Professional Documents
Culture Documents
SOPHIA XUE
Contents
1. Introduction to Fourier Series 1
2. Uniqueness of Fourier series 3
3. Convolutions 6
4. Mean-square convergence of Fourier series 11
Acknowledgments 15
References 15
means that the series on the right-hand side is the Fourier series of f .
Definition 1.3. The N th partial sum of the Fourier Series of f , for N a positive
integer, is given by
N
X
SN (f )(x) = fˆ(n)e2⇡inx/L .
n= N
th
The notion of N partial sum of the Fourier Series of f is very important in
the study of Fourier Analysis. Using the partial sums of the Fourier series, we can
view the convergence of Fourier series as the ”limit” of these symmetric sums as N
tends to infinity . Indeed, the basic question can be reformulated as follows:
Following are some simple examples to familiarize ourselves with above defini-
tions.
First, when n 6= 0,
Z ⇡
1
fˆ(n) = ✓e in✓
d✓
2⇡ ⇡
⇡ Z ⇡
1 e in✓ e in✓
= [✓ d✓]
2⇡ in ⇡ ⇡ in
in⇡ in⇡ in⇡
e +e e ein⇡
= +
2in (2⇡in)( in)
cos(n⇡) sin(n⇡)
= +
in i⇡n2
n+1
( 1)
= .
in
When n=0,
Z ⇡
1
fˆ(n) = ✓d✓ = 0.
2⇡ ⇡
Furthermore, by the alternating series test, we can easily see that the Fourier series
of f is convergent.
Corollary 2.3. If f is continuous on the circle and fˆ(n) = 0 for all n 2 Z, then
f ⌘ 0.
Moreover, this corollary is useful in answering Question 2.4 under the condition
that the Fourier series converges absolutely.
Corollary 2.4. Suppose that f is a continuous function on the circle and that
P1
the Fourier series of f is absolutely convergent, n= 1 | fˆ(n) | < 1. Then, the
Fourier series converges uniformly to f , that is,
uniformly in x.
P1 P1
Proof. Let g(x) = n= 1 fˆ(n)einx . Since n= 1 | fˆ(n) | < 1, g(x) is defined
everywhere.
By triangle-inequality, we get that
X 1
X X X
| SN (f )(x) g(x) |=| fˆ(n)einx fˆ(n)einx |=| fˆ(n)einx | | fˆ(n) | .
|n|N n= 1 |n|>N |n|>N
Since the Fourier series of f is bounded, for any ✏ > 0, we can find large N such
P
that |n|>N | fˆ(n) |< ✏. Hence, SN (f )(x) uniformly converges to g(x) as N ! 1.
Note that for large N , the Fourier coefficients of g(x) are
Z 2⇡
1
ĝ(n) = g(x)e inx dx
2⇡ 0
Z 2⇡ X
1
= ( fˆ(n)einx )e inx dx
2⇡ 0
nN
= fˆ(n).
In other words, f and g have identical Fourier coefficients. Now let’s define a new
function h = f g. By distributivity of the integral, it is easy to see that ĥ(n) = 0.
Apply Corollary 3.3 to h, we get that h = f g = 0 or f = g. Combining this
result with the uniform convergence of SN (f )(x) to g(x), we get the desired result,
i.e. SN (f )(x) uniformly converges to f (x) as N ! 1. ⇤
Now that we have proven when the Fourier series of a continuous function is
absolutely convergent, its Fourier series converge uniformly to the said function,
what conditions on f would guarantee the absolute convergence of its Fourier series?
There are also stronger versions of Corollary 2.5 such as the following.
Corollary 2.6. If f is 2⇡-periodic and has an integrable derivative, then its Fourier
series converges absolutely and uniformly to f .
More generally,
Theorem 2.7. The Fourier series of f converges absolutely (and hence uniformly
to f ) if f satisfies a Hölder condition of order ↵, with ↵ > 1/2, that is
Proof. First, we will prove that for any 2⇡-periodic and integrable function f ,
Z ⇡
ˆ 1
f (n) = f (x + ⇡/n)e inx dx,
2⇡ ⇡
and thus
Z ⇡
1
fˆ(n) = [f (x) f (x + ⇡/n)]e inx
dx.
4⇡ ⇡
By definition, we have
Z ⇡
1
fˆ(n) = f (y)e iny
dx.
2⇡ ⇡
6 SOPHIA XUE
Since f is periodic and we are integrating over the period of f , we can substitute
x + ⇡n for y without changing the bounds of integration. Hence,
Z ⇡
1
fˆ(n) = f (x + ⇡/n)e in(x+⇡/n) dx
2⇡ ⇡
Z ⇡
1
= f (x + ⇡/n)e inx e ⇡i dx
2⇡ ⇡
Z ⇡
1
= f (x + ⇡/n)e inx dx.
2⇡ ⇡
It follows
Z ⇡
1
| fˆ(n) | = | [f (x) f (x + ⇡/n)]e inx
dx |
4⇡ ⇡
Z ⇡
1 inx
| f (x) f (x + ⇡/n) || e | dx.
4⇡ ⇡
3. Convolutions
Definition 3.1. Given two 2⇡-periodic integrable functions f and g on R, their
convolution f ⇤ g on [ ⇡, ⇡] is given by
Z ⇡
1
(f ⇤ g)(x) = f (y)g(x y)dy.
2⇡ ⇡
Note that since the product of two integrable functions is again integrable, the
above integral makes sense for all x. Furthermore, since the functions are periodic,
we can change variables to see that
Z ⇡
1
(3.2) (f ⇤ g)(x) = f (x y)g(y)dy.
2⇡ ⇡
For any 2⇡-periodic integrable function f , its convolution with the N th Dirichlet
kernel is
Z ⇡ XN
1
(f ⇤ DN )(x) = f (y)( ein(x y) )dy
2⇡ ⇡ n= N
N
X Z ⇡
1 iny
= ( f (y)e dy)einx
2⇡ ⇡
n= N
N
X
= fˆ(n)einx
n= N
= SN (f )(x).
As we have shown above, Question 2.4 can be again reformulated as following:
Question 3.4. In what sense does (f ⇤ DN )(x) converge to f as N ! 1?
But before we dive in, let’s look at some nice properties of convolution.
Proposition 3.5. Suppose that f , g, and h are 2⇡-periodic integrable functions.
Then:
(i) f ⇤ (g + h) = (f ⇤ g) + (f ⇤ h)
(iii) f ⇤g =g⇤f
(iv) (f ⇤ g) ⇤ h = f ⇤ (g ⇤ h)
(v) f ⇤ g is continuous
(vi) f[
⇤ g(n) = fˆ(n)ĝ(n).
Part (i) and (ii) give linearity. Part (iii) gives commutativity while part (iv)
demonstrates associativity. Part (v) shows that convolution is a smoothing opera-
tion in the sense that even though f and g are merely integrable, f ⇤g is continuous.
Part (vi) converts convolution to multiplication, which is key in the study of Fourier
series.
Now we move on to the proof of Proposition 3.5. Part (i) and (ii) follow directly
from the linearity of integration. Part (iii) follows from Equation (3.2). Part
(iv) can be easily proven through interchanging two integral signs and changing
variables. Though Part (v) and (vi) are easily deduced if f and g are continuous,
to prove these properties under the condition that f and g are merely integrable,
we are going to need the following approximation lemma.
8 SOPHIA XUE
Lemma 3.6. Suppose f is integrable on the circle and bounded by B. Then there
exists a sequence {fk }1
k=1 of continuous functions on the circle so that
sup | fk (x) | B
x2[ pi,⇡]
Proof of Lemma 3.6. When f is real, given ✏ > 0, 9 a partition P of the interval
[ ⇡, ⇡] such that U (f, P ) L(f, P ) < ✏ since f is integrable. Now we define a step
function
g(x) = sup f (y)
xj 1 yxj
To modify the step function g and make it continuous, we will take small > 0
and define g*(x) = g(x) when the distance between x and any partition point in P
is at least . When the distance between x and any partition point in P is more than
, define g*(x) as the linear function that connects g(x ) and g(x + ). It is easy
to see that g* is continuous. Now let g*(x) = 0 when x 2 [ ⇡ , ⇡ + ]andx 2
[⇡ , ⇡ + ], which extends g* to a 2⇡-periodic function. Since g* only di↵ers from
g in N intervals of length s and is also bounded by B, we have
Z ⇡
| g(x) g*(x) | dx 4BN .
⇡
By triangle inequality,
Z ⇡ Z ⇡ Z ⇡
| f (x) g*(x) | dx | g(x) f (x) | dx + | g(x) g*(x) | dx < 2✏.
⇡ ⇡ ⇡
Let 2✏ = k1 and denote g* by fk , the sequence {fk } is what we are looking for.
When f is complex, we can apply the above proof to the real part and the
imaginary part separately to get the same result. ⇤
Applying Lemma 3.6, we can complete the proof of part (v) and (vi), the idea
here is to substitute fk and fˆk for f and fˆ respectively and use the continuity of
fk to complete the proof.
Proof of Part (v) and (vi). Let’s apply Lemma 3.6 to 2⇡-periodic integrable func-
tions f and g. We get sequences {fk } and {gk } of approximating continuous func-
tions. Note that with a bit of algebraic manipulation, we have
f ⇤g fk ⇤ gk = (f fk ) ⇤ g + fk (g gk ).
CONVERGENCE OF FOURIER SERIES 9
= fˆ(n)ĝ(n).
Hence,
Z ⇡
1
| fˆ(n) ˆ
fk (n) | = | (f (x) fk (x))e inx
dx |
2⇡ ⇡
Z ⇡
1
| f (x) fk (x) | dx
2⇡ ⇡
! 0 as k ! 1,
which means fˆk (n) ! fˆ(n) as k ! 1. Again, similarly, gˆk (n) ! ĝ(n) as k ! 1.
By substituting fˆk (n) and gˆk (n) for f (n) and g(n), we can easily get Part (vi). ⇤
Now that we have familiarized ourselves with the basic properties of convolution,
we will introduce the concept of good kernels.
Definition 3.7. A family of good kernels is a sequence of functions {Kn (x)}1
n=1
that satisfies the following properties:
The concept of good kernels is very important in the study of Fourier series
because of the nice property good kernels have, which is stated in the following
theorem.
Theorem 3.8 (Approximation to the Identity). Let {Kn (x)}1 n=1 be a family of
good kernels, and f an integrable function on the circle. Then
lim (f ⇤ Kn )(x) = f (x)
n!1
Example 3.9 (the Dirichlet kernels is not a family of good kernel). Let’s first
remind ourselves that
X sin((N + 1/2)✓)
DN (✓) = N eik✓ = .
sin(✓/2)
k= N
The last part of the equation is the closed form formula for the N th Dirichlet kernel.
x
Since sin(x) 1 for x 2 [ ⇡/2, ⇡/2],
| sin(N + 1/2)✓ |
| DN (✓) | 2 for ✓ 2 [ ⇡, pi].
✓
It follows then,
Z ⇡ Z ⇡
| sin(N + 1/2)✓ |
| DN (✓) | d✓ 4 d✓
⇡ 0 ✓
Z (N +1/2)⇡
| sin(✓) |
=4 d✓
0 ✓
Z N⇡
| sin(✓) |
4
0 ✓
X1 Z (k+1)⇡
N
| sin(✓) |
=4 d✓
k⇡ ✓
k=0
N
X1 Z (k+1)⇡
1
4 | sin(✓) | d✓
(k + 1)⇡ k⇡
k=0
N
X1
8 1
=
⇡ k+1
k=0
8
= log(N + 1)
⇡
8
logN,
⇡
R⇡
which means ⇡ | DN (✓) | d✓ is not bounded and thus the Dirichlet kernel does not
satisfy property (b) of good kernels. This result is disheartening as it suggests that
the point-wise convergence of Fourier series may even fail at points of continuity.
Z 2⇡
1
(en , em ) = ein✓ eim✓
2⇡ 0
Z 2⇡
= f rac12⇡ ei(n m)✓
d✓
0
i(n m)✓ 2⇡
1 e
=
2⇡ i(n m) 0
i(n m)2⇡
1 e e0
=
2⇡ i(n m)
1 1+0 1
=
2⇡ i(n m)
= 0.
Therefore, we have shown that the family {en }n2Z is orthonormal.
Besides the orthonormal property of the family {en }n2Z , another important ob-
servation is that the Fourier coefficients of the function f can be represented as the
inner products of f with elements of the above family:
Z 2⇡
1
(f, en ) = f (✓)e in✓ d✓ = an .
2⇡ 0
Moreover, it is also important to note that
X
SN (f ) = an en .
|n|N
With these in mind, we are ready to prove a simple but useful lemma.
CONVERGENCE OF FOURIER SERIES 13
While this lemma may not seem too interesting by itself, it has two very useful
results.
For the first result, let’s recall the Pythagorean theorem, which states that if X
and Y are orthogonal, then
||X + Y ||2 = ||X||2 + ||Y ||2 .
Let bn = an , we get
X X X X
||f an en + an en ||2 = ||f an en ||2 + || an en ||2 ,
|n|N |n|N |n|N |n|N
The second result is best known as the best approximation lemma, which states
that
Lemma 4.3 (Best approximation lemma). If f is integrable on the circle with
Fourier coefficients an , then for any complex numbers cn
X
||f SN (f )|| ||f cn en ||.
|n|N
P P
By Lemma 4.2, we know that (f |n|N an en )?(f |n|N bn en ). Hence, we
can apply the Pythagorean theorem and get the following result:
X X
||f (f cn en )||2 = ||f SN (f )||2 + ||(f bn en )||2 .
|n|N |n|N
When cn = an , bn = 0
X
||f (f cn en )|| = ||f SN (f )||.
|n|N
When cn 6= an , bn 6= 0
X Z 2⇡
1
|| bn en ||2 = | bn en |2 d✓ > 0.
2⇡ 0
|n|N
Hence, X
||f SN (f )|| < ||f cn en ||.
|n|N
⇤
Theorem 4.4 (Mean Square Convergence). Suppose f is integrable on the circle.
Then Z 2⇡
1
| f (✓) SN (f )(✓) |2 d✓ ! 0 as N ! 1.
2⇡ 0
Proof. Apply Lemma 3.6 and choose a continuous function g on the circle such that
sup | g(✓) | sup | f (✓) |= B
✓2[0,2⇡] ✓2[0,2⇡]
and Z 2⇡
| f (✓) g(✓) | d✓ < ✏2 .
0
Since supx2[ ⇡,⇡] | f (x) |= B,
Z 2⇡
1
||f g||2 = | f (✓) g(✓) |2 d✓
2⇡ 0
Z 2⇡
1
= | f (✓) g(✓) || f (✓) g(✓) | d✓
2⇡ 0
Z 2⇡
B
| f (✓) g(✓) | d✓
⇡ 0
C✏2
where C = B ⇡ is some constant.
Moreover, we can approximate g by a trigonometric polynomial P of degree M
such that
| g(✓) P (✓) |< ✏, for all ✓.
After taking squares and integrating, we get
||g P || < ✏.
1/2
It follows directly that ||f P || < C ✏. We can then get the desired result by
applying the best approximation lemma.
⇤
CONVERGENCE OF FOURIER SERIES 15
References
[1] Elias M. Stein and Rami Shakarchi. Fourier analysis-an introduction, Princeton University
Press. 2003.
[2] Maria Christina Pereyra and Lesley A. Ward. Harmonic Analysis: from Fourier to Haar