Example 4.

1: Find the FT for the signal

𝑥(𝑡) = 𝑒 −𝑎𝑡 𝑢(𝑡) 𝑎>0

∞ ∞ ∞
𝑋(𝜔) = ∫ 𝑥(𝑡)𝑒 −𝑗𝜔𝑡 𝑑𝑡 = ∫ 𝑒 −𝑎𝑡 𝑒 −𝑗𝜔𝑡 𝑑𝑡 = ∫ 𝑒 −(𝑎+𝑗𝜔)𝑡 𝑑𝑡
−∞ 0 0

−1 ∞ −1 1
𝑋(𝜔) = 𝑒 −(𝑎+𝑗𝜔)𝑡 |0 = [0 − 1] =
𝑎 + 𝑗𝜔 𝑎 + 𝑗𝜔 𝑎 + 𝑗𝜔

Since this Fourier transform is complex valued, to plot it as a function of 𝜔, we express 𝑋(𝜔)
in terms of its magnitude and phase:

1 1
|𝑋(𝜔)| = | |=
𝑎 + 𝑗𝜔 √𝑎2 + 𝜔 2

𝜔
∢𝑋(𝜔) = − tan−1 ( )
𝑎

Example 4.1: Find the FT for the signal

𝑥(𝑡) = 𝑒 −𝑎|𝑡| 𝑎>0

 𝑒 −𝑎|𝑡| = 𝑒 −𝑎𝑡 𝑢(𝑡) + 𝑒 𝑎𝑡 𝑢(−𝑡)

Solution I:

∞ 0 ∞
−𝑗𝜔𝑡 𝑎𝑡 −𝑗𝜔𝑡
𝑋(𝜔) = ∫ 𝑥(𝑡)𝑒 𝑑𝑡 = ∫ 𝑒 𝑒 𝑑𝑡 + ∫ 𝑒 −𝑎𝑡 𝑒 −𝑗𝜔𝑡 𝑑𝑡
−∞ −∞ 0

0 ∞
𝑋(𝜔) = ∫ 𝑒 𝑎−𝑗𝜔𝑡 𝑑𝑡 + ∫ 𝑒 −(𝑎+𝑗𝜔)𝑡 𝑑𝑡
−∞ 0

1 0 −1 ∞
𝑋(𝜔) = [ 𝑒 (𝑎−𝑗𝜔)𝑡 |−∞ ] + [ 𝑒 −(𝑎+𝑗𝜔)𝑡 |0 ]
𝑎 − 𝑗𝜔 𝑎 + 𝑗𝜔

1 −1
𝑋(𝜔) = [ (1 − 0)] + [ (0 − 1)]
𝑎 − 𝑗𝜔 𝑎 + 𝑗𝜔

1 1 𝑎 + 𝑗𝜔 + 𝑎 − 𝑗𝜔
𝑋(𝜔) = + =
𝑎 − 𝑗𝜔 𝑎 + 𝑗𝜔 (𝑎 − 𝑗𝜔)(𝑎 + 𝑗𝜔)

2𝑎
𝐻(𝜔) = = |𝐻(𝜔)|
𝑎2 + 𝜔2

In this case 𝑋(𝜔) is real, therefore no phase spectrum needed.

Solution II: We can also solve it using FT Table and its properties

𝑒 −𝑎|𝑡| = 𝑒 −𝑎𝑡 𝑢(𝑡) + 𝑒 𝑎𝑡 𝑢(−𝑡)

 1
From FT table 𝑒 −𝑎𝑡 𝑢(𝑡) ↔ 𝑎+𝑗𝜔

Using Time Reversal Property, find the FT of 𝑒 𝑎𝑡 𝑢(−𝑡)

1 1
𝑥(−𝑡) ↔ 𝑋(−𝜔) 𝑒 −𝑎𝑡 𝑢(𝑡) ↔ 𝑎+𝑗𝜔 𝑒 𝑎𝑡 𝑢(−𝑡) ↔ 𝑎−𝑗𝜔

1 1 2𝑎
Therefore, 𝑒 −𝑎|𝑡| ↔ 𝑎+𝑗𝜔 + 𝑎−𝑗𝜔 = 𝑎2 +𝜔2

Example 4.3: Find the FT of the unit impulse

𝑥(𝑡) = 𝛿(𝑡)

∞ ∞
𝑋(𝜔) = ∫ 𝑥(𝑡)𝑒 −𝑗𝜔𝑡 𝑑𝑡 = ∫ 𝛿(𝑡)𝑒 −𝑗𝜔𝑡 𝑑𝑡
−∞ −∞

Using the sampling property of the impulse function

∞
𝑋(𝜔) = ∫ 𝛿(𝑡)𝑒 −𝑗𝜔𝑡 𝑑𝑡 = 𝑒 −𝑗𝜔(𝑡=0) = 1
−∞

In this case 𝑋(𝜔) is real, therefore no phase spectrum needed.

That is, the unit impulse bas a Fourier transform consisting of equal contributions at
frequencies. #

𝑥 (𝑡) = 𝛿(𝑡)
𝑋(𝜔) = 1

Example 4.4: Consider the rectangular pulse signal

𝑡 1 − 𝑇1 < 𝑡 < 𝑇1
𝑥(𝑡) = 𝑟𝑒𝑐𝑡( )={
2𝑇1 0 otherwise
 ∞
𝑋(𝜔) = 𝑃𝜏 (𝜔) = ∫ 𝑥(𝑡)𝑒 −𝑗𝜔𝑡 𝑑𝑡
−∞

Since 𝑥(𝑡) = 1 for |𝑡| < 𝑇1 , and since it is zero for |𝑡| > 𝑇1,

∞
𝑋(𝜔) = ∫ 𝑥(𝑡)𝑒 −𝑗𝜔𝑡 𝑑𝑡
−∞

𝑇 −1 𝑇1 −1 2 𝑒 𝑗𝜔𝑇1 −𝑒 −𝑗𝜔𝑇1
𝑋(𝜔) = ∫−𝑇1 𝑒 −𝑗𝜔𝑡 𝑑𝑡 = 𝑗𝜔 𝑒 −𝑗𝜔𝑡 |−𝑇 = 𝑗𝜔 [𝑒 −𝑗𝜔𝑇1 − 𝑒 𝑗𝜔𝑇1 ] = 𝜔 [ ]
1 1 2𝑗

2 sin(𝑇𝜔1 )
𝑋(𝜔) = [sin(𝑇1 𝜔)] = 2𝑇1 = 2𝑇1 sinc(𝑇1 𝜔)
𝜔 𝑇1 𝜔

Note 1: For this case, the FT is real valued so we can plot it using a single plot.

Note 2: This FT can also be expressed as using normalized sinc function (Oppenheim).

𝑇𝜔
sin(𝑇1 𝜔) sin (𝜋 1 )
𝑋(𝜔) = 2𝑇1 = 2𝑇1 𝜋
𝑇1 𝜔 𝑇1 𝜔
𝜋 𝜋

sin 𝜋𝑥
sinc 𝑐 =
𝜋𝑥

𝑇1 𝜔
Here 𝑥= 𝜋

𝑇1
𝑋(𝜔) = 2𝑇1 sinc ( 𝜔)
𝜋
Example 4.5: Consider the signal 𝑥(𝑡) whose Fourier transform 𝑋(𝜔) is

𝜔 1 −𝑊 <𝜔 <𝑊
𝑋(𝜔) = 𝑟𝑒𝑐𝑡( )={
2𝑊 0 otherwise

Find the inverse transform of 𝑋(𝜔).

1 ∞
𝑥(𝑡) = ∫ 𝑋(𝜔)𝑒 𝑗𝜔𝑡 𝑑𝜔
2𝜋 −∞

1 𝑊 𝑗𝜔𝑡 1 𝑊 1 1 𝑗𝑊𝑡
𝑥(𝑡) = ∫ 1𝑒 𝑑𝜔 = 𝑒 𝑗𝜔𝑡 |−𝑊 = [𝑒 − 𝑒 −𝑗𝑊𝑡 ]
2𝜋 −𝑊 𝑡 2𝜋 𝑗𝑡

1 1 𝑒 𝑗𝑊𝑡 − 𝑒 −𝑗𝑊𝑡 1
𝑥(𝑡) = [ ] = sin(𝑊𝑡)
𝜋𝜔 2𝑗 𝜋𝑡

We can express 𝑥(𝑡) as sinc function using its normalized and unnormalized forms

𝑊 sin(𝑊𝑡) 𝑊
𝑥(𝑡) = = sinc(𝑊𝑡) Unnormalized
𝜋 𝑊𝑡 𝜋
 𝑊𝑡
𝑊 sin(𝜋 𝜋 ) 𝑊 𝑊
𝑥(𝑡) = 𝑊𝑡 = sinc ( 𝜋 𝑡) normalized
𝜋 𝜋 𝜋
𝜋

Note: Compare the fourier transform pairs

𝑡
𝑟𝑒𝑐𝑡( ) ↔ 2𝑇1 sinc(𝑇1 𝜔)
2𝑇1

𝑊 𝜔
sinc(𝑊𝑡) ↔ 𝑟𝑒𝑐𝑡 ( )
𝜋 2𝑊

Example 4.6: Fourier transform of a periodic signal

∞ 𝑘

𝑥(𝑡) = ∑ 𝑎𝑘 𝑒 𝑗𝑘𝜔0𝑡 ↔ 2𝜋 ∑ 𝑎𝑘 𝛿(𝜔 − 𝑘𝜔0 ) = 𝑋(𝜔)

𝑘=−∞ 𝑘=−∞

Consider again the square wave illustrated in Figure 4.1.

Find the FT of this square wave periodic signal.

Solution:

The Fourier series coefficients for this signal are given by

2𝑇1
𝑎0 =
𝑇

sin(𝑘𝜔0 𝑇1 )
𝑎𝑘 =
𝜋𝑘

The Fourier transform of the signal

𝑘 𝑘

𝐻(𝜔) = 2𝜋 ∑ 𝑎𝑘 𝛿(𝜔 − 𝑘𝜔0 ) = ∑ 2𝜋𝑎𝑘 𝛿(𝜔 − 𝑘𝜔0 )

𝑘=−∞ 𝑘=−∞

2𝑇1
𝑎0 =
𝑇

sin(𝑘𝜔0 𝑇1 )
𝑎𝑘 =
𝜋𝑘

For 𝑇 = 4𝑇1

2𝜋 2𝜋 𝜋
𝜔0 = = 4𝑇 𝜔0 𝑇1 =
𝑇 1 2

2𝑇1 2𝑇1
2𝜋𝑎0 = 2𝜋 = 2𝜋 =𝜋
𝑇 4𝑇1

1 𝜋 2 𝜋
2𝜋𝑎𝑘 = 2𝜋 𝑘𝜋 [sin(𝑘 2 )] = 𝑘 [sin(𝑘 2 )] 𝑘≠0

2 𝜋
2𝜋𝑎𝑘 = 𝑘 [sin(𝑘 2 )] = 0 For 𝑘 even and 𝑘 ≠ 0

𝜋
Also, sin(𝑘 2 ) alternates between ±1 for successive odd values of 𝑘. Therefore in this case

2 𝜋 2
2𝜋𝑎𝑘 = 𝑘 sin(𝑘 2 ) alternates between ± 𝑘 for successive odd values of 𝑘

1 1 1
𝑎1 = 𝑎−1 = 𝜋 𝑎3 = 𝑎−3 = − 3𝜋 𝑎5 = 𝑎−5 = 5𝜋
 2 2
2𝜋𝑎1 = 2𝜋𝑎−1 = 2 2𝜋𝑎3 = 2𝜋𝑎−3 = − 3 2𝜋𝑎5 = 2𝜋𝑎−5 = 5

Example4.7: Find the Fourier transforms of the everlasting sinusoid cos 𝜔0 𝑡.

𝑥(𝑡) = cos 𝜔0 𝑡

1
Recall the Euler formula cos 𝜔0 𝑡 = 2 [𝑒 𝑗𝜔0𝑡 + 𝑒 −𝑗𝜔0𝑡 ]

1 ⟺ 2𝜋𝛿(𝜔)

Using Frequency shifting property

1𝑒 𝑗𝜔0𝑡 ⟺ 2𝜋𝛿(𝜔 − 𝜔0 )

1𝑒 −𝑗𝜔0𝑡 ⟺ 2𝜋𝛿(𝜔 + 𝜔0 )

Therefore,

1
cos 𝜔0 𝑡 ⟺ 2 [2𝜋𝛿(𝜔 − 𝜔0 ) + 2𝜋𝛿(𝜔 + 𝜔0 )]

cos 𝜔0 𝑡 ⟺ π[𝛿(𝜔 − 𝜔0 ) + 𝛿(𝜔 + 𝜔0 )]

The spectrum of cos 𝜔0 𝑡 consists of two impulses at 𝜔0 and −𝜔0 . The result also follows from
qualitative reasoning. An everlasting sinusoid cos 𝜔0 𝑡 can be synthesized by two everlasting
exponentials, 𝑒 𝑗𝜔0𝑡 and 𝑒 −𝑗𝜔0𝑡 . Therefore, the Fourier spectrum consists of only two
components of frequencies 𝜔0 and −𝜔0.

𝑥 (𝑡) 𝑋(𝜔)

Example 4.7 (b): Find the Fourier transforms of the everlasting sinusoid cos 𝜔0 𝑡.

𝑥(𝑡) = sin 𝜔0 𝑡

1
Recall the Euler formula sin 𝜔0 𝑡 = 2𝑗 [𝑒 𝑗𝜔0𝑡 − 𝑒 −𝑗𝜔0𝑡 ]

1 ⟺ 2𝜋𝛿(𝜔)

Using Frequency shifting property

1𝑒 𝑗𝜔0𝑡 ⟺ 2𝜋𝛿(𝜔 − 𝜔0 )

1𝑒 −𝑗𝜔0𝑡 ⟺ 2𝜋𝛿(𝜔 + 𝜔0 )

Therefore,

1
sin 𝜔0 𝑡 ⟺ 2𝑗 [2𝜋𝛿(𝜔 − 𝜔0 ) − 2𝜋𝛿(𝜔 + 𝜔0 )]

π
sin 𝜔0 𝑡 ⟺ [𝛿(𝜔 − 𝜔0 ) − 𝛿(𝜔 + 𝜔0 )]
𝑗
Example 4.8: Fourier transform of a periodic signal

∞ 𝑘

𝑥(𝑡) = ∑ 𝑎𝑘 𝑒 𝑗𝑘𝜔0𝑡 ↔ 2𝜋 ∑ 𝑎𝑘 𝛿(𝜔 − 𝑘𝜔0 ) = 𝑋(𝜔)

𝑘=−∞ 𝑘=−∞

Find the Fourier transform of a periodic impulse train of the period 𝑇.

𝑥(𝑡) = ∑ 𝛿(𝑡 − 𝑘𝑇)

𝑘=−∞

We know that the Fourier series expansion of the periodic impulse train is

1
∑∞ ∞
−∞ 𝛿(𝑡 − 𝑘𝑇0 ) = ∑−∞ 𝑎𝑘 𝑒
𝑗𝑘𝜔0 𝑡
with 𝑎𝑘 = 𝑇

Therefore, the FT of this periodic impulse train using the relationship

∞ 𝑘
𝑗𝑘𝜔0 𝑡
𝑥(𝑡) = ∑ 𝑎𝑘 𝑒 ↔ 2𝜋 ∑ 𝑎𝑘 𝛿(𝜔 − 𝑘𝜔0 ) = 𝑋(𝜔)
𝑘=−∞ 𝑘=−∞
 𝑘 𝑘
2𝜋
𝑋(𝜔) = 2𝜋 ∑ 𝑎𝑘 𝛿(𝜔 − 𝑘𝜔0 ) = ∑ 𝛿(𝜔 − 𝑘𝜔0 )
𝑇
𝑘=−∞ 𝑘=−∞

2𝜋
𝜔0 =
𝑇

Example 4.9: Find the FT for the signal shown in figure

We can decompose the above signal using linearity and time shifting properties.

1
𝑥(𝑡) = 𝑥 (𝑡 − 2.5) + 𝑥2 (𝑡 − 2.5)
2 1

𝑡 1 − 1.5 < 𝑡 < 1.5

𝑥2 (𝑡) = 𝑟𝑒𝑐𝑡( ) = {
3 0 otherwise

𝑡 1 − 0.5 < 𝑡 < 0.5

𝑥1 (𝑡) = 𝑟𝑒𝑐𝑡( ) = {
1 0 otherwise
 𝑡 𝜏
From FT Table Π (𝜏) ↔ 𝜏sinc (2 𝜔)

1 1
𝑥1 (𝑡) ↔ 1sinc ( 𝜔) = sinc ( 𝜔)
2 2

3 3
𝑥2 (𝑡) ↔ 3sinc ( 𝜔) = 3 sinc ( 𝜔)
2 2

Using Time Shifting Property 𝑥(𝑡 − 𝑡0 ) ↔ 𝑋(𝜔)𝑒 −𝑗𝜔𝑡0

1
𝑥1 (𝑡 − 2.5) ↔ sinc ( 𝜔) 𝑒 −𝑗2.5𝜔
2

3
𝑥2 (𝑡 − 2.5) ↔ 3 sinc ( 𝜔) 𝑒 −𝑗2.5𝜔
2

Using linearity property

1 1 1 3
𝑥1 (𝑡 − 2.5) + 𝑥2 (𝑡 − 2.5) ↔ sinc ( 𝜔) 𝑒 −𝑗2.5𝜔 + 3 sinc ( 𝜔) 𝑒 −𝑗2.5𝜔
2 2 2 2

1 1 3
𝑋(𝜔) = sinc ( 𝜔) 𝑒 −𝑗2.5𝜔 + 3 sinc ( 𝜔) 𝑒 −𝑗2.5𝜔
2 2 2

𝑒 −𝑗2.5𝜔 1 3
𝑋(𝜔) = [sinc ( 𝜔) + 6 sinc ( 𝜔)]
2 2 2

Example 4.11: Determine the Fourier transform 𝑋(𝜔) of the unit step 𝑥(𝑡) = 𝑢(𝑡), using
integration property.

𝑥(𝑡) = 𝑢(𝑡)

Let 𝑔(𝑡) = 𝛿(𝑡)

𝑡
𝑥(𝑡) = ∫ 𝑔(𝜏)𝑑𝜏
−∞

Taking FT on both sides and application of Integration property

 1
𝑋(𝜔) = 𝐺(𝜔) + 𝜋𝐺(0)𝛿(𝜔)
𝑗𝜔

𝐺(𝜔) = 1 𝐺(0) = 1

1
𝑋(𝜔) = + 𝜋𝛿(𝜔)
𝑗𝜔

1
𝑢(𝑡) ↔ + 𝜋𝛿(𝜔)
𝑗𝜔

Example 4.12: Determine the Fourier transform 𝑋(𝜔) for the signal 𝑥(𝑡) displayed in Figure
using differentiation property.

𝑑
Taking derivative, we obtain the time domain signal 𝑔(𝑡) = 𝑑𝑡 𝑥(𝑡)

𝑑 𝑡
𝑥(𝑡) = Π ( ) − 𝛿(𝑡 + 1) − 𝛿(𝑡 − 1)
𝑑𝑡 2

Using FT of 𝛿(𝑡) and Time shifting property

𝛿(𝑡 + 1) ↔ 𝑒 𝑗𝜔 𝛿(𝑡 − 1) ↔ 𝑒 −𝑗𝜔

𝑡 𝑡 𝜏
Using FT of Π ( ) Π ( ) ↔ 𝜏sinc ( 𝜔)
𝜏 𝜏 2
 𝑡 2
Π ( ) ↔ 2sinc ( 𝜔) = 2sinc(𝜔)
2 2

Finally taking the FT on both sides of

𝑑 𝑡
𝑥(𝑡) = Π ( ) − 𝛿(𝑡 + 1) − 𝛿(𝑡 − 1)
𝑑𝑡 2

By applying the linearity and differentiation property

2
𝑗𝜔𝑋(𝜔) = sinc ( 𝜔) − 𝑒 𝑗𝜔 − 𝑒 −𝑗𝜔
2

1 𝑒 𝑗𝜔 + 𝑒 −𝑗𝜔
𝑋(𝜔) = 2sinc(𝜔) −
𝑗𝜔 𝑗𝜔

1 2 𝑒 𝑗𝜔 + 𝑒 −𝑗𝜔
𝑋(𝜔) = 2sinc(𝜔) −
𝑗𝜔 𝑗𝜔 2

2 2
𝑋(𝜔) = sinc(𝜔) − cos(𝜔)
𝑗𝜔 𝑗𝜔

Example 4.13: Using the duality (symmetry) property, find the Fourier transform 𝐺(𝜔) of the
signal

2
𝑔(𝑡) =
1 + 𝑡2

From FT table, we have a Fourier transform pair in which the Fourier transform, as a function
of , had 𝜔a form similar to that of the signal 𝑔(𝑡).

2𝑏
𝑒 −𝑏|𝑡| ↔
𝑏2 + 𝜔2

For 𝑏 = 1

2
𝑒 −|𝑡| ↔
1 + 𝜔2

Now duality property

If 𝑥(𝑡) ↔ 𝑋(𝜔) then 𝑋(𝑡) ↔ 2𝜋𝑥(−𝜔)

2
↔ 2𝜋𝑒 −|−𝜔| = 2𝜋𝑒 −|𝜔|
1 + 𝑡2

Example 4. 14: For each of the Fourier transforms shown in Figure, evaluate the following
time-domain expressions using Parseval’s theorem.

∞
𝐸 = ∫ |𝑥(𝑡)|2 𝑑𝑡
−∞

Parseval’s theorem:

∞
1 ∞
𝐸=∫ |𝑥(𝑡)|2 𝑑𝑡 = ∫ |𝑋(𝜔)|2 𝑑𝜔
−∞ 2𝜋 −∞

For Signal (a)

1 ∞
𝐸= ∫ |𝑋(𝜔)|2 𝑑𝜔
2𝜋 −∞

∞
The integral ∫−∞|𝑋(𝜔)|2 𝑑𝜔 is area under the curve |𝑋(𝜔)|2

2
1 2 √𝜋 2
𝐸= [((√𝜋) × 0.5) + (( ) × 1) + ((√𝜋) × 0.5)]
2𝜋 2

1 𝜋 𝜋 𝜋
𝐸= [( ) + ( ) + ( )]
2𝜋 2 4 2

1 2𝜋 + 𝜋 + 2𝜋 1 5𝜋 5
𝐸= [ ]= [ ]=
2𝜋 4 2𝜋 4 8
For Signal (b)

1 ∞
𝐸= ∫ |𝑋(𝜔)|2 𝑑𝜔
2𝜋 −∞

Since 𝑋(𝜔) is imaginary

1 ∞
𝐸= ∫ 𝑋(𝑗𝜔)𝑋 ∗ (𝑗𝜔)𝑑𝜔
2𝜋 −∞

0 1
1
𝐸= [∫ (−𝑗√𝜋)(𝑗√𝜋)𝑑𝜔 + ∫ (𝑗√𝜋)(−𝑗√𝜋)𝑑𝜔]
2𝜋 −1 0

0 1 0 1
1 1
𝐸= [∫ −𝑗 2 𝜋𝑑𝜔 + ∫ −𝑗 2 𝜋𝑑𝜔] = [∫ 𝜋𝑑𝜔 + ∫ 𝜋𝑑𝜔]
2𝜋 −1 0 2𝜋 −1 0

0 1
𝜋 1 1
= [∫ 𝑑𝜔 + ∫ 𝑑𝜔] = [𝜔|0−1 + 𝜔|10 ] = [0 + 1 + 1 − 0] = 1
2𝜋 −1 0 2 2

Example 4. 15: Let the impulse response of an LTI system is

ℎ(𝑡) = 𝛿(𝑡 − 𝑡0 )

Find output 𝑦(𝑡) for the input 𝑥(𝑡) using Convolution property.

Solution:

ℎ(𝑡) = 𝛿(𝑡 − 𝑡0 )

𝑦(𝑡) = 𝑥(𝑡) ∗ ℎ(𝑡)

Using Convolution property of Fourier transform

𝑌(𝜔) = 𝑋(𝜔)𝐻(𝜔)

𝛿(𝑡) ↔ 1

ℎ(𝑡) = 𝛿(𝑡 − 𝑡0 ) ↔ 𝑒 −𝑗𝜔𝑡0 = 𝐻(𝜔)

 𝐻(𝜔) = 𝑒 −𝑗𝜔𝑡0

𝑌(𝜔) = 𝑋(𝜔)𝐻(𝜔) = 𝑋(𝜔)𝑒 −𝑗𝜔𝑡0

Taking Inverse Fourier transform on both sides and by using the time shifting property
𝑥(𝑡 − 𝑡0 ) ↔ 𝑋(𝜔)𝑒 −𝑗𝜔𝑡0

𝑌(𝜔) = 𝑥(𝑡 − 𝑡0 )

Example 4. 16: Find 𝐻(𝜔) for an differentiator

Solution:

For a differentiator, the input 𝑥(𝑡) and the output 𝑦(𝑡) are related by

𝑑
𝑦(𝑡) = 𝑥(𝑡)
𝑑𝑡

Applying differentiation property and taking FT on both sides

𝑌(𝜔) = 𝑗𝜔𝑋(𝜔)

Comparing with 𝑌(𝜔) = 𝑋(𝜔)𝐻(𝜔)

𝐻(𝜔) = 𝑗𝜔

Example 4. 17: Consider an LTI system whose impulse response is ℎ(𝑡) = 𝑢(𝑡).

Verify that the LTI system is an integrator.

Solution:

1
ℎ(𝑡) = 𝑢(𝑡) ↔ 𝜋𝛿(𝜔) +
𝑗𝜔

1
𝐻(𝜔) = 𝜋𝛿(𝜔) +
𝑗𝜔

𝑌(𝜔) = 𝐻(𝜔)𝑋(𝜔)
 1
𝑌(𝜔) = [𝜋𝛿(𝜔) + ]𝑋(𝜔)
𝑗𝜔

1
𝑌(𝜔) = 𝜋𝛿(𝜔)𝑋(𝜔) + 𝑋(𝜔)
𝑗𝜔

Using sampling property of 𝜋𝛿(𝜔) 𝜋𝛿(𝜔 − 0)𝑋(𝜔) = 𝜋𝛿(𝜔 − 0)𝑋(0)

1
𝑌(𝜔) = 𝜋𝛿(𝜔)𝑋(0) + 𝑋(𝜔)
𝑗𝜔

Now comparing with the equation of integration property and taking FT on both sides

𝑡
𝑦(𝑡) = ∫ 𝑥(𝑢)𝑑𝑢
−∞

1
𝑌(𝜔) = 𝜋𝛿(𝜔)𝑋(0) + 𝑋(𝜔)
𝑗𝜔

Hence, a system with impulse response ℎ(𝑡) = 𝑢(𝑡) is an integrator.

Example 4. 19: Consider the response of an LTI system with impulse response

ℎ(𝑡) = 𝑒 −𝑎𝑡 𝑢(𝑡) 𝑎>0

to the input signal

𝑥(𝑡) = 𝑒 −𝑏 𝑢(𝑡) 𝑏>0

Find the output 𝑦(𝑡) by using property of multiplication in frequency domain in instead of
direct convolution.

Solution:

𝑌(𝜔) = 𝐻(𝜔)𝑋(𝜔)

1
𝐻(𝜔) =
𝑎 + 𝑗𝜔
 1
𝑋(𝜔) =
𝑏 + 𝑗𝜔

1
𝑌(𝜔) =
(𝑎 + 𝑗𝜔)(𝑏 + 𝑗𝜔)

Expanding 𝑌(𝜔) by partial fraction method

1 𝐴 𝐵
= +
(𝑎 + 𝑗𝜔)(𝑏 + 𝑗𝜔) 𝑎 + 𝑗𝜔 𝑏 + 𝑗𝜔

1 1
𝐴 = [(𝑎 + 𝑗𝜔)𝑌(𝜔)]|𝑎=−𝑗𝜔 = [(𝑎 + 𝑗𝜔) ]| =
(𝑎 + 𝑗𝜔)(𝑏 + 𝑗𝜔) 𝑎=−𝑗𝜔 𝑏 − 𝑎

1 1
𝐵 = [(𝑏 + 𝑗𝜔)𝑌(𝜔)]|𝑏=−𝑗𝜔 = [(𝑏 + 𝑗𝜔) ]| = = −𝐴
(𝑎 + 𝑗𝜔)(𝑏 + 𝑗𝜔) 𝑎=−𝑗𝜔 𝑎 − 𝑏

1 −1
1 𝑏 − 𝑎
𝑌(𝜔) = = + −𝑎
𝑏
(𝑎 + 𝑗𝜔)(𝑏 + 𝑗𝜔) 𝑎 + 𝑗𝜔 𝑏 + 𝑗𝜔

1 1 −1
𝑌(𝜔) = [ + ]
𝑏 − 𝑎 𝑎 + 𝑗𝜔 𝑏 + 𝑗𝜔

Taking inverse FT on both sides

1
𝑦(𝑡) = [𝑒 −𝑎𝑡 𝑢(𝑡) − 𝑒 −𝑏𝑡 𝑢(𝑡)]
𝑏−𝑎

Special Case: If 𝑎 = 𝑏

1 1
𝑌(𝜔) = =
(𝑎 + 𝑗𝜔)(𝑏 + 𝑗𝜔) (𝑎 + 𝑗𝜔)2

𝑑 1 1
We can recognize that 𝑗 𝑑𝜔 [𝑎+𝑗𝜔] = (𝑎+𝑗𝜔)2

𝑑
Therefore, we can apply the dual of differentiation property 𝑡𝑥(𝑡) ↔ 𝑗 𝑑𝜔 𝑋(𝜔)

1
𝑒 −𝑎𝑡 𝑢(𝑡) ↔
𝑎 + 𝑗𝜔
 𝑑 1
𝑡𝑒 −𝑎𝑡 𝑢(𝑡) ↔ 𝑗 [ ]
𝑑𝜔 𝑎 + 𝑗𝜔

𝑑 −𝑗
(𝑎 + 𝑗𝜔)−1 = −𝑗(𝑎 + 𝑗𝜔)−2 =
𝑑𝜔 (𝑎 + 𝑗𝜔)2

−𝑗 1
𝑡𝑒 −𝑎𝑡 𝑢(𝑡) ↔ 𝑗 =
(𝑎 + 𝑗𝜔)2 (𝑎 + 𝑗𝜔)2

𝑦(𝑡) = 𝑡𝑒 −𝑎𝑡 𝑢(𝑡)

Example 4.24:

𝑑𝑦
+ 𝑎𝑦(𝑡) = 𝑥(𝑡)
𝑑𝑡

1 1
𝐻(𝑗𝜔) = =
𝑎0 (𝑗𝜔)0 + 𝑎1 (𝑗𝜔)1 𝑎 + 𝑗𝜔

ℎ(𝑡) = 𝑒 −𝑎𝑡 𝑢(𝑡)

Example 4.25:

𝑑2 𝑦 𝑑𝑦 𝑑𝑥
2
+4 + 3𝑦(𝑡) = + 2𝑥(𝑡)
𝑑𝑡 𝑑𝑡 𝑑𝑡

𝑏0 (𝑗𝜔)0 + 𝑏1 (𝑗𝜔)1 2(𝑗𝜔)0 + 1(𝑗𝜔)1

𝐻(𝑗𝜔) = =
𝑎0 (𝑗𝜔)0 + 𝑎1 (𝑗𝜔)1 + 𝑎2 (𝑗𝜔)2 3(𝑗𝜔)0 + 4(𝑗𝜔)1 + 1(𝑗𝜔)2

2 + 𝑗𝜔 2 + 𝑗𝜔
𝐻(𝑗𝜔) = 2 2
=
3 + 4𝑗𝜔 + 𝑗 𝜔 (1 + 𝑗𝜔)(3 + 𝑗𝜔)

Then, using the method of partial-fraction expansion, we find that

2 + 𝑗𝜔 1⁄2 1⁄3
𝐻(𝑗𝜔) = = +
(1 + 𝑗𝜔)(3 + 𝑗𝜔) (1 + 𝑗𝜔) (3 + 𝑗𝜔)

1 −𝑡 1
ℎ(𝑡) = 𝑒 𝑢(𝑡) + 𝑒 −3𝑡 𝑢(𝑡)
2 2
Example 4.26: Consider the system of Example 4.25, and suppose that the input is

𝑥(𝑡) = 𝑒 −𝑡 𝑢(𝑡)

𝑌(𝑗𝜔) = 𝐻(𝑗𝜔)𝑋(𝑗𝜔)

1
𝑋(𝑗𝜔) =
𝑗𝜔 + 1

𝑗𝜔 + 2 1
𝑌(𝑗𝜔) = [ ][ ]
(𝑗𝜔 + 1)(𝑗𝜔 + 3) 𝑗𝜔 + 1

𝑗𝜔 + 2
𝑌(𝑗𝜔) =
(𝑗𝜔 + 1)2 (𝑗𝜔 + 3)

In this case the partial-fraction expansion takes the form

𝑗𝜔 + 2 𝐴11 𝐴12 𝐴21

𝑌(𝑗𝜔) = = + +
(𝑗𝜔 + 1)2 (𝑗𝜔 + 3) 𝑗𝜔 + 1 (𝑗𝜔 + 1)2 𝑗𝜔 + 3

where 𝐴11 , 𝐴12 , and 𝐴21 are constants to be determined

1 1 1
𝐴11 = 4 𝐴12 = 2 𝐴21 = − 4

1 1 1
−
𝑌(𝑗𝜔) = 4 + 2 + 4
𝑗𝜔 + 1 (𝑗𝜔 + 1)2 𝑗𝜔 + 3

1 −𝑡 1 1
𝑦(𝑡) = 𝑒 𝑢(𝑡) + 𝑡𝑒 −𝑡 𝑢(𝑡) − 𝑒 −3𝑡 𝑢(𝑡)
4 4 4

1 1 1
𝑦(𝑡) = [ 𝑒 −𝑡 + 𝑡𝑒 −𝑡 − 𝑒 −3𝑡 ] 𝑢(𝑡)
4 4 4