9.

1 Complex form of Fourier Series

It is often convenient to work with complex form of Fourier series. In deed, the complex
form of Fourier series has applications in the field of signal processing which is of great
interest to many electrical engineers.
Given the Fourier series of a function f (x) as

∞
1 X
f ∼ a0 + [an cos(nx) + bn sin(nx)] , −π < x < π (9.1)
2
n=1

with Z π
1
an = f (x) cos(nx) dx, n = 0, 1, 2 . . .
π −π
and Z π
1
bn = f (x) sin(nx) dx, n = 1, 2 . . .
π −π
We know from Euler’s formula
einx + e−inx einx − e−inx
cos(nx) = sin(nx) =
2 2i

Substituting these values of cos(nx) and sin(nx) into the equation (9.1) we obtain
∞  inx −inx inx −inx

1 X e +e e −e
f ∼ a0 + an + bn
2 2 2i
n=1
∞ h
1 1 1
X i
= a0 + (an − ibn )e inx
+ (an + ibn )e−inx
2 2 2
n=1

Let us define new coefficients as

1 1
cn = (an − ibn ), kn = (an + ibn ) (9.2)
2 2
Note that c0 = a0 /2 because b0 = 0. Then the Fourier series becomes
∞
X
cn einx + kn e−inx
 
f ∼ c0 + (9.3)
n=1

where the coefficients are given as

Z π Z π
1 1 1
cn = (an − ibn ) = f (x) [cos(nx) − i sin(nx)] dx = f (x)e−inx dx
2 2π −π 2π −π
 Lecture Notes on Fourier Transform

Z π Z π
1 1 1
kn = (an + ibn ) = f (x) [cos(nx) + i sin(nx)] dx = f (x)einx dx
2 2π −π 2π −π

From the above calculation we get kn = c−n . Substituting the value of kn into the Fourier
series (9.3) we have
∞
X
f∼ cn einx (9.4)
n=−∞

where
Z π
1
cn = f (x)e−inx dx, n = 0, ±1, ±2, . . . (9.5)
2π −π

The series on the right side of equation (9.4) is called complex form of the Fourier series.
For a function of period 2L defined in [−L, L], the complex form of the Fourier series can
analogously be derived to have
∞ Z L
X inπx 1 −inπx
f∼ cn e L , cn = f (x)e L dx, n = 0, ±1, ±2, . . .
2L −L
n=−∞

9.2 Example Problems

9.2.1 Problem 1

Find the complex Fourier series of

f (x) = ex if − π < x < π and f (x + 2π) = f (x)

Solution: We calculate the coefficients cn as

Z π Z π
1 x −inx 1
cn = e e dx = e(1−in)x dx
2π −π 2π −π
1 e(1−in)x π
1 1  π −inπ
− e−π einπ

=  = e e

2π 1 − in −π 2π 1 − in
Substituting e±inπ = cos nπ ± i sin nπ = (−1)n we get
1 1 + in 1 1 + in
cn = (−1)n sinh π = (−1)n sinh π
π (1 − in)(1 + in) π (1 + n2 )

2
 Lecture Notes on Fourier Transform

Then, the Fourier is given as

∞
sinh π X 1 + in inx
f∼ (−1)n e .
π (1 + n2 )
n=−∞

9.2.2 Problem 2

Determine the complex Fourier series representation of

f (x) = x if − l < x < l and f (x + 2l) = f (x)

.
Solution: The complex Fourier series representation of a function f (x) is given as
∞
inπx
X
f∼ cn e l

n=−∞

where Z l Z l
1 −inπx 1 −inπx
cn = f (x)e l dx = xe l dx
2l −l 2l −l
For n 6= 0, integrating by parts we get
"  l Z l #
1 −inπx −l l −inπx
cn = xe  + e l dx ,
l

2l inπ −l inπ −l

Further application of integration by parts simplifies to

 2 2

1 l −inπ l inπ l2 −inπx l

cn = − e − e − e l  ,
2l inπ inπ (inπ)2 −l
| {z }
=0

Finally, it simplifies to
(−1)n il
cn = , n = ±1, ±2, . . .
nπ
Now c0 can be calculated as Z l
1
c0 = x dx = 0
2l −l
Therefore, the Fourier series is given as
∞
il X (−1)n inπx
f∼ e l
π n
n=−∞
n6=0

3
 Lecture Notes on Fourier Transform

9.2.3 Problem 3

Show that Parseval’s identity for the complex form of Fourier series takes the form
Z π ∞
1 X
{f (x)}2 dx = |cn |2
2π −π n=−∞

Solution: For the real form of Fourier series the Parseval’s identity is given as
∞
a20 X 2
1 π
Z
2 2
+ an + b n = {f (x)} dx (9.6)
2 π −π
n=1

We know that
a0 1 1
c0 = , cn = (an − ibn ), c−n = (an + ibn )
2 2 2
We can deduce that
1 1
|cn |2 = (a2n + b2n ), |c−n |2 = (a2n + b2n ) (9.7)
4 4
Diving the equation (9.6) by 2 and then splitting the second term as
∞ ∞ Z π
a20 1 X 2
1X 1
2 2 2
{f (x)}2 dx


+ an + b n + an + b n =
4 4 4 2π −π
n=1 n=1

Using the relations (9.7) we obtain

∞ ∞ Z π
X X 1
c20 + 2
|cn | + |c−n |2 = {f (x)}2 dx
2π −π
n=1 n=1

This can be rewritten as

∞ Z π
X 1
2
|cn | = {f (x)}2 dx
2π −π
n=−∞

9.2.4 Problem 4

Given the Fourier series

∞
sinh π X
x 1 + in inx
e ∼ (−1)n e .
π (1 + n2 )
n=−∞

4
 Lecture Notes on Fourier Transform

deduce the value of

∞
X 1
n2 +1
n=−∞

Solution: From the given series we clearly have

ne
π − e−π 1 + in
cn = (−1) , n = 0, ±1, ±2, . . .
2π (1 + n2 )

These coefficients can be simplified

2
2
eπ − e−π (1 + n2 ) eπ − e−π 1
|cn |2 = =
4π 2 (1 + n2 )2 4π 2 (1 + n2 )

A simple calculation gives

Z π π
e2π − e−2π
Z
1 21
{f (x)} dx = e2x dx =
2π −π 2π −π 4π

Thus, by Parseval’s identity we have

2 ∞
e2π − e−2π eπ − e−π X 1
=
4π 4π 2 (1 + n2 )
n=−∞

Therefore, we obtain
∞
π eπ + e−π


X 1
= = π cot hπ.
(1 + n2 ) (eπ − e−π )
n=−∞