Signals and Systems

Fourier Transform of
Continuous-Time Signals
Ertem Tuncel
Professor & Chair of Electrical and Computer Engineering
University of CA, Riverside
 What about nonperiodic signals?
x(t)

t
• If the signal has finite duration, everything is fine.
• Extend the signal into a periodic one and
decompose onto .

with
 What about nonperiodic signals?
x(t)

with

• Alternatively,
 What about nonperiodic signals?

• Now define

so that

and
 What about nonperiodic signals?

• What happens if we set T to more than its

minimum possible value?

t
T
 What about nonperiodic signals?

• What happens if we set T to more than its

minimum possible value?

t
T
 What about nonperiodic signals?

• What happens if we set T to more than its

minimum possible value?

t
T
• As , we have and
 What about nonperiodic signals?

• As , we have and
• That implies
 What about nonperiodic signals?

= total area of
the rectangles
• Pictorially,

• As , where does the total area of the

rectangles go?
What about nonperiodic signals?
Continuous-time Fourier Transform
• This pair is known as the continuous-time Fourier
transform (CTFT)
 Properties
• Properties of CTFS carry over to CTFT:

imply
Linearity:
Time shifting:

Frequency shifting:

Time reversal:
Conjugation:
Linearity:
Time shifting:
Frequency shifting:
Time reversal:
Conjugation:
Convolution:

Multiplication:

Parseval's:
 Properties
• But there are also two more:
Differentiation in the time domain:

• Proof:
 Properties
Differentiation in the frequency domain:

• Proof:

Rearranging finishes the proof.

 Examples
• Find the CTFT of
• Solution:

• How do we plot this?

Real and imaginary parts separately
Magnitude and phase separately
 Examples
• If we normalize this family of functions as

we obtain from linearity that

and therefore that

Examples
Examples
 Implementation as a filter
• Note that if this is the impulse response of a
system, the system will suppress high
frequencies and boost low frequencies.

x(t) y(t)
 Implementation as a filter
• This filter is easy to implement once we
figure out that h(t) is the impulse response of
the system with differential equation

x(t) y(t)

DIFF
 Examples
• Find the CTFT of
for
• Solution: It is best to rewrite x(t) as

We already know that

Using the frequency shift property,

 Examples
• Find the CTFT of
for

• Intuitively, if is small enough, and is big

enough,
 Examples
• Indeed, when and ,
 Examples
• However, when and ,
 Examples
• And when and ,
 Implementation as a filter
• What system has an impulse response

• The homogeneous solution must be in the

form of and
• Therefore, the homogeneous part of the
differential equation must be

• The initial conditions need to be

 Implementation as a filter

• Since h(t) is not continuous at t = 0, the

most general possibility is

• Integrating both sides twice yields

 Implementation as a filter
• The system therefore has a differential
equation given by

x(t) y(t)

DIFF DIFF

DIFF
 Examples
• Find the signal x(t) whose CTFT is as below:

• Solution:
 Examples
• As an example, if ,
 Examples
• Conversely, find the CTFT of

• Solution:
 CTFT of periodic signals
• Normally, CTFS suffices in decomposing onto
complex exponentials.
• What if, just out of intellectual curiosity, we
compute the CTFT of a periodic signal?
• Since

and since CTFT is linear, it suffices to find the

CTFT of .
 CTFT of periodic signals
• Claim:

• Proof: Using the synthesis formula, we see

• Therefore,
 CTFT of periodic signals
• For example, if

then

Each impulse at is of amplitude

 CTFT of periodic signals
• Example: Find the CTFT of
x(t)
1

T t
• Solution: Recall that the CTFS coefficients were

Therefore,
 CTFT of periodic signals
• Example: Find the CTFT of
x(t)
1

T t