Lp norms of trigonometric polynomials

Jordan Bell
[email protected]
Department of Mathematics, University of Toronto
April 3, 2014
A trigonometric polynomial of degree n is an expression of the form
n
X

ck eikt ,

ck C.

k=n

Using the identity eit = cos t + i sin t, we can write a trigonometric polynomial
of degree n in the form
a0 +

n
X
k=1

ak cos kt +

n
X

bk sin kt,

ak , bk C.

k=1

For 1 p < and for a 2-periodic function f , we define the Lp norm of

f by

1/p
Z 2
1
p
kf kp =
.
|f (t)| dt
2 0
For a continuous 2-periodic function f , we define the L norm of f by
kf k = max |f (t)|.
0t2

If f is a continuous 2-periodic function, then there is a sequence of trigonometric polynomials fn such that kf fn k 0 as n [31, p. 54, Corollary 5.4].
If 1 p < and f is a continuous 2-periodic function, then

kf kp =

1
2

Z
0

1/p 
1/p
Z 2
1
|f (t)|p dt

kf kp dt
= kf k .
2 0

Jensens inequality [16, p. 44, Theorem 2.2] (cf. [30, p. 113, Problem 7.5]) tells
us that if : [0, ) R is convex, then for any function h : [0, 2] [0, ) we
have


Z 2
Z 2
1
1

h(t)dt
(h(t))dt.
2 0
2 0
1

If 1 p < q < , then : [0, ) R defined by (x) = xq/p is convex. Hence,

if 1 p < q < then for any 2-periodic function f ,
kf kp

=
=

=
=

((kf kpp ))1/q

 
1/q
Z 2
1

|f (t)|p dt
2 0

1/q
Z 2
1
(|f (t)|p )dt
2 0
1/q

Z 2
1
|f (t)|q dt
2 0
kf kq .

The Dirichlet kernel Dn is defined by

Dn (t) =

n
X

eikt = 1 + 2

k=n

n
X

cos kt.

k=1

One can show [14, p. 71, Exercise 1.1] that

kDn k1 =

4
log n + O(1).
2

(On the other hand, it can quickly be seen

that kDn k = 2n + 1, and it follows
from Parsevals identity that kDn k2 = 2n + 1.)
P
olya and Szeg
o [27, Part VI] present various problems about trigonometric polynomials together with solutions to them. A result on L norms of
trigonometric
Pn sin kt polynomials that Polya and Szego present is for the sum An (t) =
k=1 k . The local maxima and local minima of An can be explicitly determined [27, p. 74, no. 23], and it can be shown that [27, p. 74, no. 25]
Z
sin t
kAn k
dt.
t
0

Lp norms

If 1 p < q < , then [14, p. 123, Exercise 1.8] (cf. [7, p. 102, Theorem 2.6])
there is some C(p, q) such that for any trigonometric polynomial f of degree n,
we have
1
1
kf kq C(p, q)n p q kf kp .
This inequality is sharp [33, p. 230]: for 1 p < q < there is some C 0 (p, q)
Pn1
such that if Fn (t) = n1 k=0 Dk (t) (Fn is called the Fejer kernel) then
1

kFn kq > C 0 (p, q)n p q kFn kp .

Pn
Let Xn = {a0 + k=1 ak cos kt + bk sin kt : ak , bk R}, the real vector space
of real valued trigonometric polynomials of degree n, have norm
kf kXn = max{|a0 |, |a1 |, . . . , |an |, |b1 |, . . . , |bn |}.
Let Yn,p be the same vector space with the Lp norm. Ash and Ganzburg [1]
give upper and lower bounds on the operator norm of the map i : Xn Yn,p
defined by i(f ) = f .
Bernsteins inequality [14, p. 50, Exercise 7.16] states that for 1 p , if
f is a trigonometric polynomial of degree n, then
kf 0 kp nkf kp .
In the other direction, if f C 1 then
Z 2
Z t
Z 2
1
1
1
0
f (s)ds +
sf (s)ds +
(s 2)f 0 (s)ds
2 0
2 0
2 t
Z 2
Z 2
Z 2
1
1
0
=
f (s)ds +
sf (s)ds
f 0 (s)ds
2 0
2 0
t
Z 2
Z 2
2
2
1
1
1


sf (s)
f (s)ds +
f (s)ds f (s)
=
2 0
2
2 0
0
t
=f (t).
Hence
|f (t)|

1
2

|f (s)|ds +
0

Z 2
1
|f (s)|ds +
2 0
kf k1 + 2kf 0 k1 ,

1
2
Z t

s|f 0 (s)|ds +

|f 0 (s)|ds +

1
2

(2 s)|f 0 (s)|ds

|f 0 (s)|ds

so
kf k kf k1 + 2kf 0 k1 .
This is an instance of the Sobolev inequality [26].
It turns out that for a trigonometric polynomial the mass cannot be too
concentrated. More precisely, the number of nonzero terms of a trigonometric
dt
polynomial restricts how concentrated its mass can be. Let d = 2
. Thus
([0, 2]) = 1. A result of Turan [20, p. 89, Lemma 1] states that if 1 , . . . , N
PN
Z and T (t) = n=1 bn ein t , bn C, then for any closed arc I [0, 2],

N 1
2e
kT k
max |T (t)|.
tI
(I)
Nazarov [11, p. 452] shows that there is some constant A such that if E is a
closed subset of [0, 2] (not necessarily an arc), then
N

A

kT k1
max |f (T )|.
tE
(E)
3

Nazarov [23] proves that there exists some constant C such that if 0 q 2
and (E) 31 , then

1/q
Z
(E)
1
kT kq eC(N 1)(1 2 )
|T (t)|q dt
.
2 E
These results of Turan and Nazarov are examples of the uncertainty principle
[9], which is the general principle that a constrain on the support of the Fourier
transform of a function constrains the support of the function itself.
In [10], Hardy and Littlewood present inequalities for norms of 2-periodic
functions in terms of certain series formed from their Fourier coefficients. Let
ck C, k Z, be such that ck 0 as k , and define c0 , c1 , c1 , c2 , c2 , . . .
to be the absolute values of the ck ordered in decreasing magnitude. For real
r > 1, define
!1/r

r
r2
Sr (c) =
ck (|k| + 1)
.
k=

For instance, if ck = 1 for N k N and ck = 0 for |k| > N , then


1/r
PN +1
Sr (c) = 1 + 2 k=2 k r2
. Hardy and Littlewood state the result [10,
p. 164, Theorem 2] that if 1 < p 2 then there is some constant
A(p) such
P
that for any sequence c, with ck 0 as k , if f (t) = k= ck eikt and
kf kp < then
Sp (c) A(p)kf kp .
A proof of this is given in Zygmund [35, vol. II, p. 128, chap. XII, Theorem 6.3].
Asking if this inequality holds for p = 1 suggests the following question that
Hardy and Littlewood pose at the end of their paper [10, p. 168]: Is there a
constant A such that for all distinct positive integers mk , k = 1, . . . , N , we have
k

N
X

cos mk tk1 > A log N ?

k=1

McGehee, Pigno and Smith [18] prove that there is some K such that for all N ,
if n1 , . . . , nN are distinct integers and c1 , . . . , cN C satisfy |ck | 1, then
k

N
X

ck eink t k1 > K log N.

k=1

Thus
k

N
X
k=1

X
1
1
eimk t + eimk t k1 K log(2N ).
cos mk tk1 = k
2
2
k=1

PN

For k 2, define TN (t) = n=1 ein t . Since kTN k = N , for each p 1 we

have kTN kp N . Huas lemma [22, p. 116, Theorem 4.6] states that if  > 0,
then


k
kTN k2k = O N 1 2k + .
4

Huas lemma is used in additive number theory. The number of sets of integer
solutions of the equation
a r x r br

f (x1 , . . . , xn ) = N,
is equal to (cf. [12, p. 151])
X

an xn bn

a1 x1 b1

e2i(f (x1 ,...,xn )N )t dt.

Borwein and Lockhart [4]: what is the expected Lp norm of a trigonometric

polynomial of order n? Kahane [13, Chapter 6] also presents material on random
trigonometric polynomials.
Nursultanov and Tikhonov [25]: the sup on a subset of T of a trigonometric
polynomial f of degree n being lower bounded in terms of kf k , n, and the
measure of the subset.

`p norms

For a 2-periodic function f , we define f : Z C by

Z 2
1
f(k) =
eikt f (t)dt.
2 0
For 1 p < , we define the `p norm of f by
!1/p

kfkp =

|f(k)|

k=

and we define the ` norm of f by

kfk = max |f(k)|.
kZ

Parsevals identity [31, p. 80, Theorem 1.3] states that kf k2 = kfk2 .

If 1 p < , then

1/p
kfk + kfkp +
= kfkp .
If 1 p < q < , then, since for each k,

1=

X
k=

|f(k)|
kfkq

!q !1/q

|f(k)|
kfkq

k=

Hence for 1 p < p ,

kfkq kfkp .
5

1,
|f(k)|
kfkq

!p !1/q

p/q

kfkp
.
p/q
kfkq

For 1 p < , if f is a trigonometric polynomial of degree n then

!1/p
!1/p
n
n
X
X
p
p

kf kp =

= (2n + 1)1/p kfk .

|f (k)|
kf k
k=n

k=n

For 1 p < q < , we have [30, p. 123, Problem 8.3] (this is Jensens
inequality for sums)
n
X
k=n

1
|f(k)|p
2n + 1

!1/p

n
X
k=n

1
|f(k)|q
2n + 1

!1/q
,

i.e.
(2n + 1)1/p kfkp (2n + 1)1/q kfkq .
Hence for 1 < p < q < ,
1

kfkp (2n + 1) p q kfkq .

For any t,




 X
X
X

ikt
ikt 

|f (k)e | =
|f(k)| = kfk1 .
f (k)e 
|f (t)| = 


k=

k=

k=

Hence
kf k kfk1 .
For any k Z,


Z 2
Z 2
 1

1
|f(k)| = 
eikt f (t)dt
|f (t)|dt = kf k1 .
2 0
2 0
Hence
kfk kf k1 .
The Hausdorff-Young inequality [32, p. 57, Corollary 2.4] states that for 1
p 2 and p1 + 1q = 1, if f Lp then
kfkq kf kp .
The dual Hausdorff-Young inequality [32, p. 58, Corollary 2.5] states that for
1 p 2 and p1 + 1q = 1, if f Lq then
kf kq kfkq .
A survey on the Hausdorff-Young inequality is given in [6])
PN +1
For M + 1 k M + N , let ak C and let S(t) = k=M +1 ak eikt . Let
t1 , . . . , tR R, and let be such that if r 6= s then
ktr ts k ,
6

where ktk = mink |t k| is the distance from t to a nearest integer. The large
sieve [19] is an inequality of the form
R
X

|S(2tr )|2 (N, )

r=1

M
+N
X

|ak |2 .

k=M +1

A result of Selberg [19, p. 559, Theorem 3] shows that the large sieve is valid
for = N 1 + 1 .
Kristiansen [15]
Boas [2]
For F : Z/n C, its Fourier transform F : Z/n C (called the discrete
Fourier transform) is defined by
n1
1X
F (j)e2ijk/n ,
F (k) =
n j=0

0 k n 1,

and one can prove [31, p. 223, Theorem 1.2] that

F (j) =

n1
X

F (k)e2ikj/N ,

0 j n 1.

k=0

One can also prove Parsevals identity for the Fourier transform on Z/n [31,
p. 223, Theorem 1.2]. It states
n1
X

|F (k)|2 =

k=0

Let P (t) =

Pn1
k=0

n1
1X
|F (j)|2 .
n j=0

ak eikt . Define F : Z/n C by

F (j) =

n1
X

ak e2ikj/n ,

0 j n 1.

k=0

(That is, F (k) = ak .) We then have

n1
X

|ak |2 =

k=0

Thus

n1
n1
1 X   2j 2
1X
|F (j)|2 =
P
.
n j=0
n j=0
n

1/2




1
P 2j 2 .
kP k2 =
n j=0
n

n1
X

The Marcinkiewicz-Zygmund inequalities [35, vol. II, p. 28, chap. X, Theorem 7.5] state that there is a constant A such that for 1 p , if f is a
trigonometric polynomial of degree n then
!1/p
2n
1 X   2k p
f
A(2)1/p kf kp ,
2n + 1
2n + 1
k=0

and for each 1 < p < there exists some Ap such that if f is a trigonometric
polynomial of degree n then
2n

kf kp Ap

1 X   2k p
f
2n + 1
2n + 1

!1/p
.

k=0

M
ate and Nevai [17, p. 148, Theorem 6] prove that for p > 0, if Sn is a
trigonometric polynomial of degree n then

kSn k

(1 + np)e
2

1/p
kSn kp .

M
ate and Nevai [17] prove a version of Bernsteins inequality for 0 < p < 1, and
their result can be sharpened to the following [34]: For 0 < p < 1, if Tn is a
trigonometric polynomial of order n then
kTn0 kp nkTn kp .
Let supp f = {k Z : f(k) 6= 0}. A subset of Z is called a Sidon set
[28, p. 121, 5.7.2] if there exists a constant B such that for every trigonometric
polynomial f with supp f we have
kfk1 Bkf k .
Let B() be the least such B. A sequence of positive integers k is said to
be lacunary if there is a constant such that k+1 > k for all k. If k is a
lacunary sequence, then {k } is a Sidon set [21, p. 154, Corollary 6.17]. If Z
is a Sidon set, then [28, p. 128, Theorem 5.7.7] (cf. [21, p. 157, Corollary 6.19])
for any 2 < p < , for every trigonometric polynomial f with supp f we
have

kf kp B() pkf k2 ,
and
kf k2 2B()kf k1 .
Let 0 < p < . A subset E of Z is called a (p)-set if for every 0 <
r < p there is some A(E, p) such that for all trigonometric polynomials f with
supp f E we have
kf kp A(E, p)kf k2 .
(p) sets were introduced by Rudin, and he discusses them in his autobiography
[29, Chapter 28]. A modern survey of (p)-sets is given by Bourgain [5].
Bochkarev [3] proves various lower bounds on the L1 norms of certain trigonometric polynomials. Let ck C, k 1. If there are constants A and B such
that
(log k)s
(log k)s
A
|ck | B
,
k 1,
k
k

then [3, p. 58, Theorem 19]

k

n
X
k=1

ck eik

(
1
(log n)s 2 , s > 12 ,
t
k1 
log log n,
s = 12 .

Pn

ikt
If P (t) =
with ak {1, 1}, then by the Cauchy-Schwarz
k=0 ak e
inequality and Parsevals identity we have
Z 2

1
1 |P (t)|dt k1k2 kP k2 = 1 kP k2 = n + 1.
kP k1 =
2 0

Newman [24] shows that in fact we can do better than what we get using the
Cauchy-Schwarz inequality and Parsevals identity:

kP k1 < n + 0.97.


Pl1
A Fekete polynomial is a polynomial of the form k=1 kl z k , l prime, where




P
l1 k
k
eikt . Erdelyi [8] proves upper
l is the Legendre symbol. Let Pl (t) =
1/q k=1 l
 R
1
, q > 0, where I is an arc in [0, 2].
|Pl (t)|q dt
and lower bounds on |I|
I

References
[1] J. Marshall Ash and Michael Ganzburg, An extremal problem for trigonometric polynomials, Proc. Amer. Math. Soc. 127 (1999), no. 1, 211216.
[2] R. P. Boas, Jr., Inequalities for the coefficients of trigonometric polynomials. II, Nederl. Akad. Wetensch., Proc. 50 (1947), 759762 = Indagationes
Math. 9, 369372 (1947).
[3] S. V. Bochkarev, Multiplicative inequalities for the L1 -norm: applications
to analysis and number theory, Proc. Steklov Inst. Math. 255 (2006), 4964.
[4] Peter Borwein and Richard Lockhart, The expected Lp norm of random
polynomials, Proc. Amer. Math. Soc. 129 (2001), no. 5, 14631472.
[5] Jean Bourgain, p -sets in analysis: results, problems and related aspects,
Handbook of the Geometry of Banach Spaces, volume I (W. B. Johnson and
J. Lindenstrauss, eds.), North-Holland, Amsterdam, 2001, pp. 195232.
[6] P. L. Butzer, The Hausdorff-Young theorems of Fourier analysis and their
impact, J. Fourier Anal. Appl. 1 (1994), no. 2, 113130.
[7] Ronald A. DeVore and George G. Lorentz, Constructive approximation,
Die Grundlehren der mathematischen Wissenschaften, vol. 303, Springer,
1993.
[8] Tam
as Erdelyi, Upper bounds for the Lq norm of Fekete polynomials on
subarcs, Acta Arith. 153 (2012), no. 1, 8191.
9

[9] Gerald B. Folland and Alladi Sitaram, The uncertainty principle: a mathematical survey, J. Fourier Anal. Appl. 3 (1997), no. 3, 207238.
[10] G. H. Hardy and J. E. Littlewood, A new proof of a theorem on rearrangements, J. London Math. Soc. 23 (1948), no. 3, 163168.
[11] Victor Havin and Burglind Joricke, The uncertainty principle in harmonic
analysis, Ergebnisse der Mathematik und ihrer Grenzgebiete, 3. Folge,
vol. 28, Springer, 1994.
[12] Loo Keng Hua, Introduction to number theory, Springer, 1982, Translated
from the Chinese by Peter Shiu.
[13] Jean-Pierre Kahane, Some random series of functions, second ed., Cambridge University Press, 1985.
[14] Yitzhak Katznelson, An introduction to harmonic analysis, third ed., Cambridge Mathematical Library, Cambridge University Press, 2004.
[15] G. K. Kristiansen, Proof of an inequality for trigonometric polynomials,
Proc. Amer. Math. Soc. 44 (1974), 4957.
[16] Elliott H. Lieb and Michael Loss, Analysis, second ed., Graduate Studies
in Mathematics, vol. 14, American Mathematical Society, Providence, RI,
2001.
[17] Attila M
ate and Paul G. Nevai, Bernsteins inequality in Lp for 0 < p <
1 and (C, 1) bounds for orthogonal polynomials, Ann. of Math. (2) 111
(1980), no. 1, 145154.
[18] O. Carruth McGehee, Louis Pigno, and Brent Smith, Hardys inequality
and the L1 norm of exponential sums, Ann. of Math. (2) 113 (1981), no. 3,
613618.
[19] Hugh L. Montgomery, The analytic principle of the large sieve, Bull. Amer.
Math. Soc 84 (1978), no. 4, 547567.
[20]

, Ten lectures on the interface between analytic number theory and

harmonic analysis, CBMS Regional Conference Series in Mathematics,
no. 84, American Mathematical Society, Providence, RI, 1994.

[21] Camil Muscalu and Wilhelm Schlag, Classical and multilinear harmonic
analysis, volume 1, Cambridge Studies in Advanced Mathematics, no. 137,
Cambridge University Press, 2013.
[22] Melvyn B. Nathanson, Additive number theory: the classical bases, Graduate Texts in Mathematics, vol. 164, Springer, 1996.

10

[23] Fedor L. Nazarov, Complete version of Turans lemma for trigonometric

polynomials on the unit circumference, Complex analysis, operators, and
related topics: the S. A. Vinogradov memorial volume (Victor P. Havin and
Nikolai K. Nikolski, eds.), Operator Theory: Advances and Applications,
vol. 113, Birkh
auser, Basel, 2000, pp. 239246.
[24] D. J. Newman, Norms of polynomials, Amer. Math. Monthly 67 (1960),
778779.
[25] E. Nursultanov and S. Tikhonov, A Sharp Remez Inequality for Trigonometric Polynomials, Constr. Approx. 38 (2013), no. 1, 101132.
ad Benyi, The Sobolev inequality on the torus revis[26] Tadahiro Oh and Arp
ited, Publ. Math. Debrecen.
[27] George P
olya and G
abor Szego, Problems and theorems in analysis, volume II, Die Grundlehren der mathematischen Wissenschaften, vol. 216,
Springer, 1976, Translated from the German by C. E. Billigheimer.
[28] Walter Rudin, Fourier analysis on groups, Interscience Tracts in Pure and
Applied Mathematics, no. 12, Interscience Publishers, 1962.
[29]

, The way I remember it, History of Mathematics, vol. 12, American

Mathematical Society, Providence, RI, 1997.

[30] J. Michael Steele, The Cauchy-Schwarz master class: an introduction to the

art of mathematical inequalities, MAA Problem Books Series, Cambridge
University Press, 2004.
[31] Elias M. Stein and Rami Shakarchi, Fourier analysis: an introduction,
Princeton Lectures in Analysis, vol. I, Princeton University Press, 2003.
[32]

, Functional analysis: introduction to further topics in analysis,

Princeton Lectures in Analysis, vol. IV, Princeton University Press, 2011.

[33] A. F. Timan, Theory of approximation of functions of a real variable, International Series of Monographs on Pure and Applied Mathematics, vol. 34,
Pergamon Press, Oxford, 1963, Translated from the Russian by J. Berry.
[34] M. von Golitschek and G. G. Lorentz, Bernstein inequalities in Lp , 0
p +, Rocky Mountain J. Math. 19 (1989), no. 1, 145156.
[35] Antoni Zygmund, Trigonometric series, volumes I & II combined, third
ed., Cambridge Mathematical Library, Cambridge University Press, 2002.

11