ELEC 221 Lecture 08

Introducing the Fourier transform

Thurday 30 January 2025

1 / 27
Announcements

Hands-on tutorial/ Tutorial Assignment 2 tomorrow

2 / 27
Last time

Put frequency-domain concepts to use to better analyze, interpret

and design LTI systems (filters)

Moving-average (lowpass) filter:

3 / 27
Last time

From https://www.youtube.com/watch?v=OOu5KP3Gvx0
4 / 27
Last time

From https://www.youtube.com/watch?v=OOu5KP3Gvx0
5 / 27
Today

Learning outcomes:
Explain the concept of CT Fourier transform, and distinguish
it from the CT Fourier series
Compute the Fourier spectrum of a CT signal
Describe how the Fourier transform relates impulse and
frequency response of a system

6 / 27
Recap: Fourier series

So far, we have been working with the Fourier series representation

of periodic CT and DT signals:

CT synthesis equation:
→
!
x(t) = ck e jkω0 t
k=↑→

CT analysis equation:
"
1
ck = x(t)e ↑jkω0 t dt
T T

A periodic signal is composed of complex exponential signals with

integer multiples of the fundamental frequency ω0 = 2ε/T only
(harmonics).
7 / 27
The Fourier transform

The Fourier transform generalizes the Fourier series to aperiodic

signals. It involves all possible frequencies (frequency spectrum).

Fourier series:
→
! "
1
x(t) = ck e jkω0 t
, ck = x(t)e ↑jkω0 t dt
T T
k=↑→

Fourier transform:
" → " →
1
x(t) = X (jω)e jωt dω, X (jω) = x(t)e ↑jωt dt
2ε ↑→ ↑→

How do we get here?

8 / 27
Towards the Fourier transform

Consider the following square wave:

#
1, |t| < T1
x(t) =
0, T1 < |t| < T /2

Image credit: Oppenheim chapter 4.1

9 / 27
Towards the Fourier transform

#
1, |t| < T1
x(t) =
0, T1 < |t| < T /2

Let’s compute its Fourier coe!cients.

"
1
ck = x(t)e ↑jkω0 t dt
T T
Start with c0 :
"
1 T /2
c0 = x(t)dt
T ↑T /2
"
1 T1
= dt
T ↑T1
2T1
=
T
10 / 27
Towards the Fourier transform

#
1, |t| < T1
x(t) =
0, T1 < |t| < T /2
Now the ck :
" T /2
1
ck = x(t)e ↑jkω0 t dt
T ↑T /2
" T1
1
= e ↑jkω0 t dt
T ↑T1
1
= → e ↑jkω0 t |T 1
↑T1
jkω0 T
1 $ ↑jkω0 T1 %
= → e → e jkω0 T1
jkω0 T
2 sin(kω0 T1 )
=
kω0 T
11 / 27
Towards the Fourier transform

2T1 2 sin(kω0 T1 )
c0 = , ck =
T kω0 T

Let’s rearrange a bit:

#
2T1 , k=0
Tck = 2 sin(kω0 T1 )
kω0 , k ↑= 0
&
2 sin(ωT1 ) &&
= &
ω ω=kω0

The Fourier coe!cients are samples of the function 2 sin(ωT1 )/ω

taken at frequencies with spacing ω0 .

The function 2 sin(ωT1 )/ω represents the “envelope” of Tck .

12 / 27
Towards the Fourier transform

Spacing of samples at integer values of ω0 = 2ε/T .

Image credit: Oppenheim chapter 4.1

13 / 27
Towards the Fourier transform

Suppose T grows (but T1 stays the same)?

Image credit: Oppenheim chapter 4.1

14 / 27
Towards the Fourier transform

As T grows, ω0 = 2ε/T becomes smaller and smaller, so the

integer multiples of it get closer and closer together.

Image credit: Oppenheim chapter 4.1

15 / 27
Towards the Fourier transform

Consider an aperiodic signal x(t) and its periodic extension, x̃(t)

with period T .

Image credit: Oppenheim chapter 4.1

16 / 27
Motivation: Fourier transform

We know:
→
! "
1
x̃(t) = ck e jkω0 t , ck = x̃(t)e ↑jkω0 t dt
T T
k=↑→

Rewrite the coe!cients

"
1 T /2
ck = x̃(t)e ↑jkω0 t dt =
T ↑T /2

Let’s define " →

X (jω) = x(t)e ↑jωt dt
↑→
so that
1
ck = X (jkω0 )
T
17 / 27
Motivation: Fourier transform

We can put this back into our Fourier series:

→
! 1
x̃(t) = X (jkω0 )e jkω0 t
T
k=↑→

→
1 !
= X (jkω0 )e jkω0 t ω0
2ε
k=↑→

18 / 27
Motivation: Fourier transform

Two important things happen when T goes to ↓ (T ↔ ↓):

1. x̃(t) will look just like x(t) for large enough T

2. ω0 will get smaller and smaller

→
1 !
x(t) = lim x̃(t) = lim X (jkω0 )e jkω0 t ω0
T ↓→ ω0 ↓0 2ε
k=↑→
" →
1
= X (jω)e jωt dω
2ε ↑→

This is the (inverse) Fourier transform.

19 / 27
The Fourier transform

Inverse Fourier transform (synthesis equation):

" →
1
x(t) = X (jω)e jωt dω
2ε ↑→

Fourier transform (analysis equation, or Fourier spectrum):

" →
X (jω) = x(t)e ↑jωt dt
↑→

Note: Sometimes the 1/2ε prefactor appears on the spectrum, or

↗
sometimes both versions have 1/ 2ε.

20 / 27
Fourier transform and impulse response

We have actually already (unknowingly) seen the Fourier transform

when we discussed system functions and frequency response.

Consider an LTI system with impulse response h(t):

" →
x(t) ↔ y (t) = x(ϑ )h(t → ϑ )dϑ
↑→
" →
e jωt
↔ e ↑jω(t↑ε ) h(ϑ )dϑ
↑→
" →
=e jωt
e ↑jωε h(ϑ )dϑ
↑→
jωt
=e H(jω)

21 / 27
Fourier transform and impulse response

The frequency response

" →
H(jω) = h(ϑ )e ↑jωε dϑ
↑→

is the Fourier transform of the impulse response!

The impulse response

" →
1
h(t) = H(jω)e jωt dω
2ε ↑→

is the inverse Fourier transform of the frequency response.

22 / 27
Example

Compute the Fourier spectrum of:

#
cos(εt), |t| ↘ 1
x(t) =
0, |t| > 1

23 / 27
Example

#
cos(εt), |t| ↘ 1
x(t) =
0, |t| > 1
Start from the definition:

" →
X (jω) = x(t)e ↑jωt dt
↑→
" 1
= cos(εt)e ↑jωt dt
↑1
" 1 " 1
1 1
= e j(ϑ↑ω)t
dt + e ↑j(ϑ+ω)t dt
2 ↑1 2 ↑1

24 / 27
Example

" 1 " 1
1 1
X (jω) = e j(ϑ↑ω)t
dt + e ↑j(ϑ+ω)t dt
2 ↑1 2 ↑1

1 1 1 →1
= e j(ϑ↑ω)t |1↑1 + e ↑j(ϑ+ω)t |1↑1
2 j(ε → ω) 2 j(ε + ω)
= 1
2j(ω→ε)
1
(e j(ω→ε) ↑e →j(ω→ε) )↑ 2j(ω+ε) (e →j(ω+ε) ↑e j(ω+ε) )
sin(ε → ω) sin(ε + ω)
= +
ε→ω ε+ω

sin(ω) sin(ω)
= →
ε→ω ε+ω

25 / 27
Example

sin(ω) sin(ω)
X (jω) = →
ε→ω ε+ω

26 / 27
For next time

Content:
Properties of the CT Fourier transform

Recommended reading:
From today’s class: Oppenheim 4.0-4.1
Suggested problems: 4.1, 4.2a, 4.21abei, 4.22abde
For next class: Oppenheim 4.2-4.3

27 / 27