Chapter 1

Low Pass Equivalent of Bandpass

Any real, narrow band, and high frequency signal called a bandpass signal can be
represented in terms of a complex low frequency signal called the low pass equivalent
of the original bandpass signal. This chapter introduces about the low pass equivalent
of the original bandpass signal. The low pass equivalent concept can also be extended
to bandpass systems and is given in the following sections

1.1 Hilbert Transforms [1]

When the phase angles of all components of a given signal are shifted by ±90◦ , the
resulting function of time is known as the Hilbert Transform of the signal. Consider
a Fourier Transformable signal g(t) with its Fourier Transform denoted by G(f). The
Hilbert Transform of g(t) is denoted by ĝ(t) is defined by

Z∞
1 g(τ )
ĝ(t) = dτ (1.1)
π t−τ
−∞

The inverse Hilbert Transform is given by

Z∞
1 ĝ(τ )
g(t) = − dτ (1.2)
π t−τ
−∞

The functions g(t) and ĝ(t) are said to constitute a Hilbert Transform Pair. The
definition of the Hilbert Transform given in Equation (??) may be interpreted as the
convolution of g(t) with the time function 1/πt.

ĝ(t) = (1/πt) ∗ g(t) (1.3)

1
 Chapter 1. Low Pass Equivalent of Bandpass

We also know from the convolution theorem that the convolution of two functions in
the time domain is transformed into the multiplication of their Fourier transforms in
the frequency domain.


1
G̃(f ) = F G(f )
πt
= −jsgn (f ) G (f ) (1.4)

where sgn(f) is the signum function defined in the frequency domain as


 1, f > 0
sgn(f ) = (1.5)
 −1, f < 0

Equation (1.4) states that given a Fourier transformable signal g(t), we may obtain
the Fourier transform of ĝ(t) by passing g(t) through a linear time-invariant system
whose frequency response is equal to −j sgn(f ). This device or system produces a
phase shift of −90 for all positive frequencies and a phase shift of +90 degrees for all
negative frequencies. The amplitudes are unaffected by the transmission through the
device as shown in the Figure 1.1.

Figure 1.1: Magnitude and Phase Response of Hilbert Transform [1]

Dept of E&C, JNNCE Digital Communication 2

 Chapter 1. Low Pass Equivalent of Bandpass

Let us assume that g(t) is a real valued signal. The Hilbert transform has the
following properties

1. A signal g(t) and its Hilbert Transform ĝ(t) has same magnitude spectrum.
that is to say,
 
|G(f )| = Ĝ(f ) (1.6)
 

2. If ĝ(t) is the Hilbert transform of g(t), then the Hilbert transform of ĝ(t) is
−g(t). Another way of stating this property is to write

n o
arg [G(f )] = − arg Ĝ(f ) (1.7)

3. A signal g(t) and its Hilbert transform ĝ(t) are orthogonal over the entire time
interval (−∞, ∞)
Z∞
g(t)ĝ(t) = 0 (1.8)
−∞

1.2 Pre-Envelopes [1]

We can modify the frequency content of a real-valued signal g(t) such that all negative
frequency components are completely eliminated. The solution to this problem is to
define a complex valued signal g+ (t) called the pre-envelope of g(t), formally defined
as follows:
Consider a real valued signal g(t). We define the pre-envelope of the given signal
g(t) as the complex valued signal

g+ (t) = g(t) + jĝ(t) (1.9)

where ĝ(t) is the Hilbert transform of g(t). It should be noted that the given signal
g(t) is the real part of the pre-envelope g+ (t)
Let G+ (f ) denote the Fourier transform of g+ (t). Then, using Equation (1.9) and
Equation (1.4) we can write

G+ (f ) = G(f ) + sgn(f )G(f ) (1.10)

Dept of E&C, JNNCE Digital Communication 3

 Chapter 1. Low Pass Equivalent of Bandpass

Equation (1.10) can also be written as




 2G(f ), f > 0







G+ (f ) = G(0), f = 0 (1.11)









 0, f <0

(1.11) clearly shows that the pre-envelope of the signal g(t) has no frequency content
for all negative frequencies.

Figure 1.2: Pre-envelope of a Low-pass Signal

We may define the pre-envelope for negative frequencies as

g− (t) = g(t) − jĝ(t) (1.12)

The spectrum of the other pre-envelope g− (t) is given by




 2G(f ), f < 0







G− (f ) = G(0), f = 0 (1.13)









 0, f >0

Dept of E&C, JNNCE Digital Communication 4

 Chapter 1. Low Pass Equivalent of Bandpass

1.3 Complex Envelope of Band-Pass Signals [1]

Let the band-pass signal of interest be denoted by g(t) and its Fourier transform be
denoted by G(f ). The pre envelope of a g(t) is given by

g+ (t) = g̃(t)ej2πfc t (1.14)

We refer to g̃(t) as the complex envelope of the band-pass signal g(t).We find that
the spectrum of the pre-envelope g+ (t) is limited to the positive frequency band
fc − W ≤ f ≤ fc + W , as illustrated in Figure
Applying the frequency shifting property of Fourier Transform to (1.14) we get

G+ (f ) = G̃(f − fc ) (1.15)

From (1.15) and Figure we find that the spectrum of the complex envelope g̃(t) is
limited to band −W ≤ f ≤ W and centered at the origin. Thus we can say that the
complex envelope g̃(t) of a bandpass signal g(t) is a low pass signal.

Figure 1.3: (a) Magnitude spectrum of band-pass signal, pre-envelope and complex
envelope

Dept of E&C, JNNCE Digital Communication 5

 Chapter 1. Low Pass Equivalent of Bandpass

1.4 Canonical Representation of Bandpass Signals

[1] Let the band-pass signal of interest be denoted by g(t) and its Fourier transform
be denoted by G(f ). We know that from (1.9),

g+ (t) = g(t) + jĝ(t)

Therefore, the real part of the pre-envelope g+ (t) is equal to the original band-pass
signal g(t). We may therefore express the band-pass signal g(t) in terms of its corre-
sponding complex envelope as g̃(t) as follows:

g(t) = Re[g+ (t)]

Using (1.14) we get

g(t) = Re g̃(t)ej2πfc t
 
(1.16)

In general, g̃(t) is a complex-valued quantity, we emphasize this property by express-

ing it in the Cartesian form

g̃(t) = gI (t) + j gQ (t) (1.17)

Where gI (t) and gQ (t) are both real valued low pass functions. On substituting (1.17)
in (1.16) we have,

g(t) = Re [{gI (t) + jgQ (t)} {cos 2πfc t + j sin 2πfc t}]

= Re [{gI (t) cos 2πfc t − gI (t) sin 2πfc t} + j {gI (t) sin 2πfc t + gQ (t) cos 2πfc t}]

Therefore we can write,

g(t) = gI (t) cos(2πfc t) − gQ (t) sin(2πfc t) (1.18)

We refer to gI (t) as the in-phase component of the band-pass signal g(t) and refer to
gQ (t) as the quadrature-phase component of the signal g(t).

Dept of E&C, JNNCE Digital Communication 6

 Chapter 1. Low Pass Equivalent of Bandpass

On multiplying (1.18) by cos(2πfc t) we get

g(t) cos(2πfc t) = gI (t) cos2 2πfc t − gQ (t) cos 2πfc t sin 2πfc t
1 1 1
= gI (t) + gI (t) cos (4πfc t) − gQ (t) sin(4πfc t)
2 |2 {z 2 }
High F requency Components

The High frequency components present on the RHS of the above equation are elimi-
nated by passing the product signal g(t) × cos(2πfc t) through an ideal low pass filter
having a bandwidth of W. Thus at the output of the low pass filter, the 21 gI (t) appears

Figure 1.4: Scheme for Extraction of gI (t) and gq (t) from g(t)

Similarly on multiplying (1.18) by sin(2πfc t) we get

g(t) sin(2πfc t) = gI (t) cos 2πfc t sin 2πfc t − gQ (t) sin2 2πfc t
1 1 1
= − gQ (t) + gI (t) sin (4πfc t) + gQ (t) cos(4πfc t)
2 |2 {z 2 }
High F requency Components

The High frequency components present on the RHS of the above equation are elim-
inated by passing the product signal g(t) × sin(2πfc t) through an ideal low pass filter
having a bandwidth of W. Thus at the output of the low pass filter, the − 12 gQ (t)
appears.
The above procedure for Extraction of gI (t) and gq (t) from g(t) is implemented
in the Figure 1.4

Dept of E&C, JNNCE Digital Communication 7

 Chapter 1. Low Pass Equivalent of Bandpass

Figure 1.5: Scheme for Reconstruction of g(t) from gI (t) and gq (t)

The above is the implementation diagram for reconstructing the signal g(t) from
gI (t) and gq (t). Multiplication of the low-pass in-phase component gI (t) by cos(2πfc t)
and multiplication of the quadrature component gQ (t) by sin(2πfc t) represent linear
forms of modulation.

Polar Representation of Bandpass Signals [1]

(1.17) represents the cartesian form of expressing the complex envelope g̃(t). Alter-
natively we may express it in the polar form

g̃(t) = a(t)ejφ(t) (1.19)

where a(t) and φ(t) are both real valued low pass functions. From (1.16) we have,

g(t) = Re g̃(t)ej2πfc t
 

= Re a(t)ejφ(t) ej2πfc t
 

Thus we have,
g(t) = a(t) cos [2πfc t + φ(t)] (1.20)

We refer a(t) as the natural envelope or envelope of the signal g(t) and φ(t) as the
phase of the signal.

Dept of E&C, JNNCE Digital Communication 8

 Chapter 1. Low Pass Equivalent of Bandpass

From (1.19) ,we can write,

a(t) = |g̃(t)|

Thus the envelope a(t) is the magnitude of g̃(t) and is always a real valued low pass
signal. Also it can be shown that,

q
a(t) = gI (t)2 + gQ (t)2 and

 
−1 gQ (t)
φ(t) = tan
gI (t)
On the other hand we may write,

gI (t) = a(t) cos 2πfc t and

gQ (t) = a(t) sin 2πfc t

1.5 Complex Low Pass Representation of Band-

pass Systems[1]
Consider a bandpass signal x(t), with its Fourier Transform denoted by X(f ) with
the spectrum fc − W ≤ fc + W such that, fc > W . The signal x(t) is represented in
terms of in-phase and quadrature components as follows:

x(t) = xI (t) cos(2πfc t) − xQ (t) sin(2πfc t) (1.21)

Then using x̃(t) to denote the complex envelope of x(t), we may write

x̃(t) = xI (t) + j xQ (t) (1.22)

Let the signal be applied to an LTI band pass system with impulse response h(t) and
transfer function H(f ). Let the bandwidth of the system be limited to fc −B ≤ fc +B.
The impulse response h(t) is represented in terms of in-phase and quadrature

Dept of E&C, JNNCE Digital Communication 9

 Chapter 1. Low Pass Equivalent of Bandpass

components as follows:

h(t) = hI (t) cos(2πfc t) − hQ (t) sin(2πfc t) (1.23)

Then using h̃(t) to denote the complex envelope of h(t), we may write

h̃(t) = hI (t) + j hQ (t) (1.24)

By referring to (1.16) we can write h(t) as

h i
h(t) = Re h̃(t)ej2πfc t (1.25)

The above Equation can be modified as follows:

h̃(t)ej2πfc t + h˜∗ (t)e−j2πfc t

h(t) =
2
2 × h(t) = h̃(t)e j2πfc t
+ h˜∗ (t)e−j2πfc t (1.26)

where h˜∗ (t) is the complex conjugate of h̃(t). We have from the Properties of the
Fourier Transforms,
x(t) *
) X(f ) x∗ (t) *
) X ∗ (−f )

⇒ x(t)ej2πfc t *
) X(f − fc ) x∗ (t)e−j2πfc t *
) X ∗ (−f − fc )

Thus on applying FT on both sides of (1.26), we have

2H(f ) = H̃(f − fc ) + H̃ ∗ (−f − fc ) (1.27)

For the condition, f > 0, the second term in (1.27), H̃ ∗ (−f − fc ) = 0. Thus we have,

2H(f ) = H̃(f − fc ) ,f > 0 (1.28)

The (1.28) is explained in Figure 1.6. Thus for a specifed bandpass transfer function
H(f ), we may determine the H̃(f ) by taking that part of H(f ) corresponding to
positive frequencies, shifting it to the origin, and scaling it by the factor 2.

Dept of E&C, JNNCE Digital Communication 10

 Chapter 1. Low Pass Equivalent of Bandpass

Figure 1.6: Illustration of 2H(f ) = H̃(f − fc ) ,f > 0

Having determined the H̃(f ) , we decompose it into its in phase and quadrature
components, as shown by

H̃(f ) = H̃I (f ) + j H̃Q (f ) (1.29)

where the in phase component is defined by

H̃I (f ) = H̃(f ) + j H̃ ∗ (−f )

and the quadrature component is given by

1 h i
H̃Q (f ) = H̃(f ) − j H̃ ∗ (−f )
2j

To determine the complex impulse response h̃(t), we take the inverse Fourier Trans-
form of H̃(f ) obtaining Z −∞
h̃(t) = H̃(f )e(j2πfc t) df (1.30)
∞

Dept of E&C, JNNCE Digital Communication 11

 Chapter 1. Low Pass Equivalent of Bandpass

1.6 Complex Low Pass Representation of Band-

pass Signals and Systems[2]
Let us derive an analytically efficient method for determining the output of a band-
pass system driven by a corresponding band-pass signal. Let us assume that X(f )
denote the spectrum of the input signal x(t) and H(f ) denote the frequency response
of the band pass system and both are centered around fc .
Let y(t) denote the output signal of the band-pass system produced in response
to the incoming band-pass signal x(t). y(t)is also a band-pass signal, so we may
represent it in terms of its own low-pass complex envelope as

y(t) = Re ỹ(t)ej2πfc t
 
(1.31)

The output signal y(t) is related to the input signal x(t) and impulse response h(t)
of the system by the convolution integral
Z −∞
y(t) = h(τ )x(t − τ )dτ (1.32)
∞

In term of Pre envelopes, we can write

Z∞
y(t) = Re [h+ (τ )] Re [x+ (t − τ )] dτ (1.33)
−∞

(1.33) can be written as

Z∞
 
1
y(t) = Re  [x+ (τ )] [h+ (t − τ )] dτ
2
−∞
 ∞ 
Z
1
= Re  x̃(τ )ej2πfc τ h̃(t − τ )ej2πfc (t−τ ) dτ 
2
−∞
Z∞
 
1
= Re ej2πfc t x̃(τ )h̃(t − τ )dτ  (1.34)
2
−∞

Dept of E&C, JNNCE Digital Communication 12

 Chapter 1. Low Pass Equivalent of Bandpass

On comparing the RHS of (1.31) and (1.34) we have,

Z∞
2y(t) = x̃(τ )h̃(t − τ )dτ (1.35)
−∞

or using the shorthand notation for convolution we have,

2ỹ(t) = x̃(t) ∗ h̃(t) (1.36)

(1.36) states that, the complex envelope ỹ(t) of the output signal of the system is
obtained by convolving the complex low pass impulse response h̃(t) of the system
with the complex envelope x̃(t) of the input band pass signal.

Figure 1.7: Bandpass system and its low pass equivalent

Thus the analysis of a bandpass system can be simplified by replacing it with

an equivalent low pass system. The desired output y(t) is finally derived from the
complex envelope ỹ(t) by using the relation

y(t) = Re ỹ(t)ej2πfc t
 
(1.37)

Substituting the (1.22) and (1.24) in (1.36)

2ỹ(t) = [xI (t) + j xQ (t)] ∗ [hI (t) + j hQ (t)] (1.38)

Since convolution is distributive, we may rewrite (1.38) as

2ỹ(t) = [xI (t) ∗ hI (t) − hQ (t) ∗ xQ (t)] + j [xI (t) ∗ hQ (t) + hI (t) ∗ xQ (t)] (1.39)

Let the complex envelope ỹ(t) of the response be defined in terms of its in phase and
quadrature components as
ỹ(t) = yI (t) + j yQ (t) (1.40)

Dept of E&C, JNNCE Digital Communication 13

 Chapter 1. Low Pass Equivalent of Bandpass

We therefore have the in phase component yI (t) as follows:

2yI (t) = xI (t) ∗ hI (t) − hQ (t) ∗ xQ (t) (1.41)

and the quadrature component yQ (t) as follows:

2yQ (t) = xI (t) ∗ hQ (t) + hI (t) ∗ xQ (t) (1.42)

1.7 Line Coding[2]

The binary data can not be transmitted directly instead they should be converted
into electrical waveforms. This process is called as line coding. The following are
some of the different forms of line coding

Unipolar NRZ Signaling

In this line code, symbol 1 is represented by transmitting a pulse of amplitude A for

the duration of the symbol, and symbol 0 is represented by switching off the pulse as
shown in the Figure 1.8:

Figure 1.8: Unipolar NRZ Figure 1.9: PSD of Unipolar NRZ

The Power spectral density (power versus frequency )is shown in Figure 1.9. Dis-
advantages of this type of signaling are the waste of power due to the transmitted
DC level and the fact that the power spectrum of the transmitted signal does not
approach zero at zero frequency.

Dept of E&C, JNNCE Digital Communication 14

 Chapter 1. Low Pass Equivalent of Bandpass

Polar NRZ Signaling

In this second line code, symbols 1 and 0 are represented by transmitting pulses of
amplitudes +A and A, respectively, as illustrated in Figure 1.10

Figure 1.10: Unipolar NRZ Figure 1.11: PSD of Unipolar NRZ

The Power spectral density is shown in Figure 1.11.The polar NRZ line code is
relatively easy to generate, but its disadvantage is that the power spectrum of the
signal is large near zero frequency.

Unipolar RZ Signaling

In this third line code, symbol 1 is represented by a rectangular pulse of amplitude

A and half-symbol width and symbol 0 is represented by transmitting no pulse, as
illustrated in Figure 1.16

Figure 1.12: Unipolar RZ Figure 1.13: PSD of Unipolar RZ

The Power spectral density is shown in Figure 1.17. However, its disadvantage is

Dept of E&C, JNNCE Digital Communication 15

 Chapter 1. Low Pass Equivalent of Bandpass

that it requires 3 dB more power than polar RZ signaling for the same probability of
symbol error.

Bipolar RZ Signaling

This line code uses three amplitude levels, as indicated in Figure 6.24(d). Specifically,
positive and negative pulses of equal amplitude (i.e., +A and -A) are used alternately
for symbol 1, with each pulse having a half-symbol width; no pulse is always used for
symbol 0.

Figure 1.14: Bipolar RZ Figure 1.15: PSD of Bipolar RZ

A useful property of the bipolar RZ signaling is that the power spectrum of the
transmitted signal has no DC component and relatively insignificant low-frequency
components for the case when symbols 1 and 0 occur with equal probability. The
bipolar RZ line code is also called alternate mark inversion (AMI) signaling

Manchester Signaling

In this final method of signaling, illustrated in Figure 6.24e, symbol 1 is represented

by a positive pulse of amplitude A followed by a negative pulse of amplitude -A, with
both pulses being half-symbol wide. For symbol 0, the polarities of these two pulses
are reversed.
A unique property of the Manchester code is that it suppresses the DC component
and has relatively insignificant low-frequency components, regardless of the signal
statistics. This property is essential in some applications

Dept of E&C, JNNCE Digital Communication 16

 Chapter 1. Low Pass Equivalent of Bandpass

Figure 1.17: PSD of Manchester Cod-

Figure 1.16: Manchester Coding ing

Numerical Problems

1. Find the Hilbert Transform of g(t) = cos2πfc t

Since Hilbert transform introduced a phase shift of -90 for f > 0 we can write,

ˆ = cos 2πfc t − π
h i
g(t)
hπ 2i
= cos − 2πfc t
2
= sin2πfc t

2. Find the Hilbert Transform of g(t) = cos2πfc t + sin2πfc t

ˆ
h πi h πi
g(t) = cos 2πfc t − + sin 2πfc t −
hπ 2i hπ 2i
= cos − 2πfc t − sin − 2πfc t
2 2
= sin2πfc t − cos2πfc t

3. Find the Hilbert Transform of g(t) = e−j2πfc t

g(t) = cos (2πfc t) − j sin (2πfc t)

ˆ = cos 2πfc t − π − j sin 2πfc t − π
h i h i
g(t)
hπ 2i hπ 2i
= cos − 2πfc t + j sin − 2πfc t
2 2
= sin2πfc t + jcos2πfc t

= ej2πfc t

Dept of E&C, JNNCE Digital Communication 17

 Chapter 1. Low Pass Equivalent of Bandpass

1 |t| ≤
 T
2
4. Find the Hilbert Transform of g(t) =
0 elsewhere


ˆ = g(t) ∗ 1
g(t)
πt

Z +T /2
1
= 1× dτ
−T /2 π (t − τ )

Z +T /2
dτ
=
−T /2 π (t − τ )

+T /2
−1
Z
dτ
=
π −T /2 (τ − t)

−1 +T /2
= [ln(τ − t)]−T /2
π

−1
= [ln(T /2 − t) − ln(−T /2 − t)]
π

 
1 t + T /2
= ln
π t − T /2

Dept of E&C, JNNCE Digital Communication 18

 References

[1] S. Haykin, Communication systems. John Wiley & Sons, 2008.

[2] ——, Digital Communication Systems. John Wiley & Sons, 2015.

19