Part 2: Fourier Analysis

By H. Ali (HOA)

BASIC CONCEPT OF FOURIER ANALYSIS

Most of the phenomena studied in the domain of Engineering and Science are periodic in nature. For instance,
current and voltage in an alternating current circuit. These periodic functions could be analyzed into their
constituent components (fundamentals and harmonics) by a process called Fourier analysis.

FOURIER SERIES
A Fourier series is an expansion of a periodic function f(x) in terms of an infinite sum of sines and cosines. Fourier
Series makes use of the orthogonality relationships of the sine and cosine functions.
The formula of Fourier series is given as:
∞

𝑓(𝑥) = 𝑎0 + ∑(𝑎𝑛 cos 𝑛𝑥 + 𝑏𝑛 sin nx)

𝑛=1
Again, if 𝑓(𝑥) is periodic from (−𝐿) to (+𝐿).
∞
𝑛𝜋𝑥 𝑛𝜋𝑥
𝑓(𝑥) = 𝑎0 + ∑ (𝑎𝑛 cos + 𝑏𝑛 sin )
𝐿 𝐿
𝑛=1

Now, from the formula of Fourier series, it is clear that, in order to find the series, we need to first know the values
of 𝑎0 , 𝑎𝑛 and 𝑏𝑛 . So, let’s find their values.

For an:

From Fourier Series,

∞

𝑓(𝑥) = 𝑎0 + ∑(𝑎𝑛 cos 𝑛𝑥 + 𝑏𝑛 sin nx)

𝑛=1

Taking cos 𝑛𝑥 on both sides,

∞

cos 𝑛𝑥 𝑓(𝑥) = cos 𝑛𝑥 𝑎0 + cos 𝑛𝑥 ∑(𝑎𝑛 cos nx + 𝑏𝑛 sin nx)

𝑛=1
Integrating both sides,
𝜋 𝜋 𝜋 ∞

∫ cos 𝑛𝑥 𝑓(𝑥) 𝑑𝑥 = ∫ 𝑎0 cos 𝑛𝑥 𝑑𝑥 + ∫ cos 𝑛𝑥 ∑(𝑎𝑛 cos 𝑛𝑥 + 𝑏𝑛 sin nx) 𝑑𝑥

−𝜋 −𝜋 −𝜋 𝑛=1
𝜋
𝑜𝑟, ∫ cos 𝑛𝑥 𝑓(𝑥) 𝑑𝑥 = 0 + 𝑎𝑛 𝜋
−𝜋
𝟏 𝝅
∴ 𝒂𝒏 = ∫ 𝒇(𝒙) 𝐜𝐨𝐬 𝒏𝒙 𝒅𝒙
𝝅 −𝝅

For bn:

From Fourier Series,

 ∞

𝑓(𝑥) = 𝑎0 + ∑(𝑎𝑛 cos 𝑛𝑥 + 𝑏𝑛 sin nx)

𝑛=1
Taking sin 𝑛𝑥 on both sides,
∞

sin 𝑛𝑥 𝑓(𝑥) = sin 𝑛𝑥 𝑎0 + sin 𝑛𝑥 ∑(𝑎𝑛 cos 𝑛𝑥 + 𝑏𝑛 sin nx)

𝑛=1
Integrating both sides,
𝜋 𝜋 𝜋 ∞

∫ sin 𝑛𝑥 𝑓(𝑥) 𝑑𝑥 = ∫ 𝑎0 sin 𝑛𝑥 𝑑𝑥 + ∫ sin 𝑛𝑥 ∑(𝑎𝑛 cos 𝑛𝑥 + 𝑏𝑛 sin nx) 𝑑𝑥

−𝜋 −𝜋 −𝜋 𝑛=1
𝜋
𝑜𝑟, ∫ sin 𝑛𝑥 𝑓(𝑥) 𝑑𝑥 = 0 + 𝑏𝑛 𝜋
−𝜋
𝟏 𝝅
∴ 𝒃𝒏 = ∫ 𝒇(𝒙) 𝐬𝐢𝐧 𝒏𝒙 𝒅𝒙
𝝅 −𝝅

For a0:

From Fourier Series,

∞
𝑛𝜋𝑥 𝑛𝜋𝑥
𝑓(𝑥) = 𝑎0 + ∑ (𝑎𝑛 cos + 𝑏𝑛 sin )
𝐿 𝐿
𝑛=1
Taking cos 0𝑥 on both sides,
∞
𝑛𝜋𝑥 𝑛𝜋𝑥
cos 0𝑥 𝑓(𝑥) = cos 0𝑥 𝑎0 + cos 0𝑥 ∑ (𝑎𝑛 cos + 𝑏𝑛 sin )
𝐿 𝐿
𝑛=1
Integrating both sides,
𝜋 𝜋 𝜋 ∞
𝑛𝜋𝑥 𝑛𝜋𝑥
∫ cos 0𝑥 𝑓(𝑥) 𝑑𝑥 = ∫ 𝑎0 cos 0𝑥 𝑑𝑥 + ∫ cos 0𝑥 ∑ (𝑎𝑛 cos + 𝑏𝑛 sin ) 𝑑𝑥
−𝜋 −𝜋 −𝜋 𝐿 𝐿
𝑛=1
𝜋
𝑜𝑟, ∫ 𝑓(𝑥) 𝑑𝑥 = 𝑎0 [𝑥]𝜋−𝜋 + 0
−𝜋
𝟏 𝝅
∴ 𝒂𝟎 = ∫ 𝒇(𝒙) 𝒅𝒙
𝟐𝝅 −𝝅
So, in summary, the necessary formulae for any type of Fourier series dealing with REAL numbers are:

Fourier Series:
∞

𝑓(𝑥) = 𝑎0 + ∑(𝑎𝑛 cos 𝑛𝑥 + 𝑏𝑛 sin 𝑛𝑥)

𝑛=1

Different parameters to find Fourier Series:

1 𝜋
𝑎0 = ∫ 𝑓(𝑥) 𝑑𝑥
2𝜋 −𝜋
1 𝜋
𝑎𝑛 = ∫ 𝑓(𝑥) cos 𝑛𝑥 𝑑𝑥
𝜋 −𝜋
1 𝜋
𝑏𝑛 = ∫ 𝑓(𝑥) sin 𝑛𝑥 𝑑𝑥
𝜋 −𝜋

−1 ; −𝜋 < 𝑡 < 0
Example 22: If 𝑓(𝑡) = { & 𝑓(𝑡) is periodic over –π to π, find the Fourier series of 𝑓(𝑡).
+1 ; 0 < 𝑡 < 𝜋

Soln.: From Euler’s Formula,

1 𝜋
𝑎0 = ∫ 𝑓(𝑡) 𝑑𝑡
2𝜋 −𝜋
0 𝜋
1
= [∫ (−1) 𝑑𝑡 + ∫ (1) 𝑑𝑡]
2𝜋 −𝜋 0

1
= [[𝑡]0−𝜋 + [𝑡]𝜋0 ]
2𝜋
1
= (−𝜋 + 𝜋)
2𝜋
=0

Again, from Euler’s Formula,

1 𝜋
𝑎𝑛 = ∫ 𝑓(𝑡) cos 𝑛𝑡 𝑑𝑡
𝜋 −𝜋
1 0 𝜋
= [∫ (−1) cos 𝑛𝑡 𝑑𝑡 + ∫ (1) cos 𝑛𝑡 𝑑𝑡]
𝜋 −𝜋 0

1 sin 𝑛𝑡 0 sin 𝑛𝑡 𝜋
= [− [ ] +[ ] ]
𝜋 𝑛 −𝜋 𝑛 0
1
= ×0
𝜋
=0
Lastly, from Euler’s Formula,
1 𝜋
𝑏𝑛 = ∫ 𝑓(𝑡) sin 𝑛𝑡 𝑑𝑡
𝜋 −𝜋
1 0 𝜋
= [∫ (−1) sin 𝑛𝑡 𝑑𝑡 + ∫ (1) sin 𝑛𝑡 𝑑𝑡]
𝜋 −𝜋 0

1 cos 𝑛𝑡 0 cos 𝑛𝑡 𝜋
= [[ ] +[ ] ]
𝜋 𝑛 −𝜋 𝑛 0
1 1 cos 𝑛𝜋 cos 𝑛𝜋 1
= [ − − + ]
𝜋 𝑛 𝑛 𝑛 𝑛
2
= (1 − cos 𝑛𝜋)
𝑛𝜋
2
= [1 − (−1)𝑛 ]
𝑛𝜋

Now, from Fourier Series,

∞

𝑓(𝑥) = 𝑎0 + ∑(𝑎𝑛 cos 𝑛𝑡 + 𝑏𝑛 sin 𝑛𝑡)

𝑛=1
∞
2
= 0 + ∑ (0 + [1 − (−1)𝑛 ] sin 𝑛𝑡)
𝑛𝜋
𝑛=1
∞
2
=∑ [1 − (−1)𝑛 ] sin 𝑛𝑡
𝑛𝜋
𝑛=1
2 2 2
= ∙ 2 sin 𝑡 + 0 + ∙ 2 sin 3𝑡 + 0 + ∙ 2 sin 5𝑡 + ⋯ … … … … …
𝜋 3𝜋 5𝜋
4 sin 3𝑡 sin 5𝑡
= (sin 𝑡 + + + ⋯…………)
𝜋 3 5
(Ans.)

Note: If a Fourier series contains only sin θ or cos θ, then the series is called “Half Range Fourier Series.”
We can understand it clearly seeing the chart below:
Example 23: If 𝑓(𝑡) = 𝑡 − 𝑡 2 ; & 𝑓(𝑡) is periodic over –π to π, find the Fourier series of 𝑓(𝑡).

Soln.: From Euler’s Formula,

𝜋
1
𝑎0 = ∫ 𝑓(𝑡) 𝑑𝑡
2𝜋
−𝜋
𝜋
1
= ∫(𝑡 − 𝑡 2 ) 𝑑𝑡
2𝜋
−𝜋
𝜋 𝜋
1 𝑡2 𝑡3
= [[ ] − [ ] ]
2𝜋 2 −𝜋 3 −𝜋

1 𝜋2 𝜋2 𝜋3 𝜋3
= [ − − − ]
2𝜋 2 2 3 3
𝜋2
=−
3

Again, from Euler’s Formula,

1 𝜋
𝑎𝑛 = ∫ 𝑓(𝑡) cos 𝑛𝑡 𝑑𝑡
𝜋 −𝜋
𝜋
1
= ∫(𝑡 − 𝑡 2 ) cos 𝑛𝑡 𝑑𝑡 … … … … … … … (𝑖)
𝜋
−𝜋

Now, [Performing the integration part first]

∫(𝑡 − 𝑡 2 ) cos 𝑛𝑡 𝑑𝑡

= (𝑡 − 𝑡 2 ) ∫ cos 𝑛𝑡 𝑑𝑡 − ∫ [(1 − 2𝑡) ∫ cos 𝑛𝑡 𝑑𝑡] 𝑑𝑡

sin 𝑛𝑡 sin 𝑛𝑡
= (𝑡 − 𝑡 2 ) − ∫ [(1 − 2𝑡) ] 𝑑𝑡
𝑛 𝑛
sin 𝑛𝑡 1
= (𝑡 − 𝑡 2 ) − [(1 − 2𝑡) ∫ sin 𝑛𝑡 𝑑𝑡 − ∫ {(−2) ∫ sin 𝑛𝑡 𝑑𝑡} 𝑑𝑡]
𝑛 𝑛
sin 𝑛𝑡 1 cos 𝑛𝑡 cos 𝑛𝑡
= (𝑡 − 𝑡 2 ) − [(1 − 2𝑡) (− ) − ∫ {(−2) (− )} 𝑑𝑡]
𝑛 𝑛 𝑛 𝑛
sin 𝑛𝑡 1 cos 𝑛𝑡 2𝑡 cos 𝑛𝑡 2 sin 𝑛𝑡
= (𝑡 − 𝑡 2 ) − [− + − ]
𝑛 𝑛 𝑛 𝑛 𝑛2
𝑡 sin 𝑛𝑡 𝑡 2 sin 𝑛𝑡 cos 𝑛𝑡 2𝑡 cos 𝑛𝑡 2 sin 𝑛𝑡
= − + − +
𝑛 𝑛 𝑛2 𝑛2 𝑛3
From equation (i),
𝜋
1 𝑡 sin 𝑛𝑡 𝑡 2 sin 𝑛𝑡 cos 𝑛𝑡 2𝑡 cos 𝑛𝑡 2 sin 𝑛𝑡
𝑎𝑛 = [ − + − + ]
𝜋 𝑛 𝑛 𝑛2 𝑛2 𝑛3 −𝜋
1 (−1)𝑛 2𝜋(−1)𝑛 (−1)𝑛 2(−𝜋)(−1)𝑛
= [ 2 − − + ]
𝜋 𝑛 𝑛2 𝑛2 𝑛2
1 4𝜋(−1)𝑛
= [− ]
𝜋 𝑛2
4(−1)𝑛
=−
𝑛2

Lastly, from Euler’s Formula,

𝜋
1
𝑏𝑛 = ∫ 𝑓(𝑡) sin 𝑛𝑡 𝑑𝑡
𝜋
−𝜋
𝜋
1
= ∫(𝑡 − 𝑡 2 ) sin 𝑛𝑡 𝑑𝑡 … … … … … … … (𝑖𝑖)
𝜋
−𝜋

Now, [Performing the integration part first]

∫(𝑡 − 𝑡 2 ) sin 𝑛𝑡 𝑑𝑡

= (𝑡 − 𝑡 2 ) ∫ sin 𝑛𝑡 𝑑𝑡 − ∫ [(1 − 2𝑡) ∫ sin 𝑛𝑡 𝑑𝑡] 𝑑𝑡

cos 𝑛𝑡 cos 𝑛𝑡
= (𝑡 − 𝑡 2 ) (− ) − ∫ [(1 − 2𝑡) (− )] 𝑑𝑡
𝑛 𝑛
cos 𝑛𝑡 1
= (𝑡 − 𝑡 2 ) (− ) + [(1 − 2𝑡) ∫ cos 𝑛𝑡 𝑑𝑡 − ∫ {(−2) ∫ cos 𝑛𝑡 𝑑𝑡} 𝑑𝑡]
𝑛 𝑛
cos 𝑛𝑡 1 sin 𝑛𝑡 sin 𝑛𝑡
= (𝑡 − 𝑡 2 ) (− ) + [(1 − 2𝑡) ( ) − ∫ {(−2) ( )} 𝑑𝑡]
𝑛 𝑛 𝑛 𝑛
cos 𝑛𝑡 1 sin 𝑛𝑡 2𝑡 sin 𝑛𝑡 2 cos 𝑛𝑡
= (𝑡 − 𝑡 2 ) (− )+ [ − + ]
𝑛 𝑛 𝑛 𝑛 𝑛2
𝑡 cos 𝑛𝑡 𝑡 2 cos 𝑛𝑡 sin 𝑛𝑡 2𝑡 sin 𝑛𝑡 2 cos 𝑛𝑡
=− + + − +
𝑛 𝑛 𝑛2 𝑛2 𝑛3
From equation (ii),
𝜋
1 𝑡 cos 𝑛𝑡 𝑡 2 cos 𝑛𝑡 sin 𝑛𝑡 2𝑡 sin 𝑛𝑡 2 cos 𝑛𝑡
𝑏𝑛 = [− + + − + ]
𝜋 𝑛 𝑛 𝑛2 𝑛2 𝑛3 −𝜋

1 (−𝜋)(−1)𝑛 𝜋 2 (−1)𝑛 2(−1)𝑛 𝜋(−1)𝑛 𝜋 2 (−1)𝑛 2(−1)𝑛

= [ + + − − − ]
𝜋 𝑛 𝑛 𝑛3 𝑛 𝑛 𝑛3
1 2𝜋(−1)𝑛
= [− ]
𝜋 𝑛
2(−1)𝑛
=−
𝑛

We know, the Fourier series is:

∞

𝑓(𝑡) = 𝑎0 + ∑(𝑎𝑛 cos 𝑛𝑡 + 𝑏𝑛 sin 𝑛𝑡)

𝑛=1
∞
𝜋2 4(−1)𝑛 2(−1)𝑛
=− + ∑ [− cos 𝑛𝑡 − sin 𝑛𝑡]
3 𝑛2 𝑛
𝑛=1
(Ans.)

Practice Problems
0 ; −𝜋 ≤ 𝑥 < 0 1
1. If 𝑓(𝑥) = { , find the Fourier series of 𝑓(𝑥) and hence evaluate ∑∞
𝑛=1 .
𝑥 ; 0≤𝑥<𝜋 𝑛2

−1 ; −𝜋 ≤ 𝑥 < 0 𝜋 1 1
2. If 𝑓(𝑥) = { ; find the Fourier series of 𝑓(𝑥) and hence deduce =1−3+5−
+1 ; 0 ≤ 𝑥 ≤ 𝜋 4
1 1
+ − ⋯……………
7 9

3. Expand 𝑓(𝑥) = 𝑥, 0 < 𝑥 < 2 in a half range sine series.

𝜋2 cos 2𝑛𝑥
4. If 0 < 𝑥 < 𝜋, show that 𝑥(𝜋 − 𝑥) = − ∑∞
𝑛=1 ( )
6 𝑛2

5. Expand 𝑓(𝑥) = 𝑒 −𝑥 , 0 < 𝑥 < 2𝜋 in a Fourier series.

2 − 𝑥, 0 < 𝑥 < 4
6. Expand 𝑓(𝑥) = { in a Fourier series of period 8.
𝑥 − 6, 4 < 𝑥 < 8

7. If 𝑓(𝑥) = 𝑥 2 & 𝑓(𝑥) is periodic over 2π, find the Fourier series of 𝑓(𝑥) and hence evaluate
1
∑∞𝑛=1 2.
(2𝑛−1)
0 ; −𝜋 ≤ 𝑥 < 0
8. If 𝑓(𝑥) = { & 𝑓(𝑥) is periodic over 2π, find the Fourier series of 𝑓(𝑥).
sin 𝑥 ; 0 ≤ 𝑥 ≤ 𝜋

−𝑥 ; −𝜋 ≤ 𝑥 < 0
9. If 𝑓(𝑥) = { ; find the Fourier series of 𝑓(𝑥).
+𝑥 ; 0 ≤ 𝑥 ≤ 𝜋
10. Sketch 3 cycles of the function represented by:
1
0 ; −1 ≤ 𝑡 ≤ − 2
1 1
𝑓(𝑡) = cos 3𝜋𝑡 ; − 2 ≤ 𝑡 ≤ 2 and 𝑓(𝑡) = 𝑓(𝑡 + 2)
1
0 ; 2≤𝑡≤1
{
Find the Fourier Series of the above function (up to 5 terms at least).

11. Find the constant term 𝑎0 in the Fourier series corresponding to 𝑓(𝑥) = 𝑥 − 𝑥 3 in (−𝜋, 𝜋). And
what kind of function is it? Draw a graph.
+𝜋 +𝜋
12. Using ∫−𝜋 sin 𝑛𝑥 𝑑𝑥 = 0 and ∫−𝜋 cos 𝑛𝑥 𝑑𝑥 = 0, where 𝑚, 𝑛 𝜖 ℤ. Prove the following identities:
2𝜋 𝑖𝑓 𝑛 = 𝑚 = 0
𝜋
a) ∫−𝜋cos 𝑛𝑥 cos 𝑚𝑥 𝑑𝑥 = { 𝜋 𝑖𝑓 𝑛 = 𝑚 ≠ 0
0 𝑖𝑓 𝑛 ≠ 𝑚
𝜋 𝑖𝑓 𝑛 = 𝑚 = 0
𝜋 𝜋
b) ∫0 cos 𝑛𝑥 cos 𝑚𝑥 𝑑𝑥 = {2 𝑖𝑓 𝑛 = 𝑚 ≠ 0
0 𝑖𝑓 𝑛 ≠ 𝑚
𝜋 𝜋 𝑖𝑓 𝑛 = 𝑚
c) ∫−𝜋 sin 𝑛𝑥 sin 𝑚𝑥 𝑑𝑥={
0 𝑖𝑓 𝑛 ≠ 𝑚
𝜋
𝜋 𝑖𝑓 𝑛 = 𝑚
d) ∫0 sin 𝑛𝑥 sin 𝑚𝑥 𝑑𝑥 = {2
0 𝑖𝑓 𝑛 ≠ 𝑚
𝜋
e) ∫−𝜋 sin 𝑛𝑥 cos 𝑚𝑥 𝑑𝑥 = 0

13. Draw sketches and determine the Fourier Series for the following functions:
𝑥
a) 𝑠(𝑥) = 𝜋 , 𝑓𝑜𝑟 − 𝜋 < 𝑥 < 𝜋
b) 𝑠(𝑥) = 3|sin 𝑥|, 𝑓𝑜𝑟 0 ≤ 𝑥 ≤ 2𝜋
2 sin 𝑥 𝑓𝑜𝑟 0 ≤ 𝑥 < 𝜋
c) 𝑠(𝑥) = {
0 𝑓𝑜𝑟 𝜋 ≤ 𝑥 ≤ 2𝜋
1 𝑓𝑜𝑟 0 ≤ 𝑥 ≤ 𝜋
d) 𝑠(𝑥) = {
0 𝑓𝑜𝑟 𝜋 ≤ 𝑥 ≤ 2𝜋
𝐴𝑥
e) 𝑠(𝑥) = 𝐴 − 𝑓𝑜𝑟 0 ≤ 𝑥 < 𝑃
𝑃
COMPLEX FORM OF FOURIER SERIES
From MAT215 we know that,
𝑒 𝑖𝜃 = cos 𝜃 + 𝑖 sin 𝜃 … … … … (𝑖)
𝑒 −𝑖𝜃 = cos 𝜃 − 𝑖 sin 𝜃 … … … … (𝑖𝑖)
Adding equation (i) and (ii),
𝑒 𝑖𝜃 + 𝑒 −𝑖𝜃 = cos 𝜃 + 𝑖 sin 𝜃 + cos 𝜃 − 𝑖 sin 𝜃
𝑒 𝑖𝜃 + 𝑒 −𝑖𝜃 = 2 cos 𝜃
𝑒 𝑖𝜃 + 𝑒 −𝑖𝜃
∴ cos 𝜃 =
2
Subtracting equation (i) and (ii),
𝑒 𝑖𝜃 − 𝑒 −𝑖𝜃 = cos 𝜃 + 𝑖 sin 𝜃 − cos 𝜃 + 𝑖 sin 𝜃
𝑒 𝑖𝜃 + 𝑒 −𝑖𝜃 = 2𝑖 sin 𝜃
𝑒 𝑖𝜃 + 𝑒 −𝑖𝜃
∴ sin 𝜃 =
2𝑖
−𝑖(𝑒 𝑖𝜃 + 𝑒 −𝑖𝜃 )
=
2
𝑖𝑒 −𝑖𝜃 − 𝑖𝑒 𝑖𝜃
∴ sin 𝜃 =
2
Now, from the Fourier Series,
∞

𝑓(𝑡) = 𝑎0 + ∑(𝑎𝑛 cos 𝑛𝑡 + 𝑏𝑛 sin 𝑛𝑡)

𝑛=1
∞
𝑒 𝑖𝑛𝑡 + 𝑒 −𝑖𝑛𝑡 𝑖𝑒 −𝑖𝑛𝑡 − 𝑖𝑒 𝑖𝑛𝑡
= 𝑎0 + ∑ (𝑎𝑛 + 𝑏𝑛 )
2 2
𝑛=1
∞ ∞
𝑎𝑛 𝑖𝑏𝑛 𝑖𝑛𝑡 𝑎𝑛 𝑖𝑏𝑛 −𝑖𝑛𝑡
= 𝑎0 + ∑ ( − )𝑒 + ∑( + )𝑒
2 2 2 2
𝑛=1 𝑛=1
∞ −1
𝑎𝑛 𝑖𝑏𝑛 𝑖𝑛𝑡 𝑎𝑛 𝑖𝑏𝑛 𝑖𝑛𝑡
= 𝑎0 + ∑ ( − )𝑒 + ∑ ( + )𝑒
2 2 2 2
𝑛=1 𝑛=−∞
∞ −1
𝑖𝑛𝑡
𝑎𝑛 𝑖𝑏𝑛 𝑖𝑛𝑡 𝑎𝑛 𝑖𝑏𝑛 𝑖𝑛𝑡
= 𝑎0 𝑒 |𝑛=0 + ∑ ( − )𝑒 + ∑ ( + )𝑒
2 2 2 2
𝑛=1 𝑛=−∞
∞

= ∑ 𝑐𝑛 𝑒 𝑖𝑛𝑡
𝑛=−∞
∞

𝑻𝒉𝒆𝒓𝒆𝒇𝒐𝒓𝒆, 𝒕𝒉𝒆 𝒄𝒐𝒎𝒑𝒍𝒆𝒙 𝒇𝒐𝒓𝒎 𝒐𝒇 𝑭𝒐𝒖𝒓𝒊𝒆𝒓 𝑺𝒆𝒓𝒊𝒆𝒔 𝒊𝒔: 𝒇(𝒕) = ∑ 𝒄𝒏 𝒆𝒊𝒏𝒕

𝒏=−∞
Now, in order to determine the series, we need to find the value of the coefficient 𝑐𝑛 .

We got from the complex form of Fourier Series,

∞

𝑓(𝑡) = ∑ 𝑐𝑛 𝑒 𝑖𝑛𝑡
𝑛=−∞

Taking 𝑒 −𝑖𝑛𝑡 on both sides of the equation,

∞
−𝑖𝑛𝑡 −𝑖𝑛𝑡
𝑒 ∙ 𝑓(𝑡) = 𝑒 ∙ ∑ 𝑐𝑛 𝑒 𝑖𝑛𝑡
𝑛=−∞

Integrating both sides with respect to t,

𝜋 𝜋 ∞
−𝑖𝑛𝑡 −𝑖𝑛𝑡
∫𝑒 ∙ 𝑓(𝑡) 𝑑𝑡 = ∫ 𝑒 ∙ ∑ 𝑐𝑛 𝑒 𝑖𝑛𝑡
−𝜋 −𝜋 𝑛=−∞
𝜋

𝑜𝑟, ∫ 𝑓(𝑡) ∙ 𝑒 −𝑖𝑛𝑡 𝑑𝑡 = 𝑐𝑛 ∙ 2𝜋

−𝜋
𝝅
𝟏
∴ 𝒄𝒏 = ∫ 𝒇(𝒕) ∙ 𝒆−𝒊𝒏𝒕 𝒅𝒕
𝟐𝝅
−𝝅

So, in summary, the necessary formulae for Complex Form of Fourier series are:

Fourier Series:
∞

𝑓(𝑡) = ∑ 𝑐𝑛 𝑒 𝑖𝑛𝑡
𝑛=−∞

Different parameters to find Fourier Series:

𝜋
1
𝑐𝑛 = ∫ 𝑓(𝑡) ∙ 𝑒 −𝑖𝑛𝑡 𝑑𝑡
2𝜋
−𝜋

The complex form of Fourier series is algebraically simpler and more symmetric. This is why, in Engineering
Mathematics, the complex form of Fourier Series is most commonly used.

We will understand it even more clearly after completing the Parseval’s Theorem.
 −1 ; −𝜋 < 𝑡 < 0
Example 24: If (𝑡) = { & 𝑓(𝑡) is periodic over 2π, find the Fourier series of 𝑓(𝑡) using the
+1 ; 0 < 𝑡 < 𝜋
complex form.

Soln.: We know, the Complex Form of Fourier Series is given as:

∞

𝑓(𝑡) = ∑ 𝑐𝑛 𝑒 𝑖𝑛𝑡
𝑛=−∞
From Euler’s Formula,
𝜋
1
𝑐𝑛 = ∫ 𝑓(𝑡) ∙ 𝑒 −𝑖𝑛𝑡 𝑑𝑡
2𝜋
−𝜋
0 𝜋
1
= [ ∫(−1)𝑒 −𝑖𝑛𝑡 𝑑𝑡 + ∫ 1 ∙ 𝑒 −𝑖𝑛𝑡 𝑑𝑡]
2𝜋
−𝜋 0
−𝑖𝑛𝑡 0 𝜋
1 𝑒 𝑒 −𝑖𝑛𝑡
= (− [ ] +[ ] )
2𝜋 −𝑖𝑛 −𝜋 −𝑖𝑛 0
1 1 𝑒 𝑖𝑛𝜋 𝑒 −𝑖𝑛𝜋 1
= {( − )−( − )}
2𝜋 𝑖𝑛 𝑖𝑛 𝑖𝑛 𝑖𝑛
1
= (1 − 𝑒 𝑖𝑛𝜋 − 𝑒 −𝑖𝑛𝜋 + 1)
2𝑖𝑛𝜋
1
= {2 − (𝑒 𝑖𝑛𝜋 + 𝑒 −𝑖𝑛𝜋 )}
2𝑖𝑛𝜋
1
= (2 − 2 cos 𝑛𝜋)
2𝑖𝑛𝜋
1
= (1 − cos 𝑛𝜋)
𝑖𝑛𝜋
𝑖
=− {1 − (−1)𝑛 }
𝑛𝜋

Now, substituting the value of 𝑐𝑛 in the Complex Form of Fourier Series,

∞

𝑓(𝑡) = ∑ 𝑐𝑛 𝑒 𝑖𝑛𝑡
𝑛=−∞
∞
𝑖
= ∑ − {1 − (−1)𝑛 }𝑒 𝑖𝑛𝑡
𝑛𝜋
𝑛=−∞
∞
𝑖
= ∑ {(−1)𝑛 − 1}𝑒 𝑖𝑛𝑡
𝑛𝜋
𝑛=−∞
(Ans.)

Practice Problem

1. If 𝑓(𝑡) = 𝑡 − 𝑡 2 ; & 𝑓(𝑡) is periodic over 2π, find the Fourier series of 𝑓(𝑡) using the complex form.
PARSEVAL’S IDENTITY
We know, the Fourier series dealing with real numbers is given as:
∞

𝑓(𝑡) = 𝑎0 + ∑(𝑎𝑛 cos 𝑛𝑡 + 𝑏𝑛 sin 𝑛𝑡)

𝑛=1

The Parseval’s theorem deals with the square of the Fourier series, i.e., {𝑓(𝑡)}2. It is like the Pythagoras
Theorem for Fourier series.
Before we know about the Parseval’s identity, it is important to recall some of the basic integral concepts.
𝜋 0 𝑤ℎ𝑒𝑛, 𝑚 ≠ 𝑛
1. ∫−𝜋 sin 𝑛𝑥 ∙ sin 𝑚𝑥 𝑑𝑥 = {
𝜋 𝑤ℎ𝑒𝑛, 𝑚 = 𝑛
𝜋 0 𝑤ℎ𝑒𝑛, 𝑚 ≠ 𝑛
2. ∫−𝜋cos 𝑛𝑥 ∙ cos 𝑚𝑥 𝑑𝑥 = {
𝜋 𝑤ℎ𝑒𝑛, 𝑚 = 𝑛
𝜋
3. ∫−𝜋 sin 𝑛𝑥 ∙ cos 𝑚𝑥 𝑑𝑥 = 0

The Parseval’s identity states that,

𝜋 ∞
1
∫{𝑓(𝑡)}2 𝑑𝑡 = 𝑎0 2 + ∑ 𝑎𝑛 2 + 𝑏𝑛 2
𝜋
−𝜋 𝑛=1

Let’s prove the statement:

We know,
∞

𝑓(𝑡) = 𝑎0 + ∑(𝑎𝑛 cos 𝑛𝑡 + 𝑏𝑛 sin 𝑛𝑡)

𝑛=1

Squaring both sides,

∞ 2

{𝑓(𝑡)}2 = {𝑎0 + ∑(𝑎𝑛 cos 𝑛𝑡 + 𝑏𝑛 sin 𝑛𝑡)}

𝑛=1

Integrating both sides with respect to t,

𝜋 𝜋 ∞ 2

∫ {𝑓(𝑡)}2 𝑑𝑡 = ∫ {𝑎0 + ∑(𝑎𝑛 cos 𝑛𝑡 + 𝑏𝑛 sin 𝑛𝑡)} 𝑑𝑡

−𝜋 −𝜋 𝑛=1
𝜋 ∞ ∞ 2

= ∫ {𝑎0 2 + 2𝑎0 ∑(𝑎𝑛 cos 𝑛𝑡 + 𝑏𝑛 sin 𝑛𝑡) + (∑(𝑎𝑛 cos 𝑛𝑡 + 𝑏𝑛 sin 𝑛𝑡)) } 𝑑𝑡
−𝜋 𝑛=1 𝑛=1
𝜋 𝜋 ∞ 𝜋 ∞ 2

= ∫ 𝑎0 2 𝑑𝑡 + ∫ 2𝑎0 ∑(𝑎𝑛 cos 𝑛𝑡 + 𝑏𝑛 sin 𝑛𝑡) 𝑑𝑡 + ∫ (∑(𝑎𝑛 cos 𝑛𝑡 + 𝑏𝑛 sin 𝑛𝑡)) 𝑑𝑡

−𝜋 −𝜋 𝑛=1 −𝜋 𝑛=1
𝜋 ∞ 𝜋 ∞

= 𝑎0 2 [𝑡]𝜋−𝜋 + 2𝑎0 ∫ ∑(𝑎𝑛 cos 𝑛𝑡 + 𝑏𝑛 sin 𝑛𝑡) 𝑑𝑡 + ∫ (∑(𝑎𝑛 cos 𝑛𝑡 + 𝑏𝑛 sin 𝑛𝑡))
−𝜋 𝑛=1 −𝜋 𝑛=1
∞

∙ ( ∑ (𝑎𝑚 cos 𝑚𝑡 + 𝑏𝑚 sin 𝑚𝑡)) 𝑑𝑡

𝑚=1
 𝜋 ∞ ∞

= 2𝜋𝑎0 2 + 2𝑎0 ∙ 0 + ∫ (∑(𝑎𝑛 cos 𝑛𝑡 + 𝑏𝑛 sin 𝑛𝑡)) ∙ ( ∑ (𝑎𝑚 cos 𝑚𝑡 + 𝑏𝑚 sin 𝑚𝑡)) 𝑑𝑡
−𝜋 𝑛=1 𝑚=1

𝜋 ∞ ∞

= 2𝜋𝑎0 2 + ∫ {∑ ∑ (𝑎𝑛 cos 𝑛𝑡 ∙ 𝑎𝑚 cos 𝑚𝑡 + 𝑎𝑛 cos 𝑛𝑡 ∙ 𝑏𝑚 sin 𝑚𝑡 + 𝑏𝑛 sin 𝑛𝑡 ∙ 𝑎𝑚 cos 𝑚𝑡 + 𝑏𝑛 sin 𝑛𝑡

−𝜋 𝑛=1 𝑚=1

∙ 𝑏𝑚 sin 𝑚𝑡)} 𝑑𝑡

∞ ∞

= 2𝜋𝑎0 + ∑ 𝑎𝑛 𝜋 + ∑ 𝑏𝑛 2 𝜋
2 2

𝑛=1 𝑛=1
∞

= 2𝜋𝑎0 2 + ∑(𝑎𝑛 2 𝜋 + 𝑏𝑛 2 𝜋)
𝑛=1
𝜋 ∞

𝑜𝑟, ∫ {𝑓(𝑡)}2 𝑑𝑡 = 2𝜋𝑎0 + ∑(𝑎𝑛 2 𝜋 + 𝑏𝑛 2 𝜋)

2

−𝜋 𝑛=1

𝝅 ∞
𝟏
∴ ∫ {𝒇(𝒕)}𝟐 𝒅𝒕 = 𝟐𝒂𝟎 𝟐 + ∑(𝒂𝒏 𝟐 + 𝒃𝒏 𝟐 )
𝝅
−𝝅 𝒏=𝟏

FOURIER TRANSFORMATION

Fourier Transformation is a mathematical model which helps to transform the signals between two different
domains, such as transforming signal from frequency domain to time domain or vice versa. The generalization of
the complex Fourier series is known as the Fourier transformation. The term “Fourier transformation” can be used
in the mathematical function, and it is also used in the representation of the frequency domain. The Fourier
transformation helps to extend the Fourier series to the non-periodic functions, which helps us to view any
functions in terms of the sum of simple sinusoids. Fourier transformation has many applications in Engineering
and Physics, such as signal processing, RADAR, and so on.

Fourier Transformation can be of two types.

i. Forward Fourier Transformation, 𝐹. 𝑇. {𝑓(𝑡)}
ii. Inverse Fourier Transformation, 𝐼. 𝐹. 𝑇. {𝐹(𝜔)}

The formulas for both types of Fourier Transformations are given as:
∞

𝑭. 𝑻. {𝒇(𝒕)} = 𝑭(𝝎) = ∫ 𝒇(𝒕) ∙ 𝒆𝒊𝝎𝒕 𝒅𝒕

−∞
∞
𝟏
𝑰. 𝑭. 𝑻. {𝑭(𝝎)} = 𝒇(𝒕) = ∫ 𝑭(𝝎) ∙ 𝒆−𝒊𝝎𝒕 𝒅𝝎
𝟐𝝅
−∞
 1 𝑓𝑜𝑟 |𝑡| < 1 ∞ sin 𝑡
Example 25: If 𝑓(𝑡) = { ; find the Fourier Transformation of 𝑓(𝑡), hence evaluate ∫0 𝑡 𝑑𝑡.
0 𝑓𝑜𝑟 |𝑡| > 1

Soln.: We know,
∞

𝐹. 𝑇. {𝑓(𝑡)} = 𝐹(𝜔) = ∫ 𝑓(𝑡) ∙ 𝑒 𝑖𝜔𝑡 𝑑𝑡

−∞
−1 1 ∞

= ∫ 𝑓(𝑡)𝑒 𝑖𝜔𝑡 𝑑𝑡 + ∫ 𝑓(𝑡)𝑒 𝑖𝜔𝑡 𝑑𝑡 + ∫ 𝑓(𝑡)𝑒 𝑖𝜔𝑡 𝑑𝑡

−∞ −1 1
1

= 0 + ∫ 𝑓(𝑡)𝑒 𝑖𝜔𝑡 𝑑𝑡 + 0
−1
1

= ∫ 1 ∙ 𝑒 𝑖𝜔𝑡 𝑑𝑡
−1
1
𝑒 𝑖𝜔𝑡
=[ ]
𝑖𝜔 −1

𝑒 𝑖𝜔𝑡 𝑒 −𝑖𝜔𝑡
= −
𝑖𝜔 𝑖𝜔
1 𝑒 𝑖𝜔 − 𝑒 −𝑖𝜔
= ( )
𝜔 𝑖
1
= × 2 sin 𝜔
𝜔
2 sin 𝜔
=
𝜔
2 sin 𝜔
∴ 𝐹(𝜔) =
𝜔
Now,
∞
1
𝐼. 𝐹. 𝑇. {𝐹(𝜔)} = 𝑓(𝑡) = ∫ 𝐹(𝜔) ∙ 𝑒 −𝑖𝜔𝑡 𝑑𝜔
2𝜋
−∞
∞
1 2 sin 𝜔 −𝑖𝜔𝑡
𝑜𝑟, 𝑓(𝑡) = ∫ ∙𝑒 𝑑𝜔
2𝜋 𝜔
−∞
∞
1 sin 𝜔 −𝑖𝜔𝑡
= ∫ ∙𝑒 𝑑𝜔
𝜋 𝜔
−∞
1 ∞ sin 𝜔
𝑇ℎ𝑒𝑟𝑒𝑓𝑜𝑟𝑒, 𝑓(0) = 1 = 𝜋 ∫−∞ 𝑑𝜔 [𝐴𝑠 𝑔𝑖𝑣𝑒𝑛 𝑐𝑜𝑛𝑑𝑖𝑡𝑖𝑜𝑛 𝑠𝑎𝑦𝑠, 𝐴𝑡 𝑡 = 0, 𝑓(𝑡) = 1]]
𝜔
 ∞
sin 𝜔
𝑜𝑟, ∫ 𝑑𝜔 = 𝜋
𝜔
−∞
∞
sin 𝜔
𝑜𝑟, 2 ∫ 𝑑𝜔 = 𝜋
𝜔
0
∞
sin 𝜔 𝜋
𝑜𝑟, ∫ 𝑑𝜔 =
𝜔 2
0
∞
sin 𝑡 𝜋
𝑇ℎ𝑒𝑟𝑒𝑓𝑜𝑟𝑒, ∫ 𝑑𝑡 =
𝑡 2
0

(Ans.)

Relationship Between Fourier Series, Fourier Transformation & Fourier Analysis:

Fourier Series Fourier Transformation Fourier Analysis

Fourier Analysis Fourier Synthesis
Deals with periodic Deals with both periodic
functions/signals. and aperiodic
functions/signals. Breaks signals into its Synthesizes signals into
Converts a periodic Converts time domain constituent parts. its constituent parts.
function into an infinite signal into frequency
series of sine and/or domain and vice-versa.
cosine.

2 𝑓𝑜𝑟 |𝑡| < 1

Example 26: If 𝑓 (𝑡) = {1 − 𝑡 ; find the Fourier Transformation of 𝑓 (𝑡).
0 𝑓𝑜𝑟 |𝑡| > 1

Soln.: We know,
∞

𝐹. 𝑇. {𝑓(𝑡)} = 𝐹(𝜔) = ∫ 𝑓(𝑡) ∙ 𝑒 𝑖𝜔𝑡 𝑑𝑡

−∞
−1 1 ∞

= ∫ 𝑓(𝑡)𝑒 𝑖𝜔𝑡 𝑑𝑡 + ∫ 𝑓(𝑡)𝑒 𝑖𝜔𝑡 𝑑𝑡 + ∫ 𝑓(𝑡)𝑒 𝑖𝜔𝑡 𝑑𝑡

−∞ −1 1
1

= 0 + ∫ 𝑓(𝑡)𝑒 𝑖𝜔𝑡 𝑑𝑡 + 0
−1
1

= ∫(1 − 𝑡 2 ) ∙ 𝑒 𝑖𝜔𝑡 𝑑𝑡 … … … … … … (𝑖)

−1
Now,
2) 𝑖𝜔𝑡 2)
𝑒 𝑖𝜔𝑡 𝑒 𝑖𝜔𝑡
∫(1 − 𝑡 ∙𝑒 𝑑𝑡 = (1 − 𝑡 ∙ − ∫(−2𝑡) 𝑑𝑡
𝑖𝜔 𝑖𝜔
𝑒 𝑖𝜔𝑡 2 𝑒 𝑖𝜔𝑡 𝑒 𝑖𝜔𝑡
= (1 − 𝑡 2 ) ∙ + [𝑡 ∙ −∫ 𝑑𝑡]
𝑖𝜔 𝑖𝜔 𝑖𝜔 𝑖𝜔
(1 − 𝑡 2 )𝑒 𝑖𝜔𝑡 2𝑡𝑒 𝑖𝜔𝑡 2𝑒 𝑖𝜔𝑡
= − +
𝑖𝜔 𝜔2 𝑖𝜔 3

From equation (i),

1

𝐹. 𝑇. {𝑓(𝑡)} = 𝐹(𝜔) = ∫(1 − 𝑡 2 ) ∙ 𝑒 𝑖𝜔𝑡 𝑑𝑡

−1
1
(1 − 𝑡 2 )𝑒 𝑖𝜔𝑡 2𝑡𝑒 𝑖𝜔𝑡 2𝑒 𝑖𝜔𝑡
=[ − + ]
𝑖𝜔 𝜔2 𝑖𝜔 3 −1

2𝑒 𝑖𝜔 2𝑒 −𝑖𝜔 2𝑒 𝑖𝜔 2𝑒 −𝑖𝜔
=0− − + −
𝜔2 𝜔2 𝑖𝜔 3 𝑖𝜔 3
2 𝑖𝜔 −𝑖𝜔
2 (𝑒 𝑖𝜔 − 𝑒 −𝑖𝜔 )
=− (𝑒 − 𝑒 ) + ∙
𝜔2 𝜔3 𝑖
2 2
= − 2 ∙ 2 cos 𝜔 + 3 ∙ 2 sin 𝜔
𝜔 𝜔
4 sin 𝜔 4 cos 𝜔
= −
𝜔3 𝜔2
4 sin 𝜔
= 2( + cos 𝜔)
𝜔 𝜔
(Ans.)

End of Part 2: Fourier Analysis