lp (T) → Lp (T) is bounded.

72

CHAPTER 13. APPLICATIONS OF INTERPOLATION

To interpret the theorem, note Lp (T) gets smaller as p increases, and so does p (Z). Proof. The analysis operators : L1 (T) → ∞ (Z) and : L2 (T) → 2 (Z) are bounded. Observe 1 1−θ θ = + p 1 2 ⇐⇒ θ 1 =1− 2 p ⇐⇒ 1 1−θ θ = + . p ∞ 2

**Now apply the Riesz–Thorin Interpolation Theorem C.6. Convolution. The Generalized Young’s theorem says f ∗g
**

Lr (Rd )

≤ f

Lp (Rd )

g

Lq (Rd )

when

1 1 1 + = +1, p q r

1 ≤ p, q, r ≤ ∞.

**Proof. Fix g ∈ Lq (Rd ) and deﬁne T f = f ∗ g. Then T is strong (1, q) since f ∗g
**

Lq (Rd )

≤ f

L1 (Rd )

g

Lq (Rd )

**by Young’s Theorem A.3, and T is strong (q , ∞) since f ∗g
**

L∞ (Rd )

≤ f

Lq (Rd )

g

Lq (Rd ) Lq (Rd ) .

by H¨lder’s inequality. In both cases, T ≤ g o 1 1−θ θ = + p 1 q ⇐⇒ θ 1 1 1 =1− = − q p q r

Observe 1 1−θ θ = + . r q ∞

⇐⇒

Now apply the Riesz–Thorin Interpolation Theorem C.6.

**Epilogue: Fourier series in higher dimensions
**

We have studied Fourier series only on the one dimensional torus T = R/2πZ. The theory extends readily to the higher dimensional torus Td = Rd /2πZd . Summability kernels can be obtained by taking products of one dimensional kernels. Thus the higher dimensional Dirichlet kernel is Dn (t) = Dn (t1 ) · · · Dn (td )

n

=

j1 ,...,jd =−n

eijt ,

where j = (j1 , . . . , jd ), t = (t1 , . . . , td )† and † denotes the transpose operation. The Dirichlet kernel corresponds to “cubical” partial sums of multiple Fourier series, because (Dn ∗ f )(t) = =

j1 ,...,jd =−n

1 (2π)d

n

···

T T

Dn (t − τ )f (τ ) dτ1 · · · dτd

f (j)eijt .

“Spherical” partial sums of the form |j|≤n f (j)eijt can be badly behaved. For example, they can fail to converge for f ∈ Lp (Td ) when p = 2. See [Grafakos] for this theorem and more on Fourier series in higher dimensions.

73

74

CHAPTER 13. APPLICATIONS OF INTERPOLATION

Part II Fourier integrals 75 .

.

let f (t) = g(ρt). since g is not periodic. Then f (j) = 1 π g(ρt)e−ijt dt 2π −π ρπ 1 g(y)e−i(j/ρ)y dy = 2πρ −ρπ t ∈ [−π. The outer integral (“Fourier inverse”) is analogous to a Fourier series. We aim to develop a Fourier integral theory that is analogous to the theory of Fourier series. Instead we take a large piece of g and look at its Fourier series: for ρ > 0. π). for |x| < ρπ we have ∞ g(x) = f (ρ x) = j=−∞ −1 f (j)eij(ρ ∞ −1 x) 1 = 2π → 1 2π ρπ −ρπ ∞ −∞ g(y)e−i(j/ρ)y dy ei(j/ρ)x · g(y)e−iξy dy eiξx dξ j=−∞ ∞ −∞ 1 ρ as ρ → ∞. Formally.Prologue: Fourier series converge to Fourier integrals Fourier series do not apply to a function g ∈ L1 (R). and extend f to be 2π-periodic. 77 . by using Riemann sums on the ξ-integral. by changing variable. The inner integral (“Fourier transform”) is analogous to a Fourier coeﬃcient.

78 .

For f ∈ L1 (Rd ) and ξ ∈ Rd . x is a column vector. continuous functions with compact support} C0 (Rd ) = {complex-valued.1) Here ξ is a row vector.1. 79 . Banach space with norm · L∞ (Rd ) Translation fy (x) = f (x − y) Deﬁnition 14. continuous functions with f (x) → 0 as |x| → ∞}.Chapter 14 Fourier transforms: basic properties Goal Derive basic properties of Fourier transforms Reference [Katznelson] Section VI.g. 1/(1 + |x|) is in L2 (R) but not L1 (R) Cc (Rd ) = {complex-valued. and so ξx = ξ1 x1 + · · · + ξd xd equals the dot product.1 Notation f Lp (Rd ) = Rd |f (x)|p dx Nesting of Lp -spaces fails: L∞ (Rd ) ⊂ L2 (Rd ) ⊂ L1 (Rd ) due to behavior at inﬁnity e. (14. deﬁne f (ξ) = Fourier transform of f = Rd 1/p f (x)e−iξx dx.

. j=1 e . f is radial if f (Ax) = f (x) for every x and every orthogonal (“rotation and reﬂection”) matrix A.80 CHAPTER 14. g ∈ L1 (Rd ).4 (Transform of a product). If f1 .2 and that | det A| = 1. by dominated convergence. j=1 fj (xj ) has transform f (ξ) = Proof. A ∈ GL(R. d). FOURIER TRANSFORMS: BASIC PROPERTIES Theorem 14. observe |f (ξ + ω) − f (ξ)| ≤ Rd |f (x)||e−iξx ||e−iωx − 1| dx →0 as ω → 0. . . and so f is uniformly continuous. fd ∈ L1 (R) then f (x) = d d j=1 fj (ξj ). ξ. Equivalently. ω ∈ Rd . Proof. c ∈ C. Suppose A is orthogonal. y ∈ Rd .3 (Transform of a radial function). If f ∈ L1 (Rd ) is radial then f is radial. Use Fubini and the homomorphism property of the exponential: e−iξx = d −iξj xj . Recall that f is radial if it depends only on the distance to the origin: f (x) = F (|x|) for some function F . using Theorem 14. [f (x)eiωx ] (ξ) = f (ξ − ω) takes matrix dilation to its inverse. Proof. For continuity. Linearity (f + g)(ξ) = f (ξ) + g(ξ) and (cf )(ξ) = cf (ξ) Conjugation f (ξ) = f (−ξ) takes translation to modulation. Let f. . Lemma 14. [ | det A|f (Ax) ] (ξ) = f (ξA−1 ) : L1 (Rd ) → L∞ (Rd ) is bounded. Then f (Ax) = f (x) (since f is radial) and so f (ξA−1 ) = [ | det A|f (Ax) ] (ξ) = f (ξ). Exercise.2 (Basic properties). with f L∞ (Rd ) ≤ f L1 (Rd ) f is uniformly continuous If fm → f in L1 (Rd ) then fm → f in L∞ (Rd ). fy (ξ) = e−iξy f (ξ) takes modulation to translation. The convergence is independent of ξ. Corollary 14.

R 1 e−iξx −1 1[−1.5 (Diﬀerence formula).3. Corollary 14. f (ξ) = 1 2 [f (x) − f (x − πξ † /|ξ|2 )] e−iξx dx. g (ξ) = R e−x 2 /2 (−ix)e−iξx dx.1] (x)e−iξx dx = 2 and 1 − cos ξ = 2 sin2 (ξ/2) − x) cos(ξx) dx = 2ξ −2 (1 − cos ξ). so that we can use it later for the third example e−|x| .6.1] (x)e−iξx dx = dx = 2 sin(ξ)/ξ 1 (1 0 2. Proof. 1. Like Lemma 1. 1 ≤ p < ∞. Lemma 14.6 (Continuity of translation). let g(ξ) = R e−x /2 e−iξx dx be the transform we want. Next we compute for the fourth example. Proof. 2 For d = 1. Example 14.8. which is dense in Lp (Rd ).5 implies |f (ξ)| ≤ 1 f − fπξ† /|ξ|2 2 L1 (Rd ) . For ξ = 0.1. Diﬀerentiating.4 except using Cc (Rd ). which tends to zero as |ξ| → ∞ by the L1 -continuity of translation in Lemma 14. The map φ : Rd → Lp (Rd ) y → fy is continuous. 2 4. Rd where ξ † is the column vector transpose of ξ. .7 (Riemann–Lebesgue lemma). R (1 − |x|)1[−1. since ξ † /|ξ|2 has magnitude 1/|ξ| → 0. Note √ g(0) = 2π. Lemma 14. the Gaussian e−|x| /2 . Like Lemma 1. We compute the Fourier transforms in Table 14. Proof.81 Lemma 14. Thus f ∈ C0 (Rd ). f (ξ) → 0 as |ξ| → ∞. Fix f ∈ Lp (Rd ).

4. so that c1 = 1/π.8. 3. which simpliﬁes to the desired result.1] (x) e−|x| e−|x| 2 /2 = sinc2 (ξ/2) (2π)d cd (1+|ξ|2 )(d+1)/2 (2π)d/2 e−|ξ| 2 /2 Table 14. For d = 1. R e−|x| e−iξx dx = 0 e−(1+iξ)x dx + −∞ e(1−iξ)x dx = 1/(1 + iξ) + 1/(1 − iξ). Hence g (ξ) = i R e−x 2 /2 e−iξx dx e−iξx dx e−iξx dx by parts = −i R e−x 2 /2 = −ξ R e−x 2 /2 = −ξg(ξ). with the diﬀerentiation through the integral justiﬁed by using diﬀerence quotients and dominated convergence (Exercise). FOURIER TRANSFORMS: BASIC PROPERTIES dimension 1 1 d d f (x) 1[−1.1: Fourier transforms from Example 14. To handle d > 1. we need a calculus lemma that expresses a decaying exponential as a superposition of Gaussians. cd = Γ d+1 π (d+1)/2 .82 CHAPTER 14.1] (x) f (ξ) 2 sin ξ = 2 sinc ξ ξ sin(ξ/2) ξ/2 2 (1 − |x|)1[−1. note the product structure e−|x| /2 = d e−xj /2 and apply j=1 Lemma 14. 2 2 For d > 1. In the third example. ∞ 0 . The fourth example says the 2 Fourier transform of a Gaussian is a Gaussian. √ 2 Solving the diﬀerential equation yields g(ξ) = 2πe−ξ /2 .

.9 and x → √ ax ∞ a(d−1)/2 e−a/2 (2π)d/2 e−|ξ −(d+1)/2 a|2 /2 by the Gaussian in Table 14.10 (Diﬀerentiation and Fourier transforms). e Proof. For b > 0. Smoothness and decay Theorem 14. The last integral is Γ((d+1)/2).9. so that the transform equals (2π)d cd (1 + |ξ|2 )−(d+1)/2 as claimed in the Table. a e−a/2 −b2 /2a √ e da a e−b(c−1/c) e−b(c−1/c) e−b(c−1/c) e−bu 2 /2 2 /2 dc c−2 dc by letting a = bc2 by c → 1/c 2 /2 2 /2 (1 + c−2 ) dc by averaging the last two formulas where u = c − 1/c du Now we compute the Fourier transform of e−|x| as e−|x| e−iξx dx Rd 1 =√ 2π 1 =√ 2π ∞ 0 e−a/2 √ a e−|x| Rd 2 /2 e−i(ξ √ a)x √ dx ad/2 da da by Lemma 14. b ∞ 0 ∞ 0 ∞ 0 ∞ 0 ∞ −∞ −b 1 =√ 2π ∞ 0 e−a/2 −b2 /2a √ e da.83 Lemma 14.1 0 = (2π)(d−1)/2 (1 + |ξ|2 )/2 ∞ 0 u(d−1)/2 e−u du where u = a(1+|ξ|2 )/2. 1 e √ 2π √ 2 b =√ 2π √ 2 b =√ 2π √ b =√ 2π √ b =√ 2π = 1.

d. where ∂j = ∂/∂ξj for j = 1.84 CHAPTER 14. by dominated convergence with dominating function f (x)|x| ∈ L1 (Rd ). . . . 1 (b) If f ∈ Cc (Rd ) then f (ξ) = o(1/|ξ|) as |ξ| → ∞. . . and |f (ξ)| ≤ f L1 (Rd ) = O(1). Thus: takes multiplication by −ixj to diﬀerentiation. (a) If f ∈ L1 (Rd ) then f (ξ) = o(1) as |ξ| → ∞.1 (Rd )) then (∂j f )(ξ) = iξj f (ξ). Thus: takes diﬀerentiation to multiplication by iξj . Hence f (ξ) has partial derivative (−ixj f )(ξ). FOURIER TRANSFORMS: BASIC PROPERTIES 1 (a) If f ∈ Cc (Rd ) (or more generally. For (b) we compute a diﬀerence quotient. .2. and |f (ξ)| ≤ d maxj ∂j f |ξ| L1 (Rd ) = O(1/|ξ|). d. where ∂j = ∂/∂xj for j = 1. f ∈ W 1. Theorem 14. For (a) (∂j f )(x)e−iξx dx = Rd Rd f (x)(iξj )e−iξx dx by parts = iξj f (ξ). (b) If (1 + |x|)f (x) ∈ L1 (Rd ) then f is continuously diﬀerentiable. . . with (−ixj f )(ξ) = (∂j f )(ξ). . Proof. which is continuous by Theorem 14. with δ ∈ R and ej = unit vector in the j-th direction: f (ξ + δej ) − f (ξ) = δ → Rd f (x)e−iξx Rd e−iδxj − 1 dx δ f (x)e−iξx (−ixj ) dx = (−ixj f )(ξ) as δ → 0.11 (Smoothness of f and decay of f ).

7) and Theorem 14.2. Theorem 14. by Riemann–Lebesgue by Theorem 14. Given f. L1 (Rd ) ≤ f L1 (Rd ) g L1 (Rd ) . (b) For each ξ there exists j such that |ξj | ≥ |ξ|/d (since |ξ1 | + · · · + |ξd | ≥ |ξ|). (a) Use Riemann–Lebesgue (Corollary 14.13 (Convolution and Fourier transforms). If f.2 Convolution Deﬁnition 14. ≤ by Riemann–Lebesgue by Theorem 14. deﬁne their convolution (f ∗ g)(x) = Rd f (x − y)g(y) dy. ξ ∈ Rd .12. x ∈ Rd . g ∈ L1 (Rd ) then f ∗ g ∈ L1 (Rd ) with f ∗g and (f ∗ g)(ξ) = f (ξ)g(ξ). Then |f (ξ)| = |(∂j f )(ξ)| (∂j f )(ξ) ≤ iξj |ξ|/d d maxj |(∂j f )(ξ)| |ξ| = o(1/|ξ|) d maxj ∂j f L1 (Rd ) ≤ |ξ| = O(1/|ξ|).2 Or one could argue more directly using the gradient vector: |f (ξ)| = |( f )(ξ)| = o(1/|ξ|) |iξ| f L1 (Rd ) ≤ |ξ| = O(1/|ξ|).85 Proof. g ∈ L1 (Rd ).

Hence (f ∗ f ) = (f )2 . As this example illustrates.13 predicts. 3. 4. And (f ∗ g)(x) → 0 as |x| → ∞ by dominated convergence. 2. since f (x − y) → 0 as |x| → ∞. Let f = 1[−1/2. Like Theorem 1.86 CHAPTER 14. For the second. and (f ∗ f )(ξ) = sinc2 (ξ/2) by example 2 of Table 14. use Young’s Theorem A.1. then Q(ξ)f (ξ)eiξx dξ. Convolution is commutative: f ∗ g = g ∗ f . 1 ≤ p ≤ ∞. Convolution facts (similar to Chapter 2) 1. convolution is a smoothing operation. Proof. If f ∈ L1 (Rd ) and P (x) = Rd Q(ξ)eiξx dξ for some Q ∈ L1 (Rd ).3. if f ∈ C0 (Rd ) and g ∈ L1 (Rd ) then f ∗ g ∈ C0 (Rd ). Example 14. (P ∗ f )(x) = Rd (14.1. We ﬁnd f (ξ) = sinc(ξ/2) like example 1 of Table 14. If f ∈ Lp (Rd ).1/2] . It is also associative. Proof. then fm ∗ g → f ∗ g in Lp (Rd ). so that (f ∗ f )(x) = (1 − |x|)1[−1. then f ∗ g ∈ Lp (Rd ) with f ∗g Lp (Rd ) ≤ f Lp (Rd ) g L1 (Rd ) . .2) Q(ξ) Rd eiξ(x−y) f (y) dydξ by Fubini = Rd Q(ξ)eiξx f (ξ) dξ. Convolution is continuous on Lp (Rd ): if fm → f in Lp (Rd ). if f ∈ C0 (Rd ) and g ∈ L1 (Rd ) then f ∗ g is continuous because (f ∗ g)(x + z) → (f ∗ g)(x) as z → 0 by uniform continuity of f (exercise). and g ∈ L1 (Rd ). and g ∈ L1 (Rd ). and linear with respect to f and g. 1 ≤ p ≤ ∞.1] (x) by direct calculation. and hence improves the decay of the transform: sinc(ξ/2) decays like 1/ξ while sinc2 (ξ/2) decays like 1/ξ 2 . Rd (P ∗ f )(x) = Proof.14. Use linearity and Fact 2. FOURIER TRANSFORMS: BASIC PROPERTIES Thus the Fourier transform takes convolution to multiplication. as Theorem 14. For the ﬁrst claim. Proof. Further.11.

) 87 .1. (SR1) (SR2) (SR3) sup ω Rd ω→∞ |kω (x)| dx < ∞ |kω (x)| dx = 0 lim {x:|x|>δ} Some kernels further satisfy ω→∞ |x|>δ lim sup |kω (x)| = 0 (L∞ concentration) for each δ > 0.Chapter 15 Fourier integrals: summability in norm Goal Develop summability kernels in Lp (Rd ) Reference [Katznelson] Section VI. A summability kernel on Rd is a family {kω } of integrable functions such that kω (x) dx = 1 Rd (Normalization) (L1 bound) (L1 concentration) for each δ > 0.1 Deﬁnition 15. Here kω (x) does not mean the translation k(x − ω). (SR4) (Notation.

|kω (x)| dx = {x:|x|>δ} {y:|y|>ωδ} |k(y)| dy →0 as ω → ∞.1] (ξ)eiξx dξ 2π R 1 sin x = = sinc x. Show (SR1) and (SR2) by changing variable with y = ωx.1) (15.2) The Dirichlet kernel is Dω (x) = ωD(ωx) = (15.4) . Suppose k ∈ L1 (Rd ) is continuous with Put kω (x) = ω d k(ωx) Rd k(x) dx = 1. Proof.88 CHAPTER 15. For d = 1.1: Dirichlet kernel with ω = 10 2 Example 15. for ω > 0. πx (15. by dominated convergence. dy = ω d dx.2.3. For (SR3). FOURIER INTEGRALS: SUMMABILITY IN NORM 4 2 Figure 15. Example 15.3) (15. Then {kω } is a summability kernel. let D(x) = 1 1[−1. πx π ω 1 eiξx dξ 2π −ω sin(ωx) = .

4. And .2.89 4 2 Figure 15.7) (15. F is integrable since F (x) ∼ x−2 at inﬁnity. See Figure 15. In higher dimensions the Dirichlet function is d D(xj ).6) 1 = 2π sin 1 x 2 1 x 2 2 The Fej´r kernel is e 1 Fω (x) = ωF (ωx) = 2π ω = 2π ω −ω (1 − |ξ|/ω)eiξx dξ 1 ωx 2 1 ωx 2 2 (15. let F (x) = 1 2π (1 − |ξ|)1[−1. Example 15. with associated j=1 d kernel Dω (x) = j=1 Dω (xj ). (15.1. D is not integrable since |D(x)| ∼ |x|−1 at inﬁnity.1.5) by Table 14. ∴ {Dω } is not a summability kernel. For d = 1.1] (ξ)eiξx dξ R (15.2: Fej´r kernel with ω = 10 e 2 See Figure 15.8) sin .

9] for d > 1). for example.12) = cd ω −1 |x|2 + ω −2 (d+1)/2 . In higher dimensions the Fej´r function is e d kernel Fω (x) = j=1 Fω (xj ).14) = (2π)−d/2 e−|x| . e 1 F (x) = 0 Dω (x) dω in 1 dimension. P is integrable since P (x) ∼ |x|−(d+1) at inﬁnity. FOURIER INTEGRALS: SUMMABILITY IN NORM F (x) dx = R 2 lim π ρ→∞ 2 = lim π ρ→∞ 1 = lim π ρ→∞ = 1. G(x) = 1 (2π)d e−|ξ| Rd 2 /2 2 /2 eiξx dξ by Table 14. ∴ {Pω } is a summability kernel.11) (15. See Figure 15. (15. d j=1 F (xj ).5.90 CHAPTER 15.13) (15. by integrating (15.1. p.3. alRd ternatively.3 below.9) (15. P (x) = = 1 (2π)d (1 + e−|ξ| eiξx dξ by Table 14. Example 15. (15.10) directly (see [Stein and Weiss. And P (x) dx = P (0) = 1 because P (ξ) = e−|ξ| by Example 16. Example 15. ρ −ρ ρ −ρ ρ −ρ sin2 (x/2) dx x2 2 sin(x/2) cos(x/2) · (1/2) dx x sin x dx x by parts ∴ {Fω } is a summability kernel.3). with associated The Fej´r kernel is an arithmetic mean of Dirichlet kernels.6. one can integrate (15.10) Rd cd |x|2 )(d+1)/2 The Poisson kernel is Pω (x) = ω d P (ωx) = 1 (2π)d e−|ξ|/ω eiξx dξ Rd (15.1.

18) (15. Connection to Fourier integrals For f ∈ L1 (Rd ): (Dω ∗ f )(x) = 1 (2π)d f (ξ)eiξx dξ [−ω. (15. G is clearly integrable.20) 1 (Fω ∗ f )(x) = (2π)d 1 (Pω ∗ f )(x) = (2π)d 1 (Gω ∗ f )(x) = (2π)d (1 − |ξj |/ω) f (ξ)eiξx dξ [−ω.19) (15. = (2π)d/2 Rd (15.16) See Figure 15.17) (15.4.15) (15.ω]d j=1 e−|ξ|/ω f (ξ)eiξx dξ Rd e−|ξ/ω| Rd 2 /2 f (ξ)eiξx dξ .14).ω]d d G(x) dx = 1 from (15.3: Poisson kernel with ω = 10 2 The Gauss kernel is Gω (x) = ω d G(ωx) = 1 2 e−|ξ/ω| /2 eiξx dξ d (2π) Rd ωd 2 e−|ωx| /2 .91 4 2 Figure 15. and ∴ {Gω } is a summability kernel.

6. Use that if f ∈ C0 (Rd ) then f is uniformly continuous.7 (Summability in Lp (Rd ) and C0 (Rd )).4: Gauss kernel with ω = 10 2 Proof. Recall that C0 (Rd ) uses the L∞ norm. (15. then kω ∗ f → f in Lp (Rd ) as ω → ∞. The left sides of the above formulas make sense for f ∈ Lp (Rd ). Assume {kω } is a summability kernel. . FOURIER INTEGRALS: SUMMABILITY IN NORM 4 2 Figure 15.13).92 CHAPTER 15. respectively. Caution.9). Proof. Use Convolution Fact (14. Argue as for Theorem 2. (15. (b) If f ∈ C0 (Rd ) then kω ∗ f → f in C0 (Rd ) as ω → ∞. 1 ≤ p < ∞.5).2) and deﬁnitions (15. Summability in norm Theorem 15. but the right side does not: so far we have deﬁned the Fourier transform only for f ∈ L1 (Rd ). (a) If f ∈ Lp (Rd ). (15.1).

19) and diﬀerentiate through the integral. Use Fej´r summability (15. Proof.22) xd+1 |x − y|2 + x2 d+1 (d+1)/2 f (y) dy solves 2 2 2 (∂1 + · · · + ∂d + ∂d+1 )v = 0 on Rd × (0. . That is.18)–(15.7 and formulas (15. 1. ∞). Use Theorem 15. Proof. (15. with boundary value v(x.ω]d j=1 in L1 (Rd ). 0) = f (x) in the sense of Theorem 15. ∞).21) Similarly for Poisson and Gauss summability. • Uniqueness theorem: if f.21) on f and g.20). Take ω = 1/xd+1 in (15. the Fourier transform : L1 (Rd ) → L∞ (Rd ) is injective.7. That is.93 Consequences • Fej´r summability for f ∈ L1 (Rd ): e 1 (2π)d d (1 − |ξj |/ω) f (ξ)eiξx dξ → f (x) [−ω. The Poisson kernel solves Laplace’s equation in a half-space: v(x. e Connection to PDEs Fix f ∈ L1 (Rd ). g ∈ L1 (Rd ) with f = g then f = g. xd+1 ) = (P1/xd+1 ∗ f )(x) = cd Rd (15. Proof. using d+1 j=1 ∂ 2 −|ξ|xd+1 iξx (e e ) = (iξ1 )2 + · · · + (iξd )2 + (−|ξ|)2 e−|ξ|xd+1 eiξx ∂x2 j = 0. v is the harmonic extension of f from Rd to the halfspace Rd × (0.

x) = (G1/√2t ∗ f )(x) 1 2 e−|x−y| /4t f (y) dy = d/2 (4πt) Rd solves wt = ∆w for (t.7.94 CHAPTER 15.) √ Proof.20) and diﬀerentiate through the integral. . Take ω = 1/ 2t in (15. x) = f (x) in the sense of 2 2 Theorem 15. x) ∈ (0. FOURIER INTEGRALS: SUMMABILITY IN NORM For the boundary value. using ∂ − ∂t d j=1 ∂2 ∂x2 j (e−|ξ| t eiξx ) = − |ξ|2 − (iξ1 )2 − · · · − (iξd )2 e−|ξ| t eiξx = 0. (Here ∆ = ∂1 + · · · + ∂d . 2. ∞) × Rd . note ω = 1/xd+1 → ∞ as xd+1 → 0. with initial value w(0. The Gauss kernel solves the diﬀusion (heat) equation: w(t. 2 2 √ For the boundary value. note ω = 1/ 2t → ∞ as t → 0.

ˇ 95 ξ ∈ Rd . Deﬁne g (x) = ˇ 1 g(ξ)eiξx dξ (2π)d Rd 1 = g(−x). Rd x ∈ Rd . f ∈ L1 (Rd ) then f is continuous and f (x) = 1 (2π)d f (ξ)eiξx dξ.Chapter 16 Fourier transforms in L1(Rd). in view of the next theorem. g ∈ L1 (Rd ) then g is continuous and ˇ g(ξ) = Rd g (x)e−iξx dx. Theorem 16.2.1] Deﬁnition 16. (Fourier inversion) (a) If f. .1. (b) If g. and Fourier inversion Goal Fourier inversion when f is integrable Reference [Katznelson. Section VI. (2π)d We call ˇ the inverse Fourier transform.

1 displays the results.2(b). because deﬁnitions (15.96 CHAPTER 16. (a) The L1 convergence in Fej´r summability (15. and then swap x and ξ.13) express those kernels as inverse Fourier 2 transforms.13) says G(x) = g (x). change ξ → −ξ. . For example. Example 16.3.1: Fourier transforms of the Fej´r. ˆ The theorem says (f )ˇ = f and (ˇ)ˆ = g . Poisson and Gauss functions. e from Example 16. ˇ Table 16. so that G = g by Theorem 16. (b) Apply part (a) to g.ω]d (ξ) j=1 (1 − |ξj |/ω) f (ξ)eiξx dξ f (ξ)eiξx dξ by dominated convergence. (15.3.e. The Fourier transforms of the Fej´r. if we choose g(ξ) = e−|ξ| /2 then deﬁnition (15. g Proof.2(b). FOURIER INVERSION WHEN F ∈ L1 (RD ) dimension d d d f (x) F (x) = P (x) = 1 (2π)d d j=1 sin(xj /2) xj /2 2 f (ξ) F (ξ) = 1[−1.5). for some subsequence of ω-values: f (x) = lim = 1 ω→∞ (2π)d 1 (2π)d Rd d Rd 1[−ω.9) and (15. Poisson and Gauss e functions can be computed by Fourier Inversion Theorem 16. using that f ∈ L1 (Rd ).1]d (ξ) P (ξ) = e−|ξ| 2 /2 d j=1 (1 − |ξj |) cd (1+|x|2 )(d+1)/2 G(x) = (2π)−d/2 e−|x| G(ξ) = e−|ξ| 2 /2 Table 16.21) implies pointwise e convergence a.

g ∈ L2 (Rd ). for all f.3 Notation Inner product on L2 (Rd ) is f.Chapter 17 Fourier transforms in L2(Rd) Goal Extend the Fourier transform to an isometric bijection of L2 (Rd ) to itself Reference [Katznelson] Section VI. g ˆ (f )ˇ = f. 97 . The Fourier transform L2 (Rd ) → L2 (Rd ) is a bijective isometry (up to a constant factor) with f L2 (Rd ) = (2π)−d/2 f (ˇ)ˆ = g g L2 (Rd ) (Plancherel) (Parseval) (Inversion) f. we conclude the Fourier transform extends to a bounded operator from L2 to itself. g = (2π)−d f . g = Rd f (x)g(x) dx. : Theorem 17. The proof will show : L1 ∩ L2 (Rd ) → L2 (Rd ) is bounded with respect to the L2 norm.1 (Fourier transform on L2 (Rd )). Then by density of L1 ∩ L2 in L2 .

up to a constant factor. g Finally.1] (ξ).20) f (x)e−iξx f (ξ)e−|ξ/ω| /2 dξdx d ω→∞ (2π) Rd Rd 1 2 |f (ξ)|2 e−|ξ/ω| /2 dξ by Fubini. with (f )ˇ = f by Inversion Theorem 16.1) = L (2π)d By density of L1 ∩ L2 in L2 . by repeated use of Theoˆ rem 14. For f ∈ L1 ∩ L2 (Rd ). So the Fourier transform followed by the inverse transform gives the identity on the dense set L1 ∩L2 (Rd ). FOURIER TRANSFORMS IN L2 (RD ) Proof. = lim ω→∞ (2π)d Rd 1 = |f (ξ)|2 dξ by monotone convergence (2π)d Rd 1 f 2 2 (Rd ) . D = (1[−1.1] )ˇ by deﬁnition in (15. note ˇ : L2 (Rd ) → L2 (Rd ) is bounded by Deﬁnition 16.11. sin x πx .1) by density. and so D = 1[−1. or by repeating the argument for Plancherel with f.2.2.7 = lim 1 2 by (15. the Fourier transform extends to a bounded operator from L2 (Rd ) to itself. Thus the Fourier transform is an isometry.1 Inversion. Parseval follows from Plancherel by polarization. Similarly (ˇ)ˆ = g for all g ∈ L2 (Rd ). and surjective by Inversion. If f is smooth with compact support then f is bounded and decays rapidly at inﬁnity. For Inversion. since the Fourier transform is bounded.98 CHAPTER 17. Example 17. using f ∈ L∞ (Rd ). the Fourier transform is injective by Plancherel. f 2 L2 (Rd ) = lim ω→∞ f (x)(Gω ∗ f )(x) dx Rd since Gω ∗ f → f in L2 (Rd ) by Theorem 15. f changed to f.1. the Dirichlet function D(x) = belongs to L2 (R) and has D(ξ) = 1[−1. Plancherel follows from (17. In 1 dimension.1). (17. Hence f ∈ L1 (Rd ). Proof. g (and using dominated instead of monotone convergence).1] by Theorem 17. and hence on all of L2 (Rd ) by continuity.

2.4 (Hausdorﬀ–Young for Fourier transform). If f ∈ L2 (Rd ) then f 1B(n) ∈ L1 ∩ L2 (Rd ) and f 1B(n) → f in L2 (Rd ). using boundedness of : L1 (Rd ) → L∞ (Rd ) : L2 (Rd ) → L2 (Rd ) in Theorem 14. where 1 p + 1 p = 1. . The ﬁrst ﬁve Basic Properties in Theorem 14. Note the Fourier transform is well deﬁned on L1 + L2 (Rd ).1]d (ξ) Table 17. The Fourier transform : Lp (Rd ) → Lp (Rd ) is bounded for 1 ≤ p ≤ 2. Remark 17.3 (radial functions) and Lemma 14.3. from Example 17. since the L1 and L2 Fourier transforms agree on L1 ∩ L2 (Rd ). Apply the Riesz–Thorin Interpolation Theorem C. by Theorem 17. Theorem 17. for almost every ξ.20) (connection to Fourier integrals).1. Hence f (ξ) = lim (f 1B(n) )(ξ) n→∞ in L2 (Rd ). and in Theorem 17. B(n) How can this limit exist.2 still hold for the Fourier transform on Lp (Rd ). Remark 17. Proof.5. when f need not be integrable? The answer must be that oscillations of e−iξx yield cancelations that allow f (x)e−iξx to be integrated improperly.4 (product functions) and (15.1.1: Fourier transform of the Dirichlet function. 1 ≤ p ≤ 2.2. = lim n→∞ f (x)e−iξx dx.17)–(15. as above.6.99 dimension f (x) d D(x) = 1 πd sin xj d j=1 xj f (ξ) D(ξ) = 1[−1. and so do Corollary 14.

6 (Convolution and Fourier transforms).3. If f ∈ L1 (Rd ). Given f ∈ Lp (Rd ). Let m → ∞ and use the Hausdorﬀ–Young Theorem 17. Take gm ∈ L1 ∩ Lp (Rd ) with gm → g in Lp (Rd ). Then (f ∗ gm ) = f gm by Theorem 17.17)–(15.1). Here fm is the usual Fourier transform of fm ∈ L1 (Rd ). Consequence Analogue of Weierstrass trigonometric approximation: functions with compactly supported Fourier transform are dense in Lp (Rd ). then f ∗ g ∈ Lp (Rd ) and (f ∗ g) = f g.100 CHAPTER 17. and (Fω ∗ f ) = Fω f has compact support (because Fω has compact support by Table 16. noting f is bounded.4 and (15. FOURIER TRANSFORMS IN L2 (RD ) Proof. 1 ≤ p ≤ 2.7. Then fm → f in Lp (Rd ) by the Hausdorﬀ–Young Theorem 17. Corollary 17.4. . g ∈ Lp (Rd ). so that Theorem 14.2. 1 ≤ p ≤ 2.4. take fm ∈ L1 ∩ Lp (Rd ) with fm → f in Lp (Rd ). Fω ∗ f → f in Lp (Rd ) by Theorem 15. Proof. Corollary 14. Proof.20) all apply to fm .1. Lemma 14. Now let m → ∞ in those results.

e. and so does the Fej´r kernel e d in 1 dimension. and hence convergence holds at every x.1(a). The Poisson and Gauss kernels satisfy (SR4). Reference [Grafakos] Sections 2.1b. if k(x) = o(1/|x| ) as |x| → ∞ then kω (x) = ω d k(ωx) satisﬁes (SR4) (Exercise). Theorem 3.e. More generally. 101 . by using maximal functions like we did for Fourier series in Chapter 7. Assume {kω } is a summability kernel.7(b). Adapt the corresponding result on the torus. Next we aim at summability a. If f is continuous at x0 ∈ Rd then (kω ∗ f )(x0 ) → f (x0 ) as ω → ∞.. 3.Chapter 18 Fourier integrals: summability pointwise Goal Prove suﬃcient conditions for summability at a single point. Suppose either f ∈ L1 (Rd ) and {kω } satisﬁes the L∞ concentration hypothesis (SR4). But what if f is merely continuous at a point? Theorem 18. or else f ∈ L∞ (Rd ).3b If f ∈ C0 (Rd ) then kω ∗ f → f uniformly by Theorem 15.1 (Summability at a point). and a. Proof.

|B(ω)| (18.2.1) ≤ (M f )(x).4 (Majorization). then |(k ∗ f )(x)| ≤ k L1 (Rd ) (L ∗ f )(x) for all x ∈ Rd . L1/ω (y) = (1/ω)d L(y/ω) = Hence |(L1/ω ∗ f )(x)| ≤ 1 |B(ω)| |f (x − y)| dy B(ω) 1 1B(ω) (y).E. L∗ f ≤ L∗ |f | ≤ M f where M is the Hardy–Littlewood maximal operator from Chapter 6. Lemma 18. Deﬁne the Dirichlet maximal function (D∗ f )(x) = sup |(Dω ∗ f )(x)| ω ω ω Fej´r maximal function (F f )(x) = sup |(Fω ∗ f )(x)| e Poisson maximal function (P f )(x) = sup |(Pω ∗ f )(x)| Gauss maximal function (G f )(x) = sup |(Gω ∗ f )(x)| ω ω ∗ ∗ ∗ Lebesgue maximal function (L∗ f )(x) = sup |(Lω ∗ f )(x)| where L(x) = 1 1B(1) (x) |B(1)| is the normalized indicator function of the unit ball. Deﬁnition 18. . f ∈ L1 (Rd ).102 CHAPTER 18. If k ∈ L1 (Rd ) is nonnegative and radially symmetric decreasing. FOURIER INTEGRALS: SUMMABILITY A. Proof. Lemma 18.3. First.

1 ≤ p ≤ ∞. e 4 ∗ L |f | π P ∗ f ≤ L∗ f G∗ f ≤ L∗ f F ∗f ≤ for all f ∈ Lp (Rd ). Theorem 18.103 Proof. 0 if ω ≤ |y|. Assume ρ is absolutely continuous. because by (18. and Fej´r in 1 dimension). (when d = 1) .1). like we did on the torus in Lemma 7. Write k(x) = ρ(|x|) where ρ : [0. ∞) → R is nonnegative and decreasing. We ﬁrst establish a layer-cake decomposition of k. for simplicity. |B(ω)|(L1/ω ∗ f )(x) − ρ (ω) dω and so ∞ |(k ∗ f )(x)| ≤ 0 ∞ |B(ω)| − ρ (ω) dω · (L∗ f )(x) ω 0 since ρ ≤ 0 = 0 |∂B(1)|rd−1 dr − ρ (ω) dω · (L∗ f )(x) by spherical coordinates for |B(ω)| = B(ω) dy ∞ = 0 |∂B(1)|ω d−1 ρ(ω) dω · (L∗ f )(x) by parts with respect to ω (why does the ω = ∞ term vanish?) = Rd k(y) dy · (L∗ f )(x) by using spherical coordinates again.2: ∞ k(y) = ρ(|y|) = − |y| ∞ ρ (ω) dω since ρ(∞) = 0 by integrability of k =− 0 |B(ω)|L1/ω (y)ρ (ω) dω. L1/ω (y) = Hence (k ∗ f )(x) = 0 ∞ 1/|B(ω)| if ω > |y|.5 (Lebesgue dominates Poisson and Gauss in all dimensions.

even and decreasing.8 (Summability a.3]. G∗ and L∗ are weak (1. The Fej´r kernel is not majorized by a radially symmetric e decreasing integrable function. 1) and strong (p. Thus the best possible radial bound 1 would be O(|x|−2 ).3 with the weak and strong bounds on the Hardy–Littlewood maximal operator in Chapter 6. Theorem 3. Corollary 18. (The last statement is the Lebesgue diﬀerentiation theorem. with k L1 (R) = 4/π.E.4. ≤ k(x) = ω 2π 2 ω x .7. as as as as ω ω ω ω → ∞.4.5 and Lemma 18. F ∗ . which is not integrable at inﬁnity in dimensions d ≥ 2. → ∞.).3.e. 2 Note k is nonnegative.e. Remark 18. P ∗ f ≤ L∗ f by the Majorization Lemma 18. taking ω = 2 gives 1 F2 (x) = π j=1 d sin xj xj 2 .6.) . ω Fω (x) = 2π sin ω x 2 2 ω x 2 def by (15. which decays like x−2 along the x1 -axis. → ∞. see [Grafakos. when d ≥ 2. For the Fej´r kernel in dimensions d ≥ 2. Similarly G∗ f ≤ L∗ f . 1 ≤ p ≤ ∞. FOURIER INTEGRALS: SUMMABILITY A. For example.104 CHAPTER 18. When d = 1. p) on Lp (Rd ). |x| > 2/ω. since Pω is nonnegative and radially symmetric decreasing. a. a. then Fω ∗ f Pω ∗ f Gω ∗ f Lω ∗ f →f →f →f →f a. a. → ∞. P ∗ . If f ∈ Lp (Rd ). Combine Theorem 18. for 1 < p ≤ ∞.e. Hence |Fω ∗ f | ≤ k ∗ |f | ≤ (4/π)L∗ |f | by Majorization Lemma 18.8) 1. e Theorem 18. Proof.e.e. with Pω L1 (Rd ) = 1. 1/ |x| ≤ 2/ω. Proof.

e. Hence the Theorem in Chapter 6 says C = {f ∈ Lp (Rd ) : lim Fω ∗ f = f a. Because C is closed.e. which proves the result. by the part of the theorem already proved.1. on B(m). p) by Corollary 18. put g = 1B(m) f and h = f − g. Then g ∈ L1 (Rd ).e. on B(m). Assume 1 ≤ p < ∞.7. Next h ∈ L∞ (Rd ) is continuous on B(m). Argue similarly for the other kernels. F ∗ is weak (p.7. it must equal Lp (Rd ). For m ∈ N. with h = 0 there.e. When p = ∞. Since f = g + h we deduce Fω ∗ f → f a. and so Fω ∗ g → g a. Letting m → ∞ proves the result. . Hence Fω ∗ g → f a. because Fω ∗f → f uniformly by Theorem 15. Thus C is dense in Lp (Rd ) (using here that p < ∞).} ω→∞ is closed in Lp (Rd ).105 Proof. consider f ∈ L∞ (Rd ). Obviously C contains every f ∈ Cc (Rd ).. and so Fω ∗ h → h = 0 on B(m) by Theorem 18.

106 CHAPTER 18. . FOURIER INTEGRALS: SUMMABILITY A.E.

because if f ∈ ˆ L (Rd ). Sω is the “partial sum” operator for the Fourier integral. then Sω (f ) = (1[−ω. Sω : L2 (Rd ) → L2 (Rd ) p 107 .Chapter 19 Fourier integrals: norm convergence Goal Show norm convergence for Lp (Rd ) follows from boundedness of the Hilbert transform on R Reference I do not know a fully satisfactory reference for this material.ω]d f )ˇ by (15. Write Sω (f ) = Dω ∗ f where Dω (x) = j=1 d d ωD(ωxj ) = sin(ωxj ) πxj j=1 d is the Dirichlet kernel on R and D(z) = (sin z)/πz is the Dirichlet function in 1 dimension. 1 ≤ p ≤ 2. In particular. Suggestions are welcome! Deﬁnition 19.17) and Remark 17.5.1.

Thus Fω ∗ f ∈ A. if f ∈ L1 ∩ Lp (Rd ) then Fω ∗ f ∈ L1 ∩ Lp (Rd ) and (Fω ∗ f ) = Fω f has compact support by Table 16. We conclude A ⊂ {f ∈ Lp (Rd ) : lim Sω (f ) = f in Lp (Rd )} = C. we have the result already for p = 2. by the Generalized Young’s Theorem in Chapter 13. .1. 1 < p < ∞. Since Fω ∗ f → f in Lp (Rd ) by Theorem 15. First we reduce norm convergence to norm boundedness. Theorem 19. Let 1 < p < ∞ and suppose supω Sω Lp (Rd )→Lp (Rd ) < ∞. Because C is closed by Proposition 9. because Dω ∈ q L (Rd ) for each q > 1 and so Dω ∗ f ∈ Lr (Rd ) for each r ∈ (p. We claim A is dense in Lp (Rd ). We further show Sω (g) = g. note Sω (g) ∈ L2 (Rd ) because Dω ∈ L2 (Rd ) and g ∈ L1 (Rd ).2. Our goal in this Chapter is to improve the Lp summability for Fourier integrals (Fω ∗f → f in Theorem 15. which proves the theorem.2 (using the assumption that supω Sω Lp (Rd )→Lp (Rd ) < ∞). To see this fact. Proof. by boundedness of the Fourier transform and its inverse on L2 . We will prove below that Sω (f ) ∈ Lp (Rd ) when f ∈ Lp (Rd ). Sω (f ) is well deﬁned whenever f ∈ Lp (Rd ). Then Fourier integrals converge in Lp (Rd ): limω→∞ Sω (f ) − f Lp (Rd ) = 0 for each f ∈ Lp (Rd ). ω→∞ def so that C is dense in Lp (Rd ). 1 ≤ p < ∞. and L1 ∩ Lp is dense in Lp . thus Sω (g) = Dω g = 1[−ω.ω]d g = g. ∞]. FOURIER INTEGRALS: NORM CONVERGENCE is bounded.7) to Lp convergence (Dω ∗f = Sω (f ) → f in Theorem 19. As remarked above. Let A = {g ∈ L1 ∩ Lp (Rd ) : g has compact support}. Further. we conclude C = Lp (Rd ). as ω → ∞. when g ∈ A. ω]d contains the support of g. 1[−ω. we see A is dense in Lp (Rd ).108 CHAPTER 19.7. Indeed. by Table 17.4 below).ω]d f → f and so Sω (f ) → f in L2 (Rd ). But Sω (f ) need not belong to L1 (Rd ) when f ∈ L1 (Rd ) (Exercise).1 Applying Fourier inversion in L2 gives Sω (g) = g. provided ω is large enough that [−ω.

p p Lp (R) ωD(ω(x − y))g(y) dy dx = Dω ∗ g R R p ≤ Cp. (Thus our “square partial sums” for convergence of Fourier integrals can be relaxed to “rectangular partial sums”.d = sup Sω ω Lp (Rd )→Lp (Rd ) for the norm bound on the partial sum operators. proof omitted. Theorem 19.ω1 ]×···×[−ωd . . .d ≤ (Cp. .1 g p = Cp. Cp.1 (19.) Given a vector ω = (ω1 . .ωd ] is the indicator function of a rectangular box.1 p Lp (R) by deﬁnition of Cp.1 )d . For the sake of generality we allow diﬀerent ω-values in each coordinate direction. We have not yet shown that this constant is ﬁnite. Write Cp.109 Next we reduce to norm boundedness in 1 dimension. R . First observe that for g ∈ Lp (R) and ω > 0. The Fourier multiplier Dω = 1[−ω1 . deﬁne d Dω (x) = j=1 ωj D(ωj xj ). Proof. ωd ) of positive numbers.1) |g(y)|p dy.3 (Reduction to 1 dimension).

(Thus H is a Fourier multiplier operator. Boundedness in Lp (R) 1. y2 )|p dy2 dy1 R R by (19. x2 )|p dx1 dx2 R2 = R p ≤ Cp. |(Dω ∗ f )(x1 . see [Grafakos. FOURIER INTEGRALS: NORM CONVERGENCE Hence for f ∈ Lp (R2 ) and ω = (ω1 . and not “spherical” sums. y2 ). 2 Taking p-th roots gives Sω Lp (R2 )→Lp (R2 ) ≤ Cp. y2 ) dy2 |f (y1 . when working in Lp for p = 2.1]. called the Hilbert transform on R. with the property that (Hf )(ξ) = −i sign(ξ)f (ξ) when f ∈ Lp ∩ L2 (R).110 CHAPTER 19. 1 < p < ∞. We shall prove (in Chapters 20 and 21) the existence of a bounded linear operator H : Lp (R) → Lp (R). Therefore Fourier integrals and series in higher dimensions should be evaluated with “rectangular” partial sums.) 2. which proves the theorem when d = 2. Aside. Section 10.1 ω2 D(ω2 (x2 − y2 ))f (y1 .1 R R ω1 D(ω1 (x1 − y1 )) R ω2 D(ω2 (x2 − y2 ))f (y1 . y2 ) dy2 dy1 dx1 dx2 p p ω2 D(ω2 (x2 − y2 ))f (y1 . y2 ) dy2 dy1 dx2 R R R by (19. when p = 2.1) with g(y2 ) = f (y1 . ω2 ). Argue similarly for d ≥ 3. The “ball” multiplier 1B(1) (ξ) does not yield a partial sum operator with uniform norm bounds. Then the Riesz projection P : Lp (R) → Lp (R) deﬁned by 1 P f = (f + iHf ) 2 .1 .1) with g(y1 ) = R 2p ≤ Cp.

when 1 < p < ∞. e−iωx P (eiωx f ) − eiωx P (e−iωx f ) = Sω (f ). and from boundedness of the Riesz projection. 3. Observe P projects onto the positive frequencies: (P f )(ξ) = 1(0.2) ≤2 P Lp (R)→Lp (R) <∞ when 1 < p < ∞. Let 1 < p < ∞. Proof.∞) (ξ)f (ξ) [eiωx P (e−iωx f )] (ξ) = 1(ω.4 (Fourier integrals converge in Lp (Rd )).1 = sup Sω ω Lp (R)→Lp (R) f ∈ L2 (R).111 is also bounded. Hence from Theorems 19. it follows that Cp.2) applied to the dense class of f ∈ Lp ∩ L2 (R).ω] f .2 and 19.∞) (ξ)f (ξ − ω) [e−iωx P (eiωx f )] (ξ) = 1(0. which equals Sω (f ). [eiωx f ] (ξ) = f (ξ − ω) [P (eiωx f )] (ξ) = 1(0. ω→∞ It remains to prove Lp boundedness of the Hilbert transform on R. Then lim Sω (f ) − f Lp (Rd ) = 0 for each f ∈ Lp (Rd ). From (19. 4. .∞) (ω + ξ)f (ξ) = 1(−ω.∞) (ξ)f (ξ) Subtracting the last two formulas gives 1(−ω. (19.3 we conclude: Theorem 19. since i(−i sign(ξ)) = sign(ξ). Fourier inversion now completes the proof. The following formula expresses the Fourier partial sum operator in terms of the Riesz projection and some modulations: for ω > 0.∞) (ξ)f (ξ). f ∈ L2 (R).

112 CHAPTER 19. FOURIER INTEGRALS: NORM CONVERGENCE .

113 . the Riesz transform equals the Hilbert transform on R.Chapter 20 Hilbert and Riesz transforms on L2(Rd) Goal Develop spatial and frequency representations of Hilbert and Riesz transforms Reference [Duoandikoetxea] Section 4. . deﬁned by H : L2 (R) → L2 (R) f → (−i sign(ξ)f )ˇ because sign(ξ) = ξ/|ξ|. d. . The Riesz transforms on Rd are Rj : L2 (Rd ) → L2 (Rd ) f → (−i(ξj /|ξ|)f )ˇ for j = 1.1. .1 Deﬁnition 20.3 [Grafakos] Section 4. . In dimension d = 1.

1) is a singular integral. Clearly Rj L2 (Rd )→L2 (Rd ) d 2 Rj j=1 ∗ Rj ≤1 = −I = −Rj by Plancherel. by Parseval. HILBERT AND RIESZ TRANSFORMS ON L2 (RD ) Rj is bounded since the Fourier multiplier −iξj /|ξ| is a bounded function (in fact.] For ω > 0. Proof. since the convolution kernel 1/πy is not integrable. 1 2π (−i) sign(ξ)eiξy dξ = [−ω. d since j=1 (−iξj /|ξ|)2 = −1. The proposition says formally that Hf = f ∗ or p. If f ∈ L2 (R) is C 1 -smooth on an interval then (Hf )(x) = p.ω] 0 i i eiξy dξ − 2π −ω 2π 1 − cos(ωy) = . [This proof is similar to Proposition 10. R f (x − y) 1 dy πy (20.2) .3 on T. πy ω 0 eiξy dξ (20.v.114 CHAPTER 20. bounded by 1). 1 πx = −i sign(ξ). Proposition 20. The right side of (20.v.2 (Spatial representation of Hilbert transform).1) for almost every x in the interval. 1 πx Later we will justify these formulas in terms of distributions. and so was skimmed only lightly in class.

7. 1[−ω. . Convergence holds a.1) in full generality by approximating f oﬀ a neighborhood of x using functions in L1 ∩L2 .1) therefore follows from (20.ω] Hf )ˇ converges to Hf . .e.3) as ω → ∞. one deduces (20. ε<|y|<1 Finally. so that (1[−ω. Meanwhile. Rd f (x − y) cd yj dy |y|d+1 for almost every x ∈ U .ω] by Fubini 1 − cos(ωy) dy by y → x − y and (20.3 (Spatial representation of Riesz transform). If f ∈ L2 (Rd ) is C 1 -smooth on an open set U ⊂ Rd then (Rj f )(x) = p. The ﬁrst integral similarly converges to [f (x − y) − f (x)] |y|<1 1 dy. .4).3) and (20.) Proposition 20. for each j = 1. πy (20.ω] Hf )ˇ(x) dξ = = R 1 2π (−i) sign(ξ)f (ξ)eiξx dξ [−ω.4) assuming f is C 1 -smooth on a neighborhood of x (which ensures integrability of y → [f (x − y) − f (x)]/πy on |y| < 1). since (1/πy) dy = 0. Formula (20.2) πy R 1 − cos(ωy) dy = [f (x − y) − f (x)] πy |y|<1 1 − cos(ωy) + f (x − y) dy πy |y|>1 = f (x − y) by oddness of 1 − cos(ωy) /πy. The second integral converges to f (x − y) |y|>1 1 dy πy (20.v. (Obviously f belongs to L1 ∩L2 already on each neighborhood of x.ω] Hf converges to Hf in L2 (R) as ω → ∞.115 If f ∈ L1 ∩ L2 (R) then (1[−ω. .ω] f (y) 1 2π (−i) sign(ξ)eiξ(x−y) dξdy [−ω. for some subsequence of ω-values. d. by the Riemann–Lebesgue Corollary 14. .

HILBERT AND RIESZ TRANSFORMS ON L2 (RD ) Here cd = Γ (d + 1)/2 /π (d+1)/2 > 0.v.5) instead of (20.4 below.6) In class we proceeded formally. Finally. Our proof will use a truncated version of this identity: e−|ξ|δ − e−|ξ|/δ = |ξ| 1/δ δ e−|ξ|z dz.5) and that e−|ξ|z is the Fourier transform of the Poisson kernel P1/z . for some subsequence of δ values. one deduces the theorem for f ∈ L2 (Rd ) by approximating f oﬀ a neighborhood of x using functions in L1 ∩ L2 . cd yj |y|d+1 cd yj |y|d+1 = −i ξj . |ξ| Proof. For f ∈ L2 (Rd ). Applying L2 Fourier inversion yields (Rj f )(x) = lim −iξj δ→0 e−|ξ|δ − e−|ξ|/δ ˆ f ˇ(x) |ξ| in L2 (Rd ). To motivate the following proof. (Rj f )(ξ) = −i ξj f (ξ) |ξ| e−|ξ|δ − e−|ξ|/δ = lim(−iξj ) f (ξ) δ→0 |ξ| with convergence in L2 (Rd ) (by dominated convergence). For example.e. c1 = 1/π.) . observe 1 = |ξ| ∞ 0 e−|ξ|z dz (20. and hence pointwise a.6) in the proof of Lemma 20. The proposition says formally that Rj f = f ∗ or p.4 below. by Lemma 20. Thus the theorem is proved when f ∈ L1 ∩ L2 (Rd ). (Obviously f belongs to L1 ∩ L2 already on each neighborhood of x. skipping the rest of this proof and using (20. (20.116 CHAPTER 20.

so that by the L1 Fourier Inversion Theorem 16. First. and so is f .7) equals ε→0 lim f (x − y) ε<|y|<1 cd yj dy. we express it using .2 (and the deﬁnition of f for f ∈ L1 (Rd )). e−|ξ|δ − e−|ξ|/δ ˆ f ˇ(x) |ξ| 1 e−|ξ|δ − e−|ξ|/δ =− iξj f (y)e−iξy dy eiξx dξ d (2π) Rd |ξ| Rd −|ξ|δ e − e−|ξ|/δ iξy 1 iξj e dξdy =− f (x − y) (2π)d Rd |ξ| Rd −iξj after changing y → x − y. and the exponentials e−|ξ|δ and e−|ξ|/δ are square integrable. If f ∈ L1 ∩ L2 (Rd ) is C 1 -smooth on an open set U ⊂ Rd . the ﬁrst integral in (20.4. ξj /|ξ| is bounded by 1. Further. then e−|ξ|δ − e−|ξ|/δ ˆ f ˇ(x) δ→0 |ξ| cd yj = [f (x − y) − f (x)] d+1 dy + |y| |y|<1 lim −iξj f (x − y) |y|>1 cd yj dy |y|d+1 (20.117 Lemma 20.7) for almost every x ∈ U . To evaluate the inner integral. |y|d+1 Proof. Thus their product is integrable.

And . HILBERT AND RIESZ TRANSFORMS ON L2 (RD ) Poisson kernels: e−|ξ|δ − e−|ξ|/δ iξy 1 iξj e dξ (2π)d Rd |ξ| ∂ 1 e−|ξ|δ − e−|ξ|/δ iξy = e dξ ∂yj (2π)d Rd |ξ| 1/δ = δ 1/δ 1 ∂ ∂yj (2π)d e−|ξ|z eiξy dξdz Rd by identity (20.9) cd yj − f (x − y) 2 + z 2 )(d+1)/2 (|y| |y|>1 dy z=δ where we used the oddness of yj to insert f (x) in (20.12) (why?!) = δ 1/δ ∂ P1/z (y) dz ∂yj cd z ∂ 1 dz 2 + z 2 )(d+1)/2 ∂yj (|y| ∂ 1 dz 2 + z 2 )(d+1)/2 ∂z (|y| z=1/δ = δ 1/δ = δ cd yj cd yj = 2 + z 2 )(d+1)/2 (|y| .8).118 CHAPTER 20.11) by (15. which is integrable).6) by (15.8) (20. z=δ By substituting this expression into the above. expression (20. Now ﬁx a point x ∈ U . we ﬁnd e−|ξ|δ − e−|ξ|/δ ˆ f ˇ(x) −iξj |ξ| =− Rd f (x − y) cd yj (|y|2 + z 2 )(d+1)/2 z=1/δ dy z=δ z=1/δ =− |y|<1 [f (x − y) − f (x)] cd yj (|y|2 + z 2 )(d+1)/2 z=1/δ dy z=δ (20.8) converges to [f (x − y) − f (x)] |y|<1 cd yj dy |y|d+1 by dominated convergence (noting the C 1 -smoothness ensures the integrand is O(|y|) · O(1/|y|d ) = O(1/|y|d−1 ) near the origin. As δ → 0.

(Exercise: explain why the terms with z = 1/δ in (20.) For the ﬁnal claim in the lemma. . Put ∂ u(x. x ∈ Rd . xd+1 ) = (P1/xd+1 ∗ f )(x). Given a function f . expression (20. so that u is harmonic on the upper halfspace Rd × (0.8) and (20. let u(x. The Riesz transforms map the normal derivative of a harmonic function to its tangential derivatives. xd+1 =0 Then Rj v = ∂f . . d. . ∂xj j = 1.9) converges to cd yj dy |y|d+1 f (x − y) |y|>1 by dominated convergence (noting f ∈ L2 (Rd ) and yj /|y|d+1 = O(1/|y|d ) is square integrable for |y| > 1). xd+1 ) ∂xd+1 v(x) = = normal derivative of u at the boundary. ∞) with boundary value u = f when xd+1 = 0 (see Chapter 15). Formal Proof. ε<|y|<1 and use Connections to PDEs 1. using dominated convergence. . xd+1 > 0.119 as δ → 0. write |y|<1 = limε→0 the oddness of yj to remove the term with f (x). .9) vanish as δ → 0.

2. HILBERT AND RIESZ TRANSFORMS ON L2 (RD ) because (Rj v)(ξ) = −i ξj v(ξ) |ξ| ξj ∂ u(ξ. Hence (Rj Rk ∆f ) (ξ) = (−iξj ) (−iξk ) (−|ξ|2 )f (ξ) |ξ| |ξ| = −(iξj )(iξk )f (ξ) ∂ 2f =− (ξ) ∂xj ∂xk so that Rj Rk ∆f = − ∂2f . Formal Proof. ∂xj Thus we have shown the jth Riesz transform maps the normal derivative of u to its jth tangential derivative. . ∂ 2f ∂x2 j 2 (ξ) = (iξj )2 f (ξ) = −ξj f (ξ) and so summing over j gives (∆f ) (ξ) = −|ξ|2 f (ξ). ∂xj ∂xk That is. on the boundary. mixed Riesz transforms map the Laplacian to mixed partial derivatives.120 CHAPTER 20. xd+1 ) = −i |ξ| ∂xd+1 = −i = −i xd+1 =0 ξj ∂ e−|ξ|xd+1 f (ξ) |ξ| ∂xd+1 ξj (−|ξ|)f (ξ) |ξ| xd+1 =0 = iξj f (ξ) ∂f = (ξ). Mixed Riesz transforms map the Laplacian to mixed partial derivatives.

121 The above formal derivation is rigorous if. for example. with ∂ 2f ∂xj ∂xk ≤ ∆f L2 (Rd ) L2 (Rd ) since each Riesz transform has norm 1 on L2 (Rd ). by the Lp boundedness of the Riesz transform proved in the next chapter. Similar estimates hold on Lp (Rd ). . the norm of a mixed second derivative is controlled by the norms of the pure second derivatives in the Laplacian. f is C 2 -smooth with compact support. Consequently. 1 < p < ∞.

122 CHAPTER 20. HILBERT AND RIESZ TRANSFORMS ON L2 (RD ) .

Chapter 21 Hilbert and Riesz transforms on Lp(Rd) Goal Prove weak (1. just changing T to Rd and the interval I(l) to the cube Q(l). . . To ﬁnish the proof.1 (weak (1. and deduce strong (p. There exists A > 0 such that |{x ∈ Rd : |(Rj f )(x)| > ω}| ≤ for all ω > 0. d and f ∈ L1 ∩ L2 (Rd ).4 to get f = g + b. we want to show √ Rd \2 dQ(l) A f ω L1 (Rd ) |(Rj bl )(x)| dx ≤ (const.1. Apply the Calder´n–Zygmund Theorem 11. Proof. p) by interpolation and duality Reference [Duoandikoetxea] Section 5. j = 1. Note o 1 ∞ d g ∈ L ∩ L (R ) and so g ∈ L2 (Rd ). Now proceed like in the proof of Theorem 12. hence Rj g ∈ L2 (Rd ) by Chapter 20. 1) for Riesz transform.1 Theorem 21.) f 123 L1 (Rd ) . 1) on L1 ∩ L2 (Rd )).1) l . And b = f − g ∈ L2 (Rd ) so that Rj b ∈ L2 (Rd ). (21. . .

3 applied on the open set U = Rd \2 dQ(l) (where bl = 0).) Q(l) |bl (y)| dy by Lemma 21. L1 (Rd ) Now (21.2 below.4. o ≤ . HILBERT AND RIESZ TRANSFORMS ON LP (RD ) √ By Proposition 20. / bl (y) ρj (x − y) − ρj (x − c(l)) dy dx Q(l) cd (xj − yj ) dy dx |x − y|d+1 = √ Rd \2 dQ(l) where ρj (x) = cd xj |x|d+1 is the jth Riesz kernel and c(l) is the center of Q(l).124 CHAPTER 21. we have √ Rd \2 dQ(l) |Rj bl (x)| dx bl (y) Q(l) = √ Rd \2 dQ(l) √ noting x − y is bounded away from 0. since y ∈ Q(l) and x ∈ 2 dQ(l).2) ≤ (const.1) follows by summing (21. here we used that b (y) dy = 0.2) over l and recalling that b 2 f L1 (Rd ) by the Calder´n–Zygmund Theorem 11.) |( ρj )(x)| ≤ |x|d+1 √ and if x ∈ Rd \ 2 dQ(l) and y ∈ Q(l) then |x − c(l)| ≥ √ 1 side 2 dQ(l) 2 ≥ 2|y − c(l)|. the hypotheses of that lemma are satisﬁed here because (const. Hence Q(l) l √ Rd \2 dQ(l) |Rj bl (x)| dx √ Rd \2 dQ(l) ≤ Q(l) |bl (y)| |ρj (x − y) − ρj (x − c(l))| dxdy (21.

The Riesz transforms are strong (p.)|y| d+1 2|y| r = (const.125 Lemma 21. p) for 1 < p < ∞. Rj is strong (2. since |x − sy| ≥ |x| − |y| ≥ |x|/2. p) for 1 < p < 2 by Remark C. 1 ρ(x − y) − ρ(x) = 0 ∂ ρ(x − sy) ds ∂s 1 =− 0 y · ( ρ)(x − sy) ds. 1) on L1 ∩ L2 (Rd ) (and hence on all simple functions with support of ﬁnite measure) by Theorem 21. ∞ 1 d−1 r dr = (const. and Rj is weak (1.3. So Rj is strong (p. That is. By the Fundamental Theorem. |x|d+1 x ∈ Rd . Proof. Rj : Lp (Rd ) → Lp (Rd ) is bounded and linear for 1 < p < 2. If ρ ∈ C 1 (Rd \ {0}) with o |( ρ)(x)| ≤ then sup y. . We can take z = 0.) .1.2 (H¨rmander condition). Hence |ρ(x − y) − ρ(x)| dx {x:|x|≥2|y|} 1 ≤ |y| 0 |x|≥2|y| |( ρ)(x − sy)| dxds 1 dx (|x|/2)d+1 ≤ (const. |ρ(x − y) − ρ(x − z)| dx < ∞. 2) and linear. Corollary 21. by a translation.4 after Marcinkiewicz Interpolation (in Appendix C).z∈Rd {x:|x−z|≥2|y−z|} (const. by deﬁnition in Chapter 20. p) estimates.) Now we deduce strong (p.)|y| |x|≥2|y| by using the hypothesis. Proof.

The strong (p. The Riesz kernel cd (xj /|x|)/|x|d ﬁts this form. The idea is to express convolution with this kernel as an average of Hilbert transforms taken in all possible directions in Rd .2.1]. p) bound on the Riesz transform can be generalized to a whole class of convolution-type singular integral operators [Duoandikoetxea. Section 5. HILBERT AND RIESZ TRANSFORMS ON LP (RD ) ∗ For 2 < p < ∞ we use duality and anti-selfadjointness Rj = −Rj on L2 (Rd ) to reduce to the case 1 < p < 2. Alternatively. Section 4. .126 CHAPTER 21. one can instead use the method of rotations [Grafakos.2c]. for singular integral kernels of the form O(x/|x|) |x|d where O is an odd function on the unit sphere. just like in the proof of Corollary 12. since O(y) = yj is odd.

Part III Fourier series and integrals 127 .

**Chapter 22 Compactly supported Fourier transforms, and the sampling theorem
**

Goal Show band limited functions are holomorphic Prove the Kotelnikov–Shannon–Whittaker sampling theorem Reference [Katznelson] Section VI.7

ˇ Deﬁnition 22.1. We say f = g is band limited if g ∈ L1 (Rd ) has compact support. Theorem 22.2 (Band limited functions are holomorphic). Assume g ∈ L1 (Rd ) is supported in a ball B(R), and deﬁne f (z) = g (z) = ˇ 1 (2π)d g(ξ)eiξz dξ

B(R)

for z = x + iy ∈ Cd , x, y ∈ Rd . (Here ξz = ξ1 z1 + · · · + ξd zd .) Then f is holomorphic, and |f (z)| = O(eR|y| ). If in addition g ∈ L2 (Rd ) then |f (z)| = O(eR|y| / |y|). 129

130

CHAPTER 22. BAND LIMITED FUNCTIONS

Thus once more, decay of the Fourier transform (here, compact support) implies smoothness of the function (here, holomorphicity). The theorem also bounds the rate of growth of the function in the complex directions. (The function must vanish at inﬁnity in the real directions, by the Riemann– Lebesgue corollary, since g is integrable.) For example, the Dirichlet kernel D(x) = sin(x)/πx = (1[−1,1] )ˇ(x) is band limited with R = 1, in 1 dimension. Taking z = 0 + iy, we calculate ey − e−y = O(e|y| /|y|), 2πy

D(iy) =

which is better (by a factor of

|y|) than is guaranteed by the theorem.

Proof. f is well deﬁned because ξ → eiξz is bounded on B(R), for each z. And f is holomorphic because eiξz is holomorphic and f can be diﬀerentiated through the integral with respect to the complex variable z. (Exercise. Justify these claims in detail.) Clearly 1 |g(ξ)|e−ξy dξ d (2π) B(R) 1 g L1 (Rd ) eR|y| . ≤ (2π)d

|f (z)| ≤

since eiξz = eiξx e−ξy

If in addition g ∈ L2 (Rd ), then 1 g (2π)d

|f (z)| ≤

L2 (Rd ) B(R)

e−2ξy dξ

1/2

Remark 22.)eR|y| / |y|. that “sampling set” can be a lattice in Rd . and also uniformly (in L∞ (Rd )). Intuitively.4. with f supported in the cube [−ω. to the highest frequency contained in the signal f .R]d ≤ = (2R)d−1 e2R|y| − e−2R|y| 2|y| d−1 2R|y| (2R) e ≤ .131 and e−2ξy dξ = B(R) B(R) e−2ξ|y|e1 dξ by ξ → ξA for some orthogonal matrix A with Ay = |y|e1 = B(R) e−2ξ1 |y| dξ e−2ξ1 |y| dξ [−R. Assume f ∈ L2 (Rd ) is band limited. 1. the sampling rate must be high when the frequencies are high. 2. For a band limited function. sinc(ωxj − πnj ) is centered at the sampling location (π/ω)nj and rescaled to have bandwidth ω. ω]d for some ω > 0. Then d π f (x) = f n sinc(ωxj − πnj ) ω j=1 d n∈Z with the series converging in L2 (Rd ). since sinc ω(π/ω)mj − πnj = sinc π(mj − nj ) = 0. . because many samples are needed to determine a highly oscillatory function. It vanishes at all the other sampling locations (π/ω)mj .3 (Sampling theorem for band limited functions). that is. The sampling rate ω/π is proportional to the bandwidth ω. Theorem 22. Holomorphic functions are known to be determined by their values on lower dimensional sets in Cd . 2 |y| Hence |f (z)| ≤ (const.

and 3 sinc 2πx and 2 sinc 2π(x − 1/2) with dashed curves. The ﬁgure shows f with a solid curve. ξ ∈ [−π. See Remark 22. by replacing x with (π/ω)x (Exercise). f is continuous (after redeﬁning it on some set of measure zero) with f (x) = 1 (2π)d f (ξ)eiξx dξ. π]d . with ω = 2π and sampling rate ω/π = 2.1. Next.1) Thus the pointwise sampled values f (π/ω)n in the theorem are well deﬁned. (22. (22.3. f is square integrable and compactly supported. We can assume ω = π. and so is integrable. A graphical example of the sampling formula is shown in Figure 22.2) . Proof of Sampling Theorem. We will prove f (ξ) = n∈Zd f (−n)eiξn .132 CHAPTER 22.4. The dashed curves are two of the sinc functions making up the signal. for ω = 2π and f (x) = − sinc 2π(x + 1) + 2 sinc 2π(x + 1/2) + 3 sinc 2πx + 2 sinc 2π(x − 1/2) + 1 sinc 2π(x − 1) .1: Example of sampling formula in Theorem 22. BAND LIMITED FUNCTIONS 3 2 1 2 3 2 1 1 2 1 1 2 1 3 2 2 Figure 22. Rd x ∈ Rd . Hence by L1 Fourier inversion. 3.

π]d by the inversion formula (22. if we regard f as a square integrable function on the cube Td = [−π. · L2 . with convergence in L2 (Rd ) and in L1 (Rd ). ω]d ) with inner product (2π)−d ·. and it is a closed subspace (since if f = limm fm in L2 (Rd ) and fm is supported in [−ω. It is isometric. ω]d ). ω]d .ω]d (ξ)e−iξ(π/ω)n n∈Zd . under the Fourier transform.π]d ˇ(x) f (n) n∈Zd = sin(π(xj − nj )) π(xj − nj ) j=1 d with convergence in L2 (Rd ).133 with convergence in L2 ([−π. After changing n → −n in (22. consider the Paley– Wiener space P W (ω) = {f ∈ L2 (Rd ) : f is supported in [−ω. That space has orthonormal Fourier basis (π/ω)d/2 1[−ω.1). then f = limm fm is also supported in [−ω. π]d ). π]d . we have f (ξ) = n∈Zd f (n)e−iξn 1[−π. Paley–Wiener space For a deeper perspective on Sampling Theorem 22. Hence P W (ω) is a Hilbert space with the L2 inner product. ξ ∈ Rd . then its Fourier coeﬃcients are 1 f (ξ)e−iξn dξ (2π)d [−π. Applying L1 inversion gives convergence in L∞ . Thus (22.3. ω]d }.π]d 1 f (ξ)e−iξn dξ = (2π)d Rd = f (−n) since f is supported in [−π.π]d (ξ). Applying L2 inversion gives f (x) = n∈Zd f (n) e−iξn 1[−π. Clearly P W (ω) is a subspace of L2 (Rd ). Indeed.2). .2) simply expresses the Fourier series of f on the cube. to L2 ([−ω.

BAND LIMITED FUNCTIONS where the indicator function simply reminds us that we are working on the cube. . Our calculations have. essentially repeated the proof of the Sampling Theorem. we expand f= n∈Zd f. Taking the inverse Fourier transform gives an orthonormal basis of sinc functions for the Paley–Wiener space: d {gn }n∈Zd = (ω/π) d/2 j=1 sinc(ωxj − πnj ) n∈Zd . (π/ω)d/2 1[−ω.134 CHAPTER 22. Using this orthonormal basis.3. gn L2 gn . of course.ω]d e−iξ(π/ω)n (2π)d = (π/ω)d/2 f (π/ω)n L2 by Parseval by Fourier inversion. for all f ∈ P W (ω).3) simply restates the Sampling Theorem 22. Thus the orthonormal expansion (22. gn L2 = 1 f . (22.3) where the coeﬃcient is f.

and Pe(f ) is 2πZd -periodic. Given f ∈ L1 (Rd ).2. Lemma 23. then Pe(f ) = 2π(21[−π.3. If f ∈ L1 (Rd ) then the series for Pe(f )(x) converges absolutely for almost every x.1 Deﬁnition 23. with Pe(f ) L1 (Td ) ≤ f L1 (Rd ) . x ∈ Rd . Example 23. 135 . Further. its periodization is the function Pe(f )(x) = (2π)d n∈Zd f (x + 2πn). π). if f = 1[−π.Chapter 23 Periodization and Poisson summation Goal Periodize functions on Rd to functions on Td Show the Fourier series of periodization gives the Poisson summation formula References [Folland] Section 8.3 [Katznelson] Section VI. with Pe(f ) extending 2π-periodically to R.1.0) + 1[0.2π) . In 1 dimension. Pe : L1 (Rd ) → L1 (Td ) is bounded.π) ) for x ∈ [−π.

suppose f and g are bounded with compact support. Proof. PERIODIZATION AND POISSON SUMMATION The periodization has Fourier coeﬃcients Pe(f )(j) = f (j).136 CHAPTER 23. For a more direct proof. Proof. If f. That is.) . j ∈ Zd . g ∈ L1 (Rd ) then Pe(f ∗ g) = Pe(f ) ∗ Pe(g). (Thus sums and integrals can be interchanged. Lemma 23. the jth Fourier coeﬃcient of Pe(f ) equals the Fourier transform of f at j. See Problem 19 in Assignment 3.3 again and so Pe(f ∗g) = Pe(f )∗Pe(g) by the uniqueness theorem for Fourier series. below. so that the sums in the following argument are all ﬁnite rather than inﬁnite.3 by Lemma 23. We have Pe(f ∗ g) (j) = (f ∗ g) (j) = f (j) g(j) = Pe(f )(j) Pe(g)(j) = Pe(f ) ∗ Pe(g) (j) by Lemma 23.4 (Periodization of a convolution).

1) (1 + |x|)d+ε C . In particular. Pe(f ∗ g)(x) = (2π)d n∈Zd (f ∗ g)(x + 2πn) f (x + 2πn − y)g(y) dy n∈Zd Rd = (2π)d = Rd Pe(f )(x − y)g(y) dy by deﬁnition of Pe(f ) (Td + 2πm) m = m∈Zd Td Pe(f )(x − y − 2πm)g(y + 2πm) dy since Rd = Pe(f )(x − y)g(y + 2πm) dy m∈Zd Td = = using periodicity of Pe(f ) 1 Pe(f )(x − y) Pe(g)(y) dy (2π)d Td = Pe(f ) ∗ Pe(g) (x).137 For each x ∈ Rd . pass to the general case by a limiting argument. x ∈ Rd . Suppose f ∈ L1 (Rd ) is continuous and decays in space and frequency according to: C |f (x)| ≤ .2) |f (ξ)| ≤ (1 + |ξ|)d+ε for some constants C. remembering that our deﬁnition of convolution on Td has a prefactor of (2π)−d .3. ξ ∈ Rd . (23. Theorem 23. . (23. Then the periodization Pe(f ) equals its Fourier series at every point: (2π)d n∈Zd f (x + 2πn) = j∈Zd f (j)eijx . using that if fm → f in L1 (Rd ) then Pe(fm ) → Pe(f ) in L1 (Td ) by Lemma 23. Finally. ε > 0.5 (Poisson summation formula). x ∈ Rd . taking x = 0 gives (2π)d n∈Zd f (2πn) = j∈Zd f (j).

The Poisson kernel Pr on T equals the periodization of the Poisson kernel Pω on R: 1 − r2 1 ω −1 = 2π . see Chapter 4). Hence the Fourier series of Pe(f ) converges absolutely and uniformly to a continuous function. Example 23. so that it is plausible Pω periodizes to Pr for some r. since by Lemma 23. To complete the proof we will show Pe(f ) is continuous. Pe(f ) has Fourier coeﬃcients in Pe(f )(j) = j∈Zd j∈Zd 1 (Zd ). That continuous function has the same Fourier coeﬃcients as Pe(f ). and so it equals Pe(f ) a. 2 sin πx n∈Z (x + n)2 Proof. The series converges absolutely and uniformly on each ball in Rd (by using (23.4. . exercise). (just like in 1 dimension. to partially motivate these results we note Pe(Pω ∗f ) = Pe(Pω )∗ Pe(f ) by Lemma 23.6 (Periodizing the Poisson kernel). x ∈ R \ Z.138 CHAPTER 23. for then Pe(f ) equals its Fourier series everywhere (and not just almost everywhere). 1 − 2r cos x + r2 π (x + 2πn)2 + ω −2 n∈Z x ∈ R.2) |f (j)| C (1 + |j|)d+ε (const. (23. Proof.) dξ (1 + |ξ|)d+ε ≤ j∈Zd ≤ Rd <∞ by spherical coordinates. First. PERIODIZATION AND POISSON SUMMATION This Poisson summation formula relates a lattice sum of values of the function to a lattice sum of values of its Fourier transform.1).3) provided r = e−1/ω . and so Pe(f ) is continuous. Notice that Pe(f )(x) = (2π)d n∈Zd f (x + 2πn) is a series of continuous functions. Hence we obtain a series expansion for the square of the cosecant: 1 π2 = .e.3 by (23.

3). x ∈ R.1) and (23. Changing x to 2πx in (23. Example 23.7 (Periodizing the Gauss kernel). The Gauss kernel Gs (t) = −j 2 s ijt e on T equals the periodization of the Gauss kernel Gω on R: j∈Z e e−j s eijx = 2π j∈Z n∈Z 2 ω 2 2 √ e−ω (x+2πn) /2 .2) because 1 ω −1 .1.8). ξ ∈ R. which proves (23. (23.139 To prove (23. (x + n)2 + (2πω)−2 1 − 2r cos(2πx) + r2 1 1 +O 2 .3). Hence the Poisson Summation Formula says that Pe(Pω )(x) = j∈Z Pω (j)eijx e−|j|/ω eijx j∈Z = = j∈Z r|j| eijx since r = e−1/ω = Pr (x) by (2.4) . (x + n)2 2 − 2 cos(2πx) (sin πx)2 where we used monotone convergence on the left side. Pω (x) = π x2 + ω −2 Pω (ξ) = e−|ξ|/ω . observe that Pω satisﬁes decay hypotheses (23.12) and Table 16. by (15. ω ω n∈Z Since r = e−1/ω = 1 − letting ω → ∞ implies that n∈Z 1 4π 2 π2 = = .3) gives 1 1 − r2 = 2π 2 ω . 2π x ∈ R.

4) and changing s to πs yields the functional equation for the theta function. Proof.16) and Table 16.4). by (15.140 CHAPTER 23.1) and (23. 2π −(ξ/ω)2 /2 Gω (ξ) = e . 2 πs In terms of the theta function ϑ(s) = presses the functional equation n∈Z . Taking x = 0 in (23. Hence e−n n∈Z 2 πs = s−1/2 n∈Z e−n 2 π/s . x ∈ R. Hence the Poisson Summation Formula says that Pe(Gω )(x) = j∈Z Gω (j)eijx e−(j/ω) j∈Z 2 2 /2 = = j∈Z eijx √ since ω = 1/ 2s e−j s eijx = Gs (x). Decay hypotheses (23. which proves (23.1. . the last formula ex- ϑ(s) = s−1/2 ϑ(s−1 ). PERIODIZATION AND POISSON SUMMATION √ provided s > 0 and ω = 1/ 2s. e−n s > 0. ξ ∈ R.2) hold for Gω because ω 2 Gω (x) = √ e−(ωx) /2 .

1 (Qualitative uncertainty principles). then f ≡ 0. then f ≡ 0. and f is ﬁnitely supported. (a) f is a trigonometric polynomial since it has only ﬁnitely many nonzero Fourier coeﬃcients. 141 . and f is compactly supported. Thus part (a) says: a trigonometric polynomial that vanishes inﬁnitely often in T must vanish identically.Chapter 24 Uncertainty principles Goal Establish qualitative and quantitative uncertainty principles References [Goh and Micchelli] Section 2 [Jaming] Section 1 Uncertainty principles say that f and f cannot both be too localized. f has inﬁnitely many zeros in T. Proof. (b) If f ∈ L2 (Rd ) is continuous. Consequently. and so we are “uncertain” of the value of f . f vanishes on some open set. Proposition 24. (a) If f ∈ L2 (T) is continuous. if f is well localized then f is not.

z ∈ C. write f (t) = where p is the polynomial 2N an eint . If f ∈ L2 (Rd ) is continuous and f and f are supported on sets of ﬁnite measure. f is real analytic on Rd . Therefore the complementary set E = {x ∈ T : f (x + 2πn) = 0 for all n ∈ Z} has positive measure. We prove only the 1 dimensional case. here the support of f need not be compact. Let A = {x ∈ R : f (x) = 0} and B = {ξ ∈ R : f (ξ) = 0}. .2. By dilating f we can suppose |A| < 2π. Theorem 24. Since f has inﬁnitely many zeros t ∈ T. Then {x ∈ T : f (x + 2πn) = 0 for some n ∈ Z} = {x ∈ T : n∈Z 1A (x + 2πn) ≥ 1} ≤ T n∈Z 1A (x + 2πn) dx 1A (x) dx R = = |A| < |T| = 2π. UNCERTAINTY PRINCIPLES N n=−N To prove this claim. That Taylor series equals f on Rd . Then f (t) = p(eit )/eiN t p(z) = n=0 an−N z n .2 (Benedicks’ qualitative uncertainty principle). then f ≡ 0. (b) f is band limited. then the Taylor series of f centered at x0 is identically zero.142 CHAPTER 24. and so f ≡ 0. Proof. The Fundamental Theorem of Algebra implies p ≡ 0. Choose x0 ∈ Rd such that f ≡ 0 on a neighborhood of x0 .1. we see p has inﬁnitely many zeros eit on the unit circle. and hence is holomorphic on Cd by Theorem 22. In contrast to Proposition 24. In particular.

We omit the proof. |[0. Hence when ξ ∈ F . Thus Pe(f e−iξx ) equals some trigonometric polynomial Q(x) a. The Fourier coeﬃcient f (ξ + j) of Q therefore vanishes for all j. 1B (ξ + j) dξ = [0.. Nazarov’s theorem implies Benedicks’ theorem. since ξ ∈ F . .3 (Nazarov’s quantitative uncertainty principle). 1) where F has full measure. which is well deﬁned since f ∈ L1 (R).e. In particular.e. Q vanishes at inﬁnitely many points in T (using here that E has positive measure). which equals zero for but ﬁnitely many j. Theorem 24.e. The jth Fourier coeﬃcient of the periodization equals (f e−iξx ) (j) = f (ξ + j). 1). A constant Cd > 0 exists such that f 2 L2 (Rd ) = f 2 L2 (Rd ) ≤ Cd |A||B|+1 Rd \A |f (x)|2 dx + Rd \B |f (ξ)|2 dξ for all sets A.1) j∈Z R 1B (ξ) dξ = |B| < ∞. so that j∈Z 1B (ξ + j) is ﬁnite for almost every ξ ∈ [0. say for all ξ ∈ F ⊂ [0. B ⊂ Rd of ﬁnite measure and all f ∈ L2 (Rd ). Fix ξ ∈ F and consider the periodization Pe(f e−iξx )(x) = 2π n∈Z f (x + 2πn)e−iξ(x+2πn) . 1) \ F | = 0. because if f is supported in A and f is supported in B. But Pe(f e−iξx )(x) = 0 for all x ∈ E.143 Next. 1).1(a). and so Q vanishes a. and so f ≡ 0. Hence Q ≡ 0 by Proposition 24. Since f (ξ + j) = 0 for all j ∈ Z and almost every ξ ∈ [0. then the right side is zero and so f ≡ 0. the set {j ∈ Z : f (ξ + j) = 0} is ﬁnite. on E. we deduce f (ξ) = 0 a.

Then [T. The minimum is attained for α = T f. g ∈ D(T ∗ ).1) we ﬁnd [T. U ∗ f ≤ U f T ∗f + T f U ∗f . Let H be a Hilbert space. f . U ]f.1) = U f. = norm of component of T f perpendicular to f . U ]f. Let T and U be linear operators like above. Theorem 24. deﬁned on a subspace D(T ∗ ). noting for the adjoints that α= T f. f 2 . f (24.4 (Commutator estimate). U ]f. f f 2 ⇐⇒ α= T ∗ f. g = f. UNCERTAINTY PRINCIPLES Next we develop an abstract commutator inequality that leads to the Heisenberg Uncertainty Principle. T ∗ g Deﬁne ∆f (T ) = min T f − αf α∈C whenever f ∈ D(T ). Note that [T − αI. U − βI] = [T. f / f 2 .144 CHAPTER 24. T ∗ f − T f. β ∈ C. Proof. [T. f ≤ ∆f (T ∗ )∆f (U ) + ∆f (T )∆f (U ∗ ) for all f ∈ D(T U ) ∩ D(U T ) ∩ D(T ∗ ) ∩ D(U ∗ ). Write T ∗ for its adjoint. U ]. Hence by replacing T with T − αI and U with U − βI in (24. f − U T f. Here [T. Let α. f ≤ U f − βf T ∗ f − αf + T f − αf U ∗ f − βf . meaning T ∗ is linear and T f. Minimizing over α and β proves the theorem. Suppose T is a linear operator from a subspace D(T ) into H. f = T U f. U ] = T U − U T is the commutator of T and U .

U ]f = T U f − U T f d d = x · − i f (x) + i xf (x) dx dx = if (x). Here T is the position operator and U is the momentum operator.145 Example 24. 1. with D(U ) = {f ∈ L2 (R) : f ∈ L2 (R)}. The Commutator Theorem 24.2) as restricting how localized f and f can be.6.2) for all α. U ∗ = U and [T. β ∈ C. Observe T ∗ = T. around the locations α and β. Thus the Heisenberg Uncertainty Principle implies that the variance (or uncertainty) in position multiplied by the variance in momentum is at least 1/4.2) if and only if f (x) = iβx −γ(x−α)2 Ce e is a β-modulated Gaussian at α (with C ∈ C. We interpret (24. Roughly. we normalize f L2 (R) = 1 and interpret |f (x)|2 as the probability density for the position x of some particle. . Equality holds in the Heisenberg Principle (24. Remark 24. In quantum mechanics. Squaring yields the Heisenberg Uncertainty Principle: 1 f 4 4 L2 (R) ≤ R |x − α|2 |f (x)|2 dx · 1 2π |ξ − β|2 |f (ξ)|2 dξ R (24.4 implies f 2 L2 (R) ≤ 2∆f (T )∆f (U ) ≤ 2 xf − αf = 2 (x − α)f L2 (R) L2 (R) −if − βf L2 (R) 1 √ (ξ − β)f L2 (R) 2π by Plancherel. the Principle says that the more precisely one knows the position of a quantum particle.5 (Heisenberg Uncertainty Principle). (T f )(x) = xf (x) (U f )(x) = −if (x) with D(T ) = {f ∈ L2 (R) : xf (x) ∈ L2 (R)}. the less precisely one knows its momentum. and vice versa. Take H = L2 (R). γ > 0). and regard |f |2 /2π as the probability density for the momentum ξ.

A more direct proof of (24. in terms of Fourier coeﬃcients of the position density |f |2 : 1 f 4 −2 L2 (T) sup m2 (|f |2 ) (m) m∈Z 2 ≤ n∈Z |n − β|2 |f (n)|2 . 4. The Heisenberg Uncertainty Principle extends naturally to higher dimensions. One considers here a quantum particle at position eit on the unit circle. with momentum n ∈ Z. UNCERTAINTY PRINCIPLES 2. the analogous uncertainty principle says 1 2 1 m 4 2π eimt |f (t)|2 dt T 2 ≤ 1 2π |eimt − α|2 |f (t)|2 dt · T n∈Z |n − β|2 |f (n)|2 for all α.2) can be given by integrating by parts in f 2 L2 (R) = R f (x)f (x)(x − α) dx and then applying Cauchy–Schwarz. β ∈ C. On T. m ∈ Z (exercise). . 3. When α = 0 we deduce a lower bound on the localization of momentum.146 CHAPTER 24.

Part IV Problems 147 .

.

8: Fej´r’s Lemma) p q Let f ∈ L (T) and g ∈ L (T). v → u.1. Thus rapid oscillation yields weak convergence to the mean. but do not hand them in: [Katznelson] Ex. written un u weakly. (b) Then give a quick.Assignment 1 Problem 1.4. for each v ∈ H. Problem 4. Prove that p q 1 m→∞ 2π lim f (mt)g(t) dt = f (0)g(0). m ∈ N. as m → ∞. Problem 2. and deﬁne f(m) (t) = f (mt).2.5: downsampling) Let f ∈ L1 (T). Remark.2. 1. ([Katznelson] Ex. as m → ∞.1. Show f(m) f (0) = (mean value of f ) weakly in L2 (T). (Weak convergence and oscillation) Let H be a Hilbert space. e Problem 3. formal (nonrigorous) proof using the Fourier series of f . imt 2 (a) Show that e 0 weakly in L (T). T Hint. 1. Use that trigonometric polynomials are dense in Lq (T). 149 . (b) Let f ∈ L2 (T). Do the following problems. Clearly if un → u in norm (meaning un − u → 0) then un u weakly. and elementary manipulations. We say un converges weakly to u.1. 1. 1 ≤ q < ∞ and 1 + 1 = 1. Use only the deﬁnition of the Fourier coeﬃcients. 1. (a) Prove that f(m) (n) = f (n/m) if m|n and f(m) (n) = 0 otherwise. v as n → ∞. if un . where 1 < p ≤ ∞. ([Katznelson] Ex.

Problem 7. if f is smooth except for ﬁnite many corners (such as the triangular wave f (t) = |t| for t ∈ (−π. L (b) Deduce that f ∈ A(T). . (Smoothness of f implies rate of decay of f ) (a) Show that if f has bounded variation. Hint. so that Sn (f ) → f in L∞ (T). Hence the Fourier series of f converges uniformly by Chapter 4. Remark. These results cover most of the functions encountered in elementary courses. (b) Show that if f is absolutely continuous and f has bounded variation. with f ∈ C α (T) for some 0 < α < 1. In particular. t ∈ (−π.2 (T). Problem 8.2: rate of uniform summability) Assume f is H¨lder continuous. 1 − α Nα N ∈ N. functions that are smooth expect for ﬁnitely many jumps (such as the sawtooth f (t) = t. And functions that are smooth except for ﬁnitely many corners (such as the triangular wave f (t) = |t|.4) Let f be absolutely continuous on T with f ∈ L2 (T). ([Katznelson] Ex. (a) Prove that ∞ f 1 (Z) ≤ f L1 (T) + 2 n=1 1 n2 1/2 f L2 . π]) have bounded variation. f ∈ W 1. For example.3. then f (n) = O(|n|−1 ). First evaluate f 2 2 .150 Problem 5. Problem 6. t ∈ (−π. 1. That is why one encounters so many functions with Fourier coefﬁcients decaying like 1/n or 1/n2 . π]) have ﬁrst derivative with bounded variation.5. then the Fourier series converges uniformly to f . In other words. then f (n) = O(|n|−2 ). Remark. Remark. the faster σN (f ) converges to f as N → ∞. o Prove there exists C > 0 (depending on the H¨lder constant of f ) such that o σN (f ) − f L∞ ≤ 1 C . ([Katznelson] Ex. (A lacunary series) Assume 0 < α < 1. π]). Thus the “smoother” f is. 1.

show f (n) = O(|n|−β ) fails for some f ∈ C α (T). which shows how o sums of Fourier series having diﬀerent underlying periods can be used to model the heights of tides. and then f ∈ C α (T). Harald Bohr won a silver medal at the 1908 Olympics. Show f ∈ C α (T). Read Chapter 8 “Compass and Tides” from [K¨rner].151 (a) Suppose that f is continuous on T and that |f (j)| ≤ C2−nα 2n ≤|j|<2n+1 for each n ≥ 0. (Maximal function when p = 1) Deﬁne L log L(Rd ) to be the class of measurable functions for which |f (x)| log 1 + |f (x)| dx < ∞. brother of physicist Niels Bohr. Deduce that the rate n=0 of decay f (n) = O(|n|−α ) proved in Theorem 1. . Their theory was developed by the Danish mathematician Harald Bohr. n (b) Let f (t) = ∞ 2−nα ei2 t . (That is. in soccer.) Problem 9. Problem 10. Thus if the singularities of f are “logarithmically better than L1 ” then the Hardy–Littlewood maximal function belongs to L1 (at least locally). Enjoyable reading (nothing to hand in). when β > α.6 for C α (T) is sharp. loc Remark. Sums of periodic functions having diﬀerent periods are called almost periodic functions. Prove f ∈ A(T). Prove that Rd f ∈ L log L(Rd ) =⇒ M f ∈ L1 (Rd ).

152 .

Assignment 2 Problem 11 (Hilbert transform of indicator function). (In your solution. (a) Let 1 ≤ p ≤ 2. N →∞ 2π T |n|≤N 153 .b] )(t). Prove that the Fourier synthesis operator T . for t ∈ [−π. Let 1 ≤ p ≤ 2. Prove the Parseval identity 1 f (t)g(t) dt = lim f (n)g(n). Take f ∈ Lp (T) and g ∈ L1 (T) with {g(n)} ∈ p (Z). Estimate the norm of T . in Lp (T). Problem 12 (Fourier synthesis on p ). deﬁned by (T {cn })(t) = n∈Z cn eint . (b) Conclude that the Hilbert transform on T is not strong (∞. Prove that g ∈ Lp (T). where [a. b] ⊂ (−π. You may do both parts if you wish.) (b) Let 1 < p < ∞. ∞). Do part (a) or part (b). Take f ∈ Lp (T) and g ∈ Lp (T). (a) Evaluate (H 1[a. Sketch the graph. explain why the integral and sum are absolutely convergent. and establish the Parseval identity 1 2π f (t)g(t) dt = T n∈Z f (n)g(n). π) is a closed interval. Problem 13 (Parseval on Lp ). π]. Show the series converges unconditionally. Extra credit. is bounded from p (Z) to Lp (T).

Hint. (a) Show 1/p |f (n)| |n| n=0 p p−2 ≤ Cp f Lp (T) for all f ∈ Lp (T). Y = Z \ {0} with ν = counting measure weighted by n−2 . Assume u is a smooth. 2π D n∈Z Hint. (b) Show that combining the H¨lder and Hausdorﬀ–Young inequalities in o the obvious way does not prove part (a). Recall Pr denotes the Poisson kernel on T. real-valued function on a neighborhood of D. Let 1 < p ≤ 2. no credit] Assume f ∈ C ∞ (T) and show v ∈ C ∞ (D). so that v is deﬁned on the closed disk D. Problem 16 (Boundary values lose half a derivative). Problem 15 (Poisson extension). (Parts (a) and (b) show v is smooth on D and continuous on D. Use one of Green’s formulas. t ∈ T. t ∈ T. (Pr ∗f )(t) is called the harmonic extension to the disk of the boundary function f . Hence f ∈ C ∞ (T). Suppose f ∈ C(T) and deﬁne (Pr ∗ f )(t) for 0 ≤ r < 1. and remember v = v since f and v are real-valued.). Thus the task is to prove each partial derivative of v on D extends continuously to D. and write D for the open unit disk in the complex plane. and deﬁne f (t) = u(eit ) for the boundary value function of u.154 Problem 14 (Fourier analysis into a weighted space). v(reit ) = f (t) for r = 1. (a) Prove 1 | v|2 dA = |n||f (n)|2 . (c) [Optional. . Aside. (a) Show v is C ∞ smooth and harmonic (∆v = 0) in D. (b) Show v is continuous on D. and so the Poisson extension v belongs to C ∞ (D) by Problem 15(c).

(We do not have such a norm estimate on {nf (n)}. Write u = v + (u − v) and use one of Green’s formulas. compared to the original function. this result holds on arbitrary domains. As your proof reveals. Thus the boundary value loses half a derivative. and if f has one derivative (f ∈ L2 (T)) then {nf (n)} ∈ 2 (Z). then u has half a derivative in L2 on the boundary. since {|n|1/2 f (n)} ∈ 2 (Z). Boundary trace inequalities like in part (c) are important for partial differential equations and Sobolev space theory. f ∈ C ∞ (T) and so certainly f ∈ L2 (T). You might wonder. Aside. Hint. why you should bother proving the weaker result {|n|1/2 f (n)} ∈ 2 (Z) in part (c). the harmonic function has smallest Dirichlet integral. basically. The notion of fractional derivatives deﬁned via Fourier coeﬃcients can be extended to fractional derivatives in Rd . The inequality says. 2π D n∈Z Discussion. We say f has “half a derivative” in L2 . (c) Conclude 1 |n||f (n)|2 ≤ | u|2 dA.155 (b) Prove | v|2 dA ≤ D D | u|2 dA. This result is known as “Dirichlet’s principle”. then. Justiﬁcation: if f has zero derivatives (f ∈ L2 (T)) then {f (n)} ∈ 2 (Z). Aside.) This norm estimate means that the restriction map H 1 (D) → H 1/2 (∂D) u→f is bounded from the Sobolev space H 1 (D) on the disk with one derivative in L2 to the Sobolev space H 1/2 (∂D) on the boundary circle with half a derivative in L2 . by using Fourier transforms. that if a function u has one derivative u belonging to L2 on a domain. But actually you prove more in part (c): you obtain a norm estimate on {|n|1/2 f (n)} ∈ 2 (Z) in terms of the L2 norm of u. . Halfway inbetween lies the condition {|n|1/2 f (n)} ∈ 2 (Z). which implies {nf (n)} ∈ 2 (Z). It asserts that among all functions having the same boundary values. Note that in this problem.

156 Problem 17 (Measuring diameters of stars). Enjoyable reading; nothing to hand in. Read Chapter 95 “The Diameter of Stars” from [K¨rner], which shows o how the diameters of stars can be estimated using Fourier transforms of radial functions, and convolutions.

Assignment 3

Problem 18 (Adjoint of Fourier transform). Find the adjoint of the Fourier transform on L2 (Rd ). Problem 19 (Periodization, and Fourier coeﬃcients and transforms). Suppose f ∈ L1 (Rd ). (a) Prove that the periodization Pe(f )(x) = (2π)d

n∈Zd

f (x + 2πn)

of f satisﬁes Pe(f )

L1 (Td )

≤ f

L1 (Rd ) .

(b) Deduce from your argument that the series for Pe(f )(x) converges absolutely for almost every x, and that Pe(f ) is 2πZd -periodic. (c) Show that the jth Fourier coeﬃcient of Pe(f ) equals the Fourier transform of f at j: Pe(f )(j) = f (j), j ∈ Zd Problem 20 (Course summary). Write a one page description of the most important and memorable results and general techniques from this course. Be brief, but thoughtful; explain how these main results ﬁt together. You need not state the results technically — intuition is more helpful than rigor, at this stage.

157

158

Part V Appendices 159 .

.

µ) is simply the triangle inequality for Lp (X). ν) are σ-ﬁnite measure spaces. and that f (x. and apply it to norms of convolutions Minkowski’s inequality on a measure space (X. If 1 ≤ p ≤ ∞ then f (x. the norm of an integral is bounded by the integral of the norms: Theorem A.Appendix A Minkowski’s integral inequality Goal State Minkowski’s integral inequality. Suppose (X.1. Proof. Similarly. with 161 1 p + 1 q = 1. y) is measurable on the product space X × Y . µ) and (Y. Take q to be the conjugate exponent. saying that the norm of a sum is bounded by the sum of the norms: fj νj j Lp (X) ≤ j fj Lp (X) νj whenever fj ∈ Lp (X) and the constants νj are nonnegative. y) Lp (X) dν(y) whenever the right side is ﬁnite. y) dν(y) Y Lp (X) ≤ Y f (x. Then for all .

the convolution f ∗ g is well deﬁned a. y) dν(y) g(x) dµ(x) X Y ≤ Y X |f (x. Lp (T) whenever the right sides are ﬁnite. (f ∗ g)(t) = 2π T whenever the integral makes sense. g L1 (Rd ) . In particular. y) Lp (X) dν(y) · g Lq (X) .162 g ∈ Lq (X).e. Deﬁnition A. Theorem 6. Theorem A. Argue similarly on T. whenever the integral makes sense. Deﬁne the convolution of functions f and g on T by 1 f (t − τ )g(τ ) dτ. Lp (Rd ) g L1 (Rd ) .2. y)||g(x)| dµ(x)dν(y) |f (x. x ∈ Rd .1. Then f ∗ g Lp (T) ≤ f f ∗ g Lp (Rd ) ≤ f g L1 (T) . f ∗g Lp (Rd ) = Rd f (· − y)g(y) dy f (· − y) Rd Lp (Rd ) ≤ = f Lp (Rd ) |g(y)| dy Lp (Rd ) by Minkowski’s integral inequality. Fix 1 ≤ p ≤ ∞. Proof. Theorem A.14]). Deﬁne the convolution of functions f and g on Rd by (f ∗ g)(x) = Rd f (x − y)g(y) dy. Now the theorem follows from the dual characterization of the norm on Lp (X) (see [Folland. t ∈ T. whenever f ∈ Lp and g ∈ L1 . y)|p dµ(x) Y X 1/p ≤ = Y g Lq (X) dν(y) by H¨lder o f (x. .3 (Young’s theorem). MINKOWSKI’S INTEGRAL INEQUALITY f (x. APPENDIX A.

Lemma B. Let α > 0. write E(λ) = {x ∈ X : |f (x)| > λ} for the level set of f above level λ.Appendix B Lp norms and the distribution function Goal Express Lp -norms in terms of the distribution function Given a σ-ﬁnite measure space (X. X (B.1. (B.3) . X (B. when r = 0 < p < ∞ and α = 1. formula (B.2) In particular. µ) and a measurable function f on X. The distribution function of f is µ(E(λ)).1) expresses the Lp -norm in terms of the distribution function: ∞ 0 pλp−1 µ(E(λ)) dλ = X |f (x)|p dµ(x) = f 163 p Lp (X) .1) If −∞ < p < r < ∞ then ∞ 0 λp−r−1 E(λ/α)c |f (x)|r dµ(x)dλ = αp−r r−p |f (x)|p dµ(x). If −∞ < r < p < ∞ then ∞ 0 λp−r−1 E(λ/α) |f (x)|r dµ(x)dλ = αp−r p−r |f (x)|p dµ(x).

Thus we have proved (B. Write E = {(x. λ)|f (x)|r dµ(x)dλ ∞ 0 = X |f (x)|r |f (x)| X r 0 λp−r−1 1E (x. ∞) : |f (x)| > λ}.164 APPENDIX B. λ) dλdµ(x) λp−r−1 dλdµ(x) by Fubini since λ < |f (x)| on E |f (x)| = = X |f (x)|r 1 |f (x)|p−r dµ(x) p−r since p − r > 0. .2) is similar. λ) ∈ X × (0. λ) ∈ E ⇔ x ∈ E(λ). so that (x. LP NORMS AND THE DISTRIBUTION FUNCTION Proof. Then the left hand side of (B. and (B.1). We can assume α = 1 without loss of generality. by changing variable with λ → αλ.1) equals ∞ 0 λp−r−1 X 1E (x.

ν) are measure spaces. and all y in the underlying set. assuming either weak endpoint bounds (Marcinkiewicz) or strong endpoint bounds (Riesz–Thorin) References [Folland] Chapter 6 [Grafakos] Section 1. Let 1 ≤ p0 < p1 ≤ ∞ and suppose (X.3 Deﬁnition C. then T is strong (p. p1 ). Theorem C.2 (Marcinkiewicz Interpolation). p) whenever p0 < p < p1 . g in the domain of T . If T is weak (p0 . 165 .Appendix C Interpolation Goal Interpolation of operators on Lp spaces. p0 ) and weak (p1 . all c ∈ C. µ) and (Y. Assume T : Lp0 + Lp1 (X) → {measurable functions on Y } is sublinear.1. An operator is sublinear if |T (f + g)(y)| ≤ |(T f )(y)| + |(T g)(y)| |T (cf )(y)| = |c||(T f )(y)| for all f.

p1 ) estimates. Let α > 0. |f (x)|p0 dµ(x) |f (x)|p1 dµ(x). h ∈ Lp1 (X) since |h|p1 ≤ |f |p (λ/α)p1 −p . We have proved strong (p.2). p). Then ν ({y ∈ Y : |T f (y)| > λ}) ≤ ν ({y ∈ Y : |T g(y)| > λ/2}) + ν ({y ∈ Y : |T h(y)| > λ/2}) by sublinearity p0 p1 A0 A1 g Lp0 (X) h Lp1 (X) ≤ + by the weak estimates on T λ/2 λ/2 = (2A0 )p0 λ−p0 {x:|f (x)|>λ/α} and h = f 1{x:|f (x)|≤λ/α} . Write A0 . p /(p −p ) p /(p −p ) Choosing α = 2A11 1 0 /A00 1 0 gives simple constants: Tf Lp (Y ) ≤ 2p1/2 1 1 + p − p0 p1 − p 1/p A1−θ Aθ f 0 1 Lp (X) (C. |T f | ≤ |T g| + |T h|.2) blows up as p approaches p0 or p1 .1) and formulas (B. λ > 0. Split f into “large” and “small” parts: g = f 1{x:|f (x)|>λ/α} Notice that g ∈ Lp0 (X) since |g|p0 ≤ |f |p (λ/α)p0 −p . INTERPOLATION Proof.1). {x:|f (x)|≤λ/α} ∞ + (2A1 )p1 λ−p1 Therefore Tf p Lp (Y ) (C. (B. p p0 p1 Note the estimate in (C. A1 for the constants in the weak (p0 . Consider f ∈ Lp (X). . By sublinearity. Case 1. Hence f = g + h ∈ Lp0 + Lp1 (X).166 APPENDIX C. p0 ) and (p1 .1) = 0 pλp−1 ν ({y ∈ Y : |T f (y)| > λ}) dλ αp−p0 f p − p0 p Lp (X) ≤ p(2A0 )p0 + p(2A1 )p1 αp−p1 f p1 − p p Lp (X) by (C.2) 1 where 0 < θ < 1 is determined by expressing p as a convex combination of 1 and p11 : p0 1−θ θ 1 = + . Assume p1 < ∞.

Given a measure space (X. p0 ) and weak (p1 . Let α = 2A1 . write Σ(X) = {simple functions on X with support of ﬁnite measure}. . Proof. p). And T f = T g + T h by linearity. Suppose T : Σ(X) → {measurable functions on Y } is linear. and the sets Fj have ﬁnite measure and are disjoint. Remark C. we deduce T has a unique extension to a bounded linear operator on Lp (X). 1]. That is. Hence the proof of Marcinkiewicz Interpolation gives a strong (p. then so are g = f 1{|f |>λ/α} and h = f 1{|f |≤λ/α} . (This extension step uses linearity of T .5 (Hadamard’s Three Lines). f ∈ Σ(X) provided f = αj 1Fj where the sum is ﬁnite. ∞).3. If f is simple with support of ﬁnite measure. and so Th L∞ (Y ) ≤ A1 λ λ = α 2 by deﬁnitions of h and α. µ). By density of Σ(X) in Lp (X) (using here that p < p1 implies p < ∞). 1−θ θ Then |H(z)| ≤ B0 B1 whenever Re(z) = θ ∈ [0. We have T h L∞ (Y ) ≤ A1 h L∞ (X) . Deﬁnition C.167 Case 2. p) on Lp (X) whenever p0 < p < p1 . p1 ) on the simple functions in Σ(X). Let B0 = supRe(z)=0 |H(z)| and B1 = supRe(z)=1 |H(z)|. then T is strong (p. Hence ν ({y ∈ Y : |T f (y)| > λ}) ≤ ν ({y ∈ Y : |T g(y)| > λ/2}) because |T f | ≤ |T g| + |T h|. Assume H(z) is holomorphic on U = {z ∈ C : 0 < Re(z) < 1} and continuous and bounded on U = {z ∈ C : 0 ≤ Re(z) ≤ 1}. because weak (∞.) Our next interpolation result needs: Lemma C.4 (Linear Operators). Next we weaken the hypotheses of Marcinkiewicz Interpolation. Now argue like in Case 1 to get strong (p. ∞) is deﬁned to mean strong (∞. Then Marcinkiewicz Interpolation still holds: if T is weak (p0 . αj ∈ C \ {0}. Assume p1 = ∞. p) bound for T on Σ(X).

Let Bθ = supRe(z)=θ |H(z)|.168 APPENDIX C. q1 ≤ ∞.) Proof. p1 . then add ε to H and argue as above. and (X. . −(y 2 +1)/m Hence Gm → 0 as |z| → ∞ in U . q) whenever 1 1 1 1 1 1 . Then G(z) = H(z) 1−z z B0 B1 1−z z is holomorphic on U and bounded on U . Assume B0 > 0 and B1 > 0. so that Bθ ≤ B0 B1 by the Lemma. Speciﬁcally. If T is strong (p0 . Let 1 ≤ p0 . Let ε → 0. then further assume ν is semi-ﬁnite.) Suppose T : Lp0 + Lp1 (X) → Lq0 + Lq1 (Y ) is linear. Show that θ → log Bθ is convex. µ) and (Y. q1 ). (If q0 = q1 = ∞. = (1 − θ) . since H is bounded and |B0 B1 | = 2 −1)/m 1−Re(z) Re(z) B0 B1 ≥ min(B0 . p q p0 q0 p1 q1 for some 0 < θ < 1. which proves the lemma. INTERPOLATION 1−θ θ (Exercise. Let Gm = G(z)e(z . ν) be measure spaces. B1 ) > 0. Therefore the Maximum Principle applied to Gm in U says sup |Gm (z)| ≤ sup |Gm | z∈U ∂U ∪{∞} = sup |Gm | ∂U ≤ sup |G| ∂U ≤ 1. Theorem C. If B0 = 0 or B1 = 0. Then Gm is holomorphic on U with |Gm (z)| = |G(z)|e−(y ≤ |G(z)|e 2 +1)/m e(x 2 −1)/m where z = x + iy since x2 ≤ 1 on U . T Lp (X)→Lq (Y ) ≤ T 1−θ Lp0 (X)→Lq0 (Y ) T θ Lp1 (X)→Lq1 (Y ) . q0 .6 (Riesz–Thorin Interpolation). since |H| ≤ B0 on {Re(z) = 0} and |H| ≤ B1 on {Re(z) = 1}. Letting m → ∞ gives |G(z)| ≤ 1. q0 ) and (p1 . m > 0. then T is strong (p. +θ .

2. if θ = 0 then (p. we need the value of T f to be independent of the choice of decomposition f = f0 + f1 . This deﬁnition is independent of the decomposition.1 1 p Figure C. and the two deﬁnitions agree on the intersection Lp0 ∩ Lp1 (X). q1 ). p0 q0 0 1. A subtle aspect of the theorem is that when we assume T maps Lp0 + Lp1 (X) into Lq0 + Lq1 (Y ). as follows. q) = (p1 . For suppose f = f0 + f1 . q) = (p0 . p q 1 1 .169 1 q 1 1 .7. In applications of the theorem. q0 ). Then f0 − f0 = f1 − f1 ∈ Lp0 ∩ Lp1 (X) . In particular. Recall from measure theory that Lp ⊂ Lp0 + Lp1 .1: Parameters in the Riesz–Thorin theorem. Then one deﬁnes T on f = f0 + f1 ∈ Lp0 + Lp1 by T f = T f0 + T f1 . and if θ = 1 then (p. Remark C. usually one has T deﬁned and linear on p0 L (X) and Lp1 (X). 1.1. by splitting f ∈ Lp into large and small parts. p1 q1 1 1 . f1 ∈ Lp1 (X)} consists of all sums of functions in Lp0 and Lp1 . The relationship between the p and q parameters is shown in Figure C. The space Lp0 + Lp1 (X) = {f0 + f1 : f0 ∈ Lp0 (X).

INTERPOLATION and so T f0 − f0 = T f1 − f1 .3) applied to T f on Y . say f = αj 1Fj and g = βj 1Gj . so that p0 = p1 = p. Next suppose p0 = p1 . q) bound. Fix θ ∈ (0. 1).170 APPENDIX C. 3. q0 ) and (p1 . deﬁne p p (1 − z) + z. Let f ∈ Σ(X) and g ∈ Σ(Y ). where on the left side we use T deﬁned on Lp0 (X) and on the right side we use T on Lp1 (X). Riesz–Thorin says that Lp0 ∩ Lp1 ⊂ Lp with f where f 1 p Lp (X) ≤ f 1−θ Lp0 (X) f θ Lp1 (X) (C. For z ∈ C. Then T f Lq (Y ) ≤ T f 1−θ(Y ) T f θ q1 (Y ) L Lq0 by (C. p0 p1 q q Q (z) = (1 − z) + z q0 q1 P (z) = . q1 ) bounds can be applied directly to give the (p. Linearity of T now yields T f0 + T f1 = T f0 + T f1 so that the deﬁnition of T f is independent of the decomposition of f . Then at the end we will prove the bound for f ∈ Lp (X). First suppose p0 = p1 . We will prove an Lp → Lq bound on T f for f ∈ Σ(X). so that p < ∞.3) = = 1−θ p0 + θ . When T = identity. Now the (p0 . p1 Here is a direct proof: p Lp (X) |f |p dµ X = X |f |p(1−θ) |f |pθ dµ |f |p(1−θ)·p0 /p(1−θ) dµ X p(1−θ) Lp0 (X) p(1−θ)/p0 X ≤ = f |f |pθ·p1 /pθ dµ pθ/p1 by H¨lder o f pθ Lp1 (X) Proof of Riesz–Thorin Interpolation. which ﬁxes p and q.

here. o as one sees by substituting (C. Lq (Y ) . (The in Q does not denote a derivative.) Let fz (x) = |f (x)|P (z) ei arg f (x) .171 where 1 q + 1 q = 1. Re(z) = 0 ⇒ Re P (z) = p . p0 ) and (p1 .) The right side belongs to Lq0 ∩ Lq1 (Y ) by the strong (p0 . And H is holomorphic. Similarly.5) ≤ (const. since P (θ) = 1 and Q (θ) = 1.5) into (C.6) and taking the sums outside the integral.4) and (C. y ∈ Y. z ∈ U .6) is well-deﬁned and bounded for z ∈ U . Note fθ = f and gθ = g. by H¨lder. |αj |P (z) ei arg αj (T 1Fj ) (C. Next. Hence the function H(z) = Y (T fz )(y)gz (y) dν(y) (C. Clearly gz = |βj |Q (z) ei arg βj 1Gj . (C. since 1Fj ∈ Lp0 ∩ Lp1 (X). so that |T fz | ≤ |αj |Re P (z) |T 1Fj | |T 1Fj | for z ∈ U .4) Therefore gz (y) is bounded for y ∈ Y. p0 Re Q (z) = q q0 ⇒ |fz |p0 = |f |p0 Re P (z) = |f |p |gz |q0 = |g|q0 Re Q (z) = |g|q ⇒ fz gz Lp0 (X) Lq0 (Y ) = f = g p/p0 Lp (X) q /q0 Lq (Y ) (valid even when p0 = ∞ or q0 = ∞) ⇒ |H(z)| ≤ T fz Lq0 (Y ) gz Lq0 (Y ) by H¨lder o |H(z)| ≤ T Lp0 (X)→Lq0 (Y ) f p/p0 Lp (X) g q /q1 . T fz = by linearity. p1 ) bounds. and it has support (independent of z) with ﬁnite measure. e . gz (y) = |g(y)| Q (z) i arg g(y) x ∈ X.

hn = fn 1X\E . Lq (Y ) Hence by the Hadamard Three Lines Lemma C. and |gn | ≤ |g|.e. So ﬁx f ∈ Lp (X) and let E = {x : |f (x)| > 1}. INTERPOLATION Lp1 (X)→Lq1 (Y ) f p/p1 Lp (X) g q /q1 .7) which is the desired strong (p. 1−θ Lp0 (X)→Lq0 (Y ) θ Lp1 (X)→Lq1 (Y ) f Lp (X) . Theorem 2. Now the dual characterization of the norm on Lq (Y ) implies Tf Lq (Y ) ≤ T 1−θ Lp0 (X)→Lq0 (Y ) T θ Lp1 (X)→Lq1 (Y ) f Lp (X) . and h = f 1X\E .172 Similarly.7). p) bound. Theorem 6. Hence T hn → T h in Lq1 (Y ).e.7) from f ∈ Σ(X) to f ∈ Lp (X). Also hn → h in Lp1 (X) by dominated convergence (or. gn = fn 1E . |hn | ≤ |h|. by the uniform convergence fn → f on X \ E). Suppose p0 < p1 . fn = gn + hn . by swapping p0 and p1 if necessary. Choose a sequence of simple functions fn ∈ Σ(X) with |fn | ≤ |f | and fn → f at every point. such a sequence exists by [Folland. By passing to a subsequence we can further suppose T gn → T g pointwise a. we have T fn → T f pointwise a. and with fn → f uniformly on X\E.14] for the dual characterization of the norm. Re(z) = 1 ⇒ |H(z)| ≤ T APPENDIX C.10]. and so Tf Lq (Y ) ≤ lim inf T fn n Lq (Y ) by Fatou’s lemma T T θ Lp1 (X)→Lq1 (Y ) ≤ T = T 1−θ Lp0 (X)→Lq0 (Y ) lim inf fn n Lp (X) by (C. p) bound on the simple functions. (See [Folland. if p1 = ∞.e. and so T gn → T g in Lq0 (Y ). (C. By passing to a subsequence we can suppose T hn → T h a. g | = |H(θ)| ≤ T 1−θ Lp0 (X)→Lq0 (Y ) T θ Lp1 (X)→Lq1 (Y ) f Lp (X) g Lq (Y ) . Deﬁne g = f 1E . Therefore by linearity of T . the strong (p.) We must extend this bound (C. which uses semi-ﬁniteness of ν when q = ∞. if z = θ then | T f. Then gn → g in Lp0 (X) by dominated convergence. so that f = g + h.5 and a short calculation.

.173 since fn → f in Lp (X) by dominated convergence. p) bound for all f ∈ Lp (X). We have proved the desired strong (p. and so the proof is complete.

INTERPOLATION .174 APPENDIX C.

Grafakos. Jaming. New Jersey. Edwards and G. Gaudry. CRCPress. Prentice Hall. 2001. Micchelli. Second edition. J. American Mathematical Society. a [Jaming] P. I. S.Bibliography [Benedetto] J. Ergebnisse der Mathematik und ihrer Grenzgebiete. Nazarov’s uncertainty principles in higher dimension. 8:335–373. 14. McKean. Appl. Folland. John Wiley and Sons. 1999. Uncertainty principles in Hilbert spaces. [Grafakos] L. 2001. Gr¨chenig. P. o Birkh¨user. 2002.. 2007. Pure and Applied Mathematics (New York). Modern Techniques and their Applications. Approx. [Duoandikoetxea] J. Inc. Benedetto. Graduate Studies in Mathematics. Springer–Verlag. 2001. B. New York. [Dym and McKean] H. 175 . Goh and C. A. Academic Press. [Folland] G. Classical and Modern Fourier Analysis. New York-London. o [Gr¨chenig] K. [Goh and Micchelli] S. Band 90. [Edwards and Gaudry] R. RI. 1972. 29. Fourier Anal. J. A Wiley-Interscience Publication. Harmonic Analysis and Applications. No. Boston. 2003. Dym and H. Littlewood-Paley and Multiplier Theory. Providence. Berlin–New York. Fourier Analysis. E. Duoandikoetxea. Fourier series and integrals. Transl. Upper Saddle River. Probability and Mathematical Statistics. CruzUribe. Foundations of Time–Frequency Analysis. Real Analysis. 1977. Theory 149:30–41. J. D.

1. Princeton Lectures in Analysis. Dover Publications. Princeton University Press. [Zygmund] A. Singular Integrals and Diﬀerentiability Properties of Functions. A. Katznelson. N. Princeton Mathematical Series. . Cambridge Mathematical Library. K¨rner. 1970. An introduction. Princeton University Press. 1990. No. Foreword by R. Princeton. Stein and G. Third edition. Fourier analysis. M. Strichartz. M. Cambridge University Press. S.. 32. New York. 1971. Second corrected edition. Cambridge. Princeton. Stein.. M. Weiss. [Stein and Shakarchi] E. Princeton Mathematical Series. Fourier Analysis. World Scientiﬁc Publishing Company. Shakarchi. Feﬀerman.J. 2003. Cambridge. Rudin. [Strichartz] R. Introduction to Fourier Analysis on Euclidean Spaces. Fourier Analysis on Groups. Cambridge Unio o versity Press. Inc. 2003. I. Trigonometric Series. 1989.. Stein and R. No. [K¨rner] T. Zygmund. Princeton University Press. [Rudin] W. 30. Princeton. W. II. A Guide to Distribution Theory and Fourier Transforms. N.176 BIBLIOGRAPHY [Katznelson] Y. New York. NJ. 2002. Vol. Second edition. [Stein] E. [Stein and Weiss] E. An Introduction to Harmonic Analysis. 1976.J. Wiley-Interscience.

Are you sure?

This action might not be possible to undo. Are you sure you want to continue?

We've moved you to where you read on your other device.

Get the full title to continue

Get the full title to continue listening from where you left off, or restart the preview.

scribd