ELEC3400 Lecture Notes

José De Doná
June 6, 2015

Abstract
These notes are based on the excellent classic textbook by B. P. Lathi: Signal Processing & Linear Sys-
tems, Oxford University Press 1998 (or the International Edition 2009/2010). There is no claim to originality
nor should they be viewed as a substitute to reading the textbook. On the contrary, they are intended as a brief
summary and a study-guide to be used in conjunction with reading the main text. (These notes may be regarded
as the analogue of taking painkillers, while reading the textbook would be the equivalent to the real solution to
the problem—undergoing root canal therapy.)

Introductory circuit 4,5'16,-(#$%&/7'1(

We will begin our presentation with a basic circuit that

should be familiar to you from your previous Electrical
Circuits courses. Consider the circuit shown in Fig-
ure 1. The two basic rules to understand this circuit
theoretically are the ideal properties of the Operational 3 2 3
Amplifier (op-amp); namely: 3
1. Infinite open-loop gain; which implies that in
this circuit the differential input voltage of the
op-amp is zero. 2 2
2. Infinite input impedance; which implies that the Figure 1: Inverting amplifier.
input current to the op-amp is zero.
Using these two basic rules, we can derive the following equations. Since the current entering the inverting input
(–) of the op-amp is zero (Rule 2 above), then Kirchhoff’s current law gives:
i1 (t) = i2 (t). (1)
vi (t) − v− (t) v− (t) − vo (t)
Ohm’s law applied to both resistors gives i1 (t) = and i2 (t) = respectively. By Rule 1
Ri Rf
above we have v− (t) = v+ (t).
Since we can take the conductor with negative polarity in the
circuit of Figure 1 as the reference voltage, we can assume it to
be zero volts, and hence v− (t) = v+ (t) = 0, which implies
vi (t) vo (t)
i1 (t) = , i2 (t) = − . (2)
Ri Rf
Substituting i1 (t) and i2 (t) from Equations (2) into Equation (1)
and rearranging we obtain:
vo (t) Rf
=− . (3)
vi (t) Ri
If the input voltage vi (t) is, for example, a unit step function as
shown in the top part of Figure 2, then Equation (3) tells us that
Figure 2: Step response. the output voltage is given by the bottom plot of Figure 2.

1
Example of analogue filter
One of the most fundamental topics that are !"#$%&'()*(+,#&)-(./&0'1(
covered in this Signal Processing course is Filter Design.
In order to obtain a useful filter we will now
make a modification to the circuit of Fig-
ure 1 by introducing an extra element, a ca-
pacitor C, as shown in Figure 3. Rule 2 3 2
above and Kirchhoff’s current law give:
i1 (t) = i2 (t) + i3 (t). (4)
Recalling that Rule 1 above implies that
v− (t) = v+ (t) = 0, we then have that the 3 2 3
voltage across the capacitor C is vc (t) = 3
v− (t) − vo (t) = −vo (t), and hence the cur-
rent through the capacitor is:
2 2
dvc (t) dvo (t) Figure 3: Analog filter.
i3 (t) = C = −C . (5)
dt dt
Since Equations (2) still apply, substituting i1 (t) and i2 (t) from those equations and i3 (t) from Equation (5) into
Equation (4) and rearranging we obtain:

dvo (t) 1 1
+ vo (t) = − vi (t). (6)
dt CRf CRi

We conclude that the differential equation (6) describes the re-

lationship between the input voltage vi (t) and the output volt-
age vo (t) in the circuit of Figure 3. Before going ahead with
the mathematical treatment of this differential equation, let’s do
some intuitive analysis. Suppose the circuit is initially at rest
with voltages vi = 0 and vo = 0; then, at some time instant—
that without loss of generality we call t = 0—we apply a con-
stant input voltage vi (t) = 1 volts for t ≥ 0 (i.e., vi is a unit
step function). So, in (6) we have that at time t = 0+ (i.e., an
infinitesimal instant after t = 0), vo (0+ ) = 0 (since the voltage
!"#$%&''
across the capacitor cannot change instantaneously), vi (0+ ) = 1,
(&)*+,'' dvo (0+ ) 1
and then =− . So, vo (t) will start to decrease from
Figure 4: Step response. dt CRi
vo (0) = 0 with a negative slope −1/CRi (see Figure 4). As soon
as vo (t) becomes negative we have:
dvo (t) 1 1 1
=− − vo (t) > − ,
dt CRi CRf CRi
| {z }
positive

1
and we can see that the slope starts to ‘back off’ from the initial value − and becomes smaller in magnitude.
CRi
This situation continues until vo (t) reaches a value such that:
1 1
− − vo = 0,
CRi CRf

Rf dvo (t)
that is, vo = − , when = 0, and the circuit attains a new equilibrium (see Figure 4). Comparing
Ri dt
Figure 4 with Figure 2 we can see that the introduction of the capacitor C makes the circuit function as a lowpass

2
filter, since it ‘smoothes out’ the sharp edge of the step signal (in other words, it opposes sudden changes, i.e.,
high frequencies).
We will now solve the differential equation (6). First apply the integrating factor et/CRf to obtain

et/CRf
 
t/CRf dvo (t) 1
e + vo (t) = − vi (t),
dt CRf CRi

d  et/CRf
vo (t)et/CRf = −

which can be rewritten using the product rule as vi (t). Integrating both sides,
dt CRi
t Z t
d  1
Z
τ /CRf
eτ /CRf vi (τ )dτ.

vo (τ )e dτ = −
0 dτ CRi 0

Solving the left-hand side integral and considering the initial condition vo (0) = 0 we obtain,
Z t
t/CRf 0/CRf t/CRf 1
vo (t)e − vo (0) e = vo (t)e =− eτ /CRf vi (τ )dτ,
| {z } CRi 0
=0

and finally, rearranging,

% M& G "! N
t
1
Z % M& G "! N
vo (t) = − e(τ −t)/CRf vi (τ )dτ. (7) A % ) G ""
CRi 0 %
G & ) ) @ % & ""
%
G & )
2
A % ) G "" 2
Equation (7) is a convolution integral, which is a > <=
> 80,
fundamental topic to be studied soon. We will re-
visit this equation in coming sections.
Operational amplifier circuits like the one we
have just analysed are ubiquitous in analog filter Figure 5:Figure Butterworth filter with
12: Second-order a 3 dB-cutoff
Butterworth frequency of
design example.
design. A typical analog filter is illustrated in Fig- 1000 rad/sec (159 Hz). [Source: http://ocw.mit.edu]
ure 5 (Butterworth filters will be studied later).generate the following filter transfer function
122828246505000
H(s) =
(s2 + 468.4s + 429300)(s2 + 178.9s + 988300)(s + 289.5)
The unit impulse function = 2
429300
× 2
988300
×
289.5
s + 468.4s + 429300 s + 178.9s + 988300 s + 289.5
We will now start reviewing and developing more fundamental
We implement the filtertopics that will
as two cascade allow ussections
second-order analyse andwith
(each predict
a gain the
of Klp = 0.
as above, and a single first-order non-inverting section with a gain
behaviour of equations like (6) and (7) that describe the models of dynamical systems (and, in particular, filters). of 4. We will use the tw
op-amp circuit
We start by formally defining the Let all capacitors have a value of 0.47 µF.
unit impulse (Dirac delta) function
(1) For the first section a = 468.4, b = 429300:
δ(t) as:
1 1
R2 = = = 4, 542 Ω,
aC2 468.4 × 0.47 × 10−6
δ(t) = 0 for t 6= 0, 2a 2 × 468.4
Z ∞
(8) R1 = R3 = = = 4, 642 Ω
bC1 429, 300 × 0.47 × 10−6
δ(t)dt = 1.
−∞ (2) For the second section a = 178.9, b = 988, 300:

We can immediately realise that the Figure 6: Impulse 1as limit of a1gate function.
R2 = = = 11, 893 Ω,
aC2 178.9 × 0.47 × 10−6
function δ(t) has unusual features. 2a 2 × 178.9
= R3an
R1 have = area = = 770 Ω
For example, how can a function that is zero almost everywhere bC1 equal988, 300to !!!!!"#$%&'!!!!!
one?
× 0.47 × 10We−6 can imagine

this function as the limit of a gate function, as shown in Figure 6. Note that the area underneath each of the
functions in the limiting process of Figure 6 is(2) For the first-order section K = 4 Ωc = 289.5:
always equal to 1 and hence the limit has area 1 (and, in particular,
at t = 0 it shoots to infinity!). Strictly speaking, δ(t) is not a function 1
R1 = but something
=
1known as a ‘generalised
= 7, 349 Ω,
function’, which is rather defined by its effect on other functions instead of by its value at ×every
Ω c C 289.5 × 0.47 10−6 instant of time.
Let R2 = 1, 500 Ω,
In particular, the effect on other functions is characterised by the following property.
R = (K − 1)R = 3 × 1, 500 = 4, 500 Ω
3 2

!!!!!"#$%&'!!!!!
and the design is complete. The final circuit is shown in Fig. 13.
3
10
 Sampling property of the unit impulse function
Let f (t) be a function and consider the following integral:
Z ∞ Z ∞ Z ∞
f (τ )δ(τ )dτ = f (0) δ(τ ) dτ = f (0) δ(τ )dτ = f (0),
−∞ −∞ | {z } −∞
since δ(τ )=0 for τ 6=0 | {z }
=1

and we thus conclude that Z ∞

f (τ )δ(τ )dτ = f (0). (9)
−∞

So, the area under the product f (t)δ(t) equals the value of the function f (t) at the instant when the impulse is
located.
Consider now a shifted version of the impulse, δ(t − τ ) (here we
think of δ(t − τ ) as a function of τ shifted by t). Notice first, using (8)
and making the change of variables t − τ = x, dτ = −dx, that:
Z ∞ Z −∞ Z ∞
δ(t − τ )dτ = δ(x)(−dx) = δ(x)dx = 1.
−∞ ∞ −∞

Applying (8) we also have:

δ(t − τ ) = 0 for t − τ 6= 0, i.e., for τ 6= t,

Figure 7: Sampling property.
so, the impulse δ(t − τ ) shoots to infinity at τ = t (see Figure 7).
Consider the following integral:
Z ∞ Z ∞ Z ∞
f (τ )δ(t − τ )dτ = f (t) δ(t − τ ) dτ = f (t) δ(t − τ )dτ = f (t).
−∞ −∞ | {z } −∞
since δ(t−τ )=0 for τ 6=t | {z }
=1

So, we have the sampling property of the impulse function:

t Z ∞
oximated f (τ )δ(t − τ )dτ = f (t). (10)
sum of −∞
angular
es The integral (10) also has a very interesting interpretation.
We can approximate the function f (t) as a sum of rectangu-
lar pulses of width ∆τ as shown in Figure 8. Then, in the
Figure 8: Approximation by rectangular pulses.
ulse limit as ∆τ approaches zero we have:
onse
∞
X Z ∞
lim f (n∆τ )δ(t − n∆τ )∆τ = f (τ )δ(t − τ )dτ = f (t). (11)
yed ∆τ →0 −∞
n=−∞
ulse
onse
The impulse response h(t) of a filter
h pulse in
In the previous section we reached the conclusion that we can decompose a
has strength
function
a) equal to f (t) into individual components given by impulses at every time t !!!!!"#$%&'!!!!!
(as expressed by (11) and interpreted graphically in Figure 8). If we now
apply the signal f (t) to the input of a linear filter (or any other linear system), Figure 9: Linear filter.
as shown in Figure 9, we would expect, by the very fundamental property of
a linear system, the output of the filter y(t) to be the superposition of the responses to each individual impulse
l response
btainedcomponent.
by We will soon see that this is indeed the case.
ming all the !!!!!"#$%&'!!!!!
ponents 4
erposition)
 But, first, what is the response of a linear system to an impulse?
When we apply an impulse to a system it is like applying a sud-
den surge of current to a capacitor in a circuit. The capacitor
!!!!!"#$%&'!!!!! will charge suddenly and then it discharges through the circuit
following a transient. Exactly the same thing happens in a linear
Figure 10: Impulse response h(t). system and you can imagine the impulse response as a combi-
nation of decaying exponentials (as in the case of the capacitor
discharging in the circuit). Thus, denoting by h(t) the impulse response, we have a situation as represented in
Figure 10. The impulse response can be obtained from the differential equation that defines the filter (or any
other linear system), and this can be read in Section 2.3 of Lathi’s textbook. However, a much simpler method is
afforded by the Laplace transform and thus we will defer the treatment of this particular topic. For now, we just
assume that the impulse response h(t) of a filter is given to us, e.g., as shown in Figure 10, or that we can easily
find it by using Laplace transform methods.

System response to an external input

We shall consider that the initial conditions of our filter are equal to zero (if they are not zero, their contribution
to the total response will just be a decaying transient). We will use the principle of superposition (satisfied by
all linear systems). We first express the input f (t) in terms of impulses. Figure 8 above shows f (t) as a sum of
rectangular pulses of width ∆τ . We will use this ‘decomposition’ to obtain the overall response to f (t). This is
done graphically in Figure11.

Input
approximated
as a sum of
rectangular
pulses

Impulse
response

Delayed
impulse
response

Each pulse in
has strength
(area) equal to

Total response
is obtained by
summing all the
components
(superposition)

Figure 11: System response to input f (t).

As shown in Figure11, the output is obtained as the sum (superposition) of the responses to the individual

5
components:
∞
X Z ∞
y(t) = lim f (n∆τ )h(t − n∆τ )∆τ = f (τ )h(t − τ )dτ,
∆τ →0 −∞
n=−∞

and we have obtained a fundamental relationship,

Z ∞
y(t) = f (τ )h(t − τ )dτ, (12)
−∞

that tells us that knowing the impulse response h(t) of a linear system we can determine the response y(t) to any
input f (t) (that is, provided we can solve Equation (12)).

The convolution integral

Equation (12) above is the convolution of f (t) with h(t) and it is an all-important relationship in signals and
systems. Formally, the convolution of two functions f1 (t) and f2 (t) is denoted f1 (t) ∗ f2 (t) and is given by:
Z ∞
f1 (t) ∗ f2 (t) = f1 (τ )f2 (t − τ )dτ.
−∞

Properties of convolution
1. Commutative: f1 (t) ∗ f2 (t) = f2 (t) ∗ f1 (t).
2. Distributive: f1 (t) ∗ [f2 (t) + f3 (t)] = f1 (t) ∗ f2 (t) + f1 (t) ∗ f3 (t).
3. Associative: f1 (t) ∗ [f2 (t) ∗ f3 (t)] = [f1 (t) ∗ f2 (t)] ∗ f3 (t).
4. Shift property: if f1 (t) ∗ f2 (t) = c(t), then f1R(t) ∗ f2 (t − T ) = c(t − T ).
∞
5. Convolution with an impulse: f (t) ∗ δ(t) = −∞ f (τ )δ(t − τ )dτ = f (t) (see Equation (10) above).
6. Width property: if the durations (widths) of f1 (t) and f2 (t) are T1 and T2 respectively, then the duration
(width) of f1 (t) ∗ f2 (t) is T1 + T2 (see Figure12).

Figure 12: Width property of convolution.

Causality
In general, we will consider inputs to a system that start at t = 0 and that are zero before that:

f (t) = 0 for t < 0.

If the system is causal, its response to an impulse δ(t) (which is located at t = 0) cannot begin before t = 0.
Therefore, for a causal system:
h(t) = 0 for t < 0.
Consider again the output of a filter given in (12) above. We have:

f (τ ) = 0 for τ < 0,

and,
h(t − τ ) = 0 for t − τ < 0, i.e., for τ > t.

6
Hence, the product f (τ )h(t − τ ) is 0 for τ < 0 and τ > t, and we can write the output equation (12) of a causal
filter as: Z t
 f (τ )h(t − τ )dτ for t ≥ 0,
y(t) = f (t) ∗ h(t) =
 0
0 for t < 0.
This result shows that if the impulse response h(t) is causal, then the filter response is also causal.

Graphical understanding of convolution

It is important to have a good grasp of convolution in a graphical way. Consider:
Z ∞
y(t) = f (t) ∗ h(t) = f (τ )h(t − τ )dτ.
−∞

One critical point is to realise that this integral is performed with respect to τ , so that t is a parameter (like a
constant). We plot f (τ ) (that is easy, just substitute t → τ in f (t)). Next we flip h(τ ) to obtain h(−τ ) and then
we shift it (slide) by t time units: h(t − τ ), and evaluate the area under the product of f (τ )h(t − τ ) to obtain
the value of y(t). Then we repeat the process over and over again for different values of t. This process will be
illustrated in the following section for the response of the analog filter shown in Figure 3.

Example of graphical convolution

Earlier on we obtained that the response of the differential equation (6) when the initial condition is vo (0) = 0 is
given by the integral (7), repeated here for convenience:
t
1
Z
vo (t) = − e(τ −t)/CRf vi (τ )dτ. (7)
CRi 0

Equation (7) is an example of the convolution equation (12) with

 −t/CR
 e f
(
vi (t) for t ≥ 0, − for t ≥ 0,
f (t) = h(t) = CRi
0 for t < 0, 
0 for t < 0,

as it is straightforward to verify that

f (t) ∗ h(t) equals (7). Let us now
see how convolution works graphi-
cally for the case we analysed intu-
itively before when vi (t) is given by
a unit step function (recall Figure 4).
In other words, we are going to con-
volve the two functions illustrated in Figure 13: f (t) and h(t).
Figure 13. We can use the commuta-
tive property of convolution (Property 1 above) to flip and slide either f (t) or h(t). We will do it with f (t) as it is
simpler. We can thus evaluate: Z t
vo (t) = h(τ )f (t − τ )dτ.
0
The procedure is illustrated in Figure 14. We start (top graphs) by plotting h and f as functions of τ . Then we
flip f (τ ) and slide (shift) it by t for different values of t (left-hand side plots). On the right-hand side we show
the product h(τ )f (t − τ ) (and the areas being evaluated) for each value of t considered. The result of the area
under h(τ )f (t − τ ), evaluated at each value of t, is shown in the bottom plot, which gives us the output voltage
vo (t) of the filter (compare with Figure 4).

7
 Figure 14: Graphical convolution.

We can see, in particular, from the previous graphical analysis that the final value of vo (t) is the total area
under h(τ ) from τ = 0 to τ = ∞. Indeed,
Z ∞
e−τ /CRf CRf  −τ /CRf τ =∞  Rf
vo, final = − dτ = e τ =0
=− ,
0 CRi CRi Ri
and we conclude that the DC gain of a linear filter (or any other linear system) is the area underneath its impulse
response.

The Fourier transform

Formally, the Fourier transform of a signal f (t) is given by:
Z ∞
F (ω) = f (t)e−jωt dt, (13)
−∞

and the inverse Fourier transform of F (ω) is given by:

∞
1
Z
f (t) = F (ω)ejωt dω. (14)
2π −∞

We say that f (t) and F (ω) are a Fourier transform pair, and write it as f (t) ⇐⇒ F (ω) or, equivalently, F (ω) =
F[f (t)] and f (t) = F−1 [F (ω)].
The Fourier representation is a way of expressing a signal in terms of everlasting sinusoidals (or exponentials).
The Fourier spectrum of a signal indicates the relative amplitudes and phases of the sinusoids that are required to
synthesise that signal. We can see from (14) that the contribution of the exponential function ejωt within a band

8
 1
dω is equal to F (ω)dω = F (ω)dF, where dF is the bandwidth in hertz. (Recall that ω = 2πF, where ω is
2π
angular frequency in rad/sec and F is frequency in hertz = cycles/sec.) Clearly, F (ω) is the ‘spectral density’
per unit bandwidth (in hertz). F (ω) is commonly
referred to as the spectrum (of Fourier spectrum) of f (t).
Using Euler’s formula ejθ = cos θ + j sin θ in (13) we obtain:
Z ∞ Z ∞ Z ∞
F (ω) = f (t) (cos ωt − j sin ωt) dt = f (t) cos ωt dt − j f (t) sin ωt dt,
−∞ −∞ −∞


and hence we can see that, in general, F (ω) is a complex function, and we have amplitude |F (ω)| and phase


F (ω) spectra:
F (ω) = |F (ω)|ej F (ω) . (15)

Conjugate symmetry property

From (13) we have, for negative frequencies:
Z ∞ Z ∞ Z ∞ Z ∞
jωt
F (−ω) = f (t)e dt = f (t) (cos ωt + j sin ωt) dt = f (t) cos ωt dt + j f (t) sin ωt dt.
−∞ −∞ −∞ −∞

If f (t) is a real function we can see, by comparing the expressions for F (ω) and F (−ω) above, that:

|F (−ω)| = |F (ω)|,
(16)
F (−ω) = − F (ω),

so that the amplitude is an even function and the phase is an odd function of ω. In other words, the spectrum for
negative frequencies, F (−ω), is the complex conjugate of the spectrum for positive frequencies, F (ω); denoted
F (−ω) = [F (ω)]∗ . For this reason, this is known as the conjugate symmetry property of the Fourier transform
and is valid for any real function f (t).

Example 1
u(t)
Consider the function f (t) = e−at u(t) shown on the left-hand side of Fig-
ure 16 (below, where u(t) is the unit step function shown in Figure 15,
1 for t ≥ 0,
u(t) =
0 for t < 0.
!
Direct computation from (13) yields: Figure 15: Unit step function.
Z ∞ Z 0 Z ∞
F (ω) = e−at u(t)e−jωt dt = e−at × 0 × e−jωt dt + e−at × 1 × e−jωt dt
Z−∞∞ Z ∞ −∞ 0
1  ∞ 
= e−at e−jωt dt = e−(a+jω)t dt = − e−(a+jω)t 0 .
0 0 a + jω
Notice that lim e−(a+jω)t = lim e−at −jωt
| e {z } = 0 if a > 0.
t→∞ t→∞
always magnitude = 1
Therefore we have:
1
F (ω) = , for a > 0.
a + jω
Computing the amplitude and phase of F (ω):
1 −1 ω
 
|F (ω)| = √ , F (jω) = − tan ,
a2 + ω 2 a
which are shown on the right-hand side of Fig-
Figure 16: Fourier transform pair for Example 1. ure 16.

9
Existence of the Fourier transform
We saw in Example 1 above that the Fourier transform of f (t) = e−at u(t) does not exist if a < 0. Clearly, not all
signals are Fourier transformable. The existence of the Fourier transform is assured if f (t) satisfies the Dirichlet
conditions, the first of which is:
Z ∞ Z ∞
−jωt
|F (ω)| ≤ |f (t)e |dt = |f (t)|dt < ∞ .
−∞ −∞
| {z }
First Dirichlet condition

The second Dirichlet condition is that, in any finite interval, f (t) may have only a finite number of maxima and
minima and a finite number of finite discontinuities. Any function that can be generated in practice (e.g., in the
laboratory) satisfies the Dirichlet conditions and, thus, has a Fourier transform.

Linearity of the Fourier transform

It follows from (13) that, if f1 (t) ⇐⇒ F1 (ω) and f2 (t) ⇐⇒ F2 (ω), then, for any constants a1 and a2 ,

a1 f1 (t) + a2 f2 (t) ⇐⇒ a1 F1 (ω) + a2 F2 (ω).

Linear system response using the Fourier transform

If h(t) is the impulse response of a linear time invariant (LTI) system—e.g., a linear filter—and we apply an
exponential function ejωt to its input, then we have from (12) that the output is:
Z ∞ Z ∞
jωt
y(t) = h(t) ∗ e = h(τ )ejω(t−τ )
dτ = ejωt
h(τ )e−jωτ dτ = ejωt H(ω).
−∞
| −∞ {z }
H(ω)= Fourier transform of h(t)

If we now have a general input function f (t) and we express it as an infinite sum of exponential functions (this is
precisely (14)) we have: Z ∞
1
f (t) = F (ω)ejωt dω.
2π −∞
The output corresponding to f (t) is:
Z ∞ Z ∞  Z ∞ 
1 jω(t−τ )
y(t) = h(t) ∗ f (t) = h(τ )f (t − τ )dτ = h(τ ) F (ω)e dω dτ
−∞ −∞ 2π −∞
Z ∞ Z ∞ 
1 −jωτ
= F (ω) h(τ )e dτ ejωt dω
2π −∞ (17)
| −∞ {z }
H(ω)= Fourier transform of h(t)
∞
1
Z
= F (ω)H(ω)ejωt dω.
2π −∞

So we have an analogous case to that of Figure 11, where we decomposed the input f (t) as a sum of its impulse
components, and then we recomposed the output, using superposition, as an infinite sum of the responses to the
individual impulses (thus obtaining the convolution integral (12)). Here we have:
System response to an individual expo-
ejωt −→ H(ω)ejωt .
nential:
Input signal f (t) formed as an infinite 1
R∞
f (t) = 2π −∞
F (ω)ejωt dω.
sum of exponential components:
The output y(t) is the sum of the re- 1
R∞
f (t) −→ y(t) = 2π −∞
F (ω)H(ω)ejωt dω.
sponses to the exponential components:

10
 Thus, transmission of a signal through a linear system can be viewed as transmission of the various sinusoidal
components of the signal through the system.
Equation (17) also tells us that y(t) is the inverse Fourier transform of the product F (ω)H(ω). That is,
Z ∞ Z ∞
1 jωt 1
y(t) = Y (ω)e dω = F (ω)H(ω) ejωt dω.
2π −∞ 2π −∞ | {z }
Y (ω)

Thus, we conclude that the Fourier transform of y(t) is:

Y (ω) = F (ω)H(ω). (18)

Equation (18) is a fundamental relationship, and says that the Fourier transform of the output of a linear system
is given by the product of the Fourier transform of the input and the Fourier transform of the impulse response of
the system.

Frequency response of a system

Equation (18) tells us, in other words, that the output spectrum is obtained by multiplying the input spectrum by
the spectral response of the system (also called spectral shaping, or modification, of the signal by the system).
Recalling that Fourier transforms are in general complex functions (see (15)), and using elementary properties
of complex numbers, we obtain from (18):

|Y (ω)| = |F (ω)||H(ω)|, (19)

Y (ω) = F (ω) + H(ω). (20)

So, during transmission through a channel, the input signal amplitude spectrum |F (ω)| is changed to
|F (ω)||H(ω)|, and the input signal phase spectrum F (ω) is changed to F (ω) + H(ω).
An input signal spectral component of frequency ω is modified in amplitude by a factor |H(ω)| and is shifted
in phase by an angle H(ω).
The plots of |H(ω)| and H(ω) as functions of ω show at a glance how the system modifies the amplitudes
and phases of various sinusoidal inputs. For this reason, H(ω) (the Fourier transform of the impulse response) is
called the frequency response of the system (a fundamental concept).

Fourier transforms of some useful functions

Gate or pulse function
Consider the pulse function represented in Figure 17, defined as:

   0 if |t| > a/2,
t
f (t) = rect = 1/2 if |t| = a/2,
a 
1 if |t| < a/2. Figure 17: Gate or pulse function.


Using (13) we have:

Z ∞ Z a/2
−jωt
F (ω) = f (t)e dt = e−jωt dt
−∞ −a/2
1 −jωa/2
2 sin (ωa/2) sin (ωa/2)
=− e − ejωa/2 = =a = a sinc(ωa/2).
jω ω (ωa/2)
The spectrum F (ω) is shown in Figure 18; and the amplitude and phase components of the spectrum are shown
in Figure 19. As can be seen, the spectrum is mainly concentrated within the first lobe (from ω = 0 to ω = 2π/a).
So, an estimate of the bandwidth of the pulse function of width a seconds is 2π/a rad/sec, or 1/a in hertz (note
the reciprocal relationship between the pulse’s width and its bandwidth).

11
 Figure 18: Spectrum of pulse function. Figure 19: Amplitude and phase spectra of pulse function.

Unit impulse function δ(t)

By the previous result (Figure 18) and the limit process of Figure 6 we have the situation represented in Figure 20.
So, we would expect the Fourier transform of the unit impulse to be a flat spectrum with amplitude equal 1.

F F F

Figure 20: Fourier transform of the unit impulse as a limit process.

Indeed, the sampling property of the impulse function (see (9)) gives:
Z ∞
F[δ(t)] = δ(t)e−jωt dt = e−jω×0 = 1,
−∞

and we have thus found this Fourier transform pair:

δ(t) ⇐⇒ 1.

Exercise 1
Show that F [2πδ(ω)] = 1; i.e., 1 ⇐⇒ 2πδ(ω).
−1

Exercise 2
Show that F [2πδ(ω − ω0 )] = ejω0 t ; i.e., ejω0 t ⇐⇒ 2πδ(ω − ω0 ).
−1

Exercise 3
Show that cos ω0 t ⇐⇒ π [δ(ω + ω0 ) + δ(ω − ω0 )] .

12
 Unit step function u(t)
The unit step function u(t) was defined in Example 1, on Page 9 above (see
Figure 15). Direct application of the Fourier transform formula (13) yields
an indeterminate result. So, we approach the problem by a limit process.
Consider:
F u(t) lim e−at u(t),
F = a→0 Figure 21: Step function as limit of
(see Figure 21). Using the result obtained in Example 1, Page 9, we have exponentials.
 
1 a ω 1
U (ω) = F [u(t)] = lim F e u(t) = lim
 −at 
= lim 2 2
−j 2 2
= πδ(ω) + ,
a→0 a→0 a + jω a→0 a +ω a +ω jω
and we thus obtain this Fourier transform pair:
1
u(t) ⇐⇒ πδ(ω) + . (21)
jω
In deriving the last limit we have taken into account the fact that the
function a/(a2 + ω 2 ) has very interesting properties. It looks like the
plot on Figure 22 and, as a → 0, it approaches zero for all ω 6= 0.
However, its area is always equal to
Z ∞
a −1 ω 
 ∞
2 2
dω = tan = π.
−∞ a + ω a −∞

a
Figure 22: Limit of the function Hence, as a → 0, the function 2 → πδ(ω).
a/(a2 + ω 2 ) as a → 0. a + ω2

Properties of the Fourier transform

Time-frequency duality
Comparing (13) and (14) on Page 8 we can see that the direct and inverse Fourier transforms are remarkably
similar, the only minor differences are the 2π factor and the opposite signs of the exponents. This similarity gives
a ‘duality’ relationship between the time and frequency representations. One of the consequences is the following
symmetry property.

Symmetry property
If f (t) ⇐⇒ F (ω), then F (t) ⇐⇒ 2πf (−ω).

Example 2
On Page 11 we obtained (see Figure 23):
 
t  ωa 
rect ⇐⇒ a sinc .
a | {z 2 }
| {z }
f (t) F (ω) Figure 23: Pulse function and its transform.
Using the symmetry property we obtain (see Fig-
ure 24):
−ω
   
at ω 
a sinc ⇐⇒ 2π rect = 2π rect ,
2 a a
| {z } | {z }
F (t) 2πf (−ω)

and we have just obtained the impulse response (i.e.,

the inverse Fourier transform of its spectral response,
Figure 24: Ideal lowpass filter spectral response and im-
see Page 11) of an ideal lowpass filter.
pulse response.
13
Scaling property
1 ω 
If f (t) ⇐⇒ F (ω), then f (at) ⇐⇒ F .
|a| a
This property means that time compression by a factor of a causes the spectrum to be expanded by the same
factor a, and vice versa. The effect of scaling can be appreciated in Figures 23 and 24, and also in the limit process
illustrated in Figure 20. It also implies that a signal’s duration and its bandwidth are reciprocal (as already noted
on Page 11).

Time-shifting property
If f (t) ⇐⇒ F (ω), then f (t − t0 ) ⇐⇒ F (ω)e−jωt0 .
This shows that delaying a signal by t0 seconds does not change its amplitude spectrum. The phase spectrum,
however, is changed by −ωt0 .

Frequency-shifting property
If f (t) ⇐⇒ F (ω), then f (t)ejω0 t ⇐⇒ F (ω − ω0 ).
This property allows us, for example, to
understand the phenomenon of amplitude
modulation. Multiplication of a signal f (t)
by a sinusoid cos ω0 t amounts to modulat-
ing the sinusoid’s amplitude. The sinusoid
cos ω0 t is called the carrier and the signal
f (t) is the modulating signal. Since,

1 jω0 t
+ e−jω0 t ,


cos ω0 t = e
2
we obtain: Figure 25: Amplitude modulation.
1
f (t) cos ω0 t ⇐⇒ [F (ω − ω0 ) + F (ω + ω0 )] .
2
So, multiplication of a signal by a sinusoid of frequency ω0 shifts the spectrum F (ω) by ±ω0 (see Figure 25).

Convolution
Time convolution: f1 (t) ∗ f2 (t) ⇐⇒ F1 (ω)F2 (ω).
If f1 (t) ⇐⇒ F1 (ω) and
f2 (t) ⇐⇒ F2 (ω), then we have: 1
Frequency convolution: f1 (t)f2 (t) ⇐⇒ F1 (ω) ∗ F2 (ω).
2π

Time differentiation and time integration

If f (t) ⇐⇒ F (ω), then:

Time differentiation: Time integration:

t
df (t) F (ω)
Z
⇐⇒ jωF (ω), f (τ )dτ ⇐⇒ + πF (0)δ(ω).
dt −∞ jω
2
d f (t)
⇐⇒ (jω)2 F (ω),
dt2
..
.
dn f (t)
⇐⇒ (jω)n F (ω).
dtn

14
 Notice that we can obtain the Fourier transform of the unit step function u(t) using the time integration
property. Indeed,
f (t) = δ(t) ⇐⇒ F (ω) = 1,
and,
t
1
Z
u(t) = δ(τ )dτ ⇐⇒ + πδ(ω).
−∞ jω
(Compare this last result with (21).)

Ideal and practical filters

Let us consider an ideal lowpass filter (LPF) with bandwidth W . We can obtain the inverse Fourier transform of
H(ω) from the result of Example 2 on Page 13:

W  ω 
h(t) = sinc (W t) ⇐⇒ H(ω) = rect . (22)
π 2W
We can see from Figure 26 that h(t) is non
causal, since h(t) 6= 0 for t < 0 (see the def-
inition of a causal system on Page 6). How can
we obtain a filter with lowpass characteristics but
that can be realised with a causal system?
To tackle this question let us first analyse the
conditions for a distortionless transmission. As Figure 26: Ideal lowpass filter.
can be seen from the frequency response of a sys-
tem in (19)–(20), during transmission through a system some frequency components may be boosted in amplitude
while others are attenuated. The relative phases of the various components also change. In general, the output
waveform will be different from the input waveform. Transmission through a channel is considered distortionless
if the input and the output have identical wave shapes within a multiplicative constant and with the output possibly
(a) Ideal lowpass filter (b) Ideal highpass filter
delayed with respect to the input (i.e., a time-delay is tolerated). Thus, for distortionless transmission, the output
y(t) of a transmission channel, corresponding to an input f (t), must be:

y(t) = kf (t − td ). (23)

From the time-shifting property (see Page 14) we have:

(b) Ideal bandpass filter
Y (ω) = kF (ω)e−jωtd = F (ω)H(ω) . (24)
| {z }
from (18)

Hence:
|H(ω)| = k,
H(ω) = ke−jωtd ⇒
H(ω) = −ωtd .
Therefore, for distortionless transmission the amplitude response |H(ω)| must be a constant and the phase re-
sponse H(ω) must be a linear function of ω with slope −td (td is the time delay). If the slope of H(ω) is not
constant, then different frequency components in the input signal undergo different amounts of time delay and
the output is not a replica of the input waveform. Generally, the human ear is sensitive to amplitude distortion but
relatively insensitive to phase distortion. This is the reason why the manufacturers of audio equipment make only
available the |H(ω)| characteristics of their systems. For video signals the situation is the opposite. The human
eye is sensitive to phase distortion but is relatively insensitive to amplitude distortion. Phase distortion causes
different time delays in different picture elements, resulting in a smeared picture.
Ideal filters allow distortionless transmission of a certain band of frequencies and suppress the remaining fre-
quencies. Figure 27 represents the spectral response of ideal lowpass, highpass and bandpass filters, respectively.

15
 (a) Ideal lowpass filter (b) Ideal highpass filter

(b) Ideal bandpass filter

Figure 27: Ideal filters.

For(a)the lowpass
Ideal lowpass filter
filter of Figure 27(a) (b)we have:
Ideal highpass filter
 ω 
H(ω) = rect e−jωtd .
2W
By Equation (22) and the time-shifting property
(see Page 14) we obtain:
ghpass filter  ω 
h(t) = F
−1
h i
−jωtd
rect e
2W Figure 28: Impulse response of ideal lowpass filter.
(b) Ideal bandpass filter
W
= sinc (W (t − td )) .
π
The impulse response h(t) is shown in Figure 28. Clearly the filter is still non causal and hence not realisable.
One practical approach is to cut off the tail of
h(t) for t < 0:
ĥ(t) = h(t)u(t),
where u(t) is the unit step function defined in
Example 1 on Page 9 (see Figure 15). The
truncated signal ĥ(t) is shown in Figure 29.
Figure 29: Truncated impulse response of ideal lowpass filter. If td is sufficiently large, ĥ(t) will be a close
approximation of h(t) and the resulting filter
Ĥ(ω) will be a good approximation of the ideal filter. A glance at Figure 28 shows that a delay td of three to four
times π/W will make ĥ(t) a reasonably close version of h(t).
We have seen that, for a physically realisable system, h(t) must be causal. That is,
h(t) = 0 for t < 0.
In the frequency domain, this condition is equivalent to the Paley-Wiener criterion, which states that the neces-
sary and sufficient condition for the amplitude response |H(ω)| to be realisable is:
Z ∞
| ln|H(ω)| |
dω < ∞. (25)
−∞ 1 + ω2

Note that if H(ω) = 0 over any finite band (that is, | ln|H(ω)| | = ∞) then condition (25) is not satisfied.
Therefore, for a physically realisable system, H(ω) may be zero at some discrete frequencies, but it cannot be
zero over any finite band. Hence, the ideal characteristics of Figure 27 are not realisable.

16
Case-study of a practical filter
We obtained earlier that the circuit represented in Figure 3 can be described by the differential equation (6):

dvo (t) 1 1
+ vo (t) = − vi (t).
dt CRf CRi

Applying the Fourier transform to the above equation and using the time differentiation property (see Page 14)
we obtain:
1 1
jωVo (ω) + Vo (ω) = − Vi (ω),
CRf CRi
that is,
1
− CR
Vo (ω) = i
1 Vi (ω).
jω + CR
| {z f}
H(ω)

By inspecting the above equation and recalling (18) we conclude that the frequency response of the filter (i.e., the
Fourier transform of its impulse response) is:
1
− CR
H(ω) = i
1 . (26)
jω + CRf

(If you need further convincing, consider that the Fourier transform of the unit impulse function, see Page 12, is
F[δ(t)] = 1. Hence, when the input is vi (t) = δ(t), the Fourier transform Vo (ω) of the output—i.e., the Fourier
transform of the impulse response—is precisely H(ω), since Vo (ω) = H(ω)Vi (ω) = H(ω) × 1 = H(ω).)
From Example 1 on Page 9 we can deduce that the expression for the impulse response in the time domain is:
1 − CR1 f t
h(t) = − e u(t),
CRi
which coincides with the expression we obtained on Page 7 after having solved the differential equation.
On Page 5 we mentioned that the impulse response can be readily obtained using Laplace transforms. We can
see from the above derivations that the same can be achieved by using Fourier transforms.
To plot the frequency response of the filter (see Page 11)
we need to compute the amplitude and phase of H(ω).
Rewriting (26) we obtain:

−(Rf /Ri )
H(ω) = ,
jCRf ω + 1

and, computing the amplitude of H(ω),

Rf /Ri
|H(ω)| = p ,
(CRf ω)2 + 1

and the phase of H(ω),

H(ω) = π − tan−1 (CRf ω).

The amplitude and phase of the frequency response are

shown in Figure 30. We can verify that the frequency
response corresponds to a lowpass filter, as we had con- Figure 30: Frequency response of the practical filter
cluded previously from the intuitive analysis performed corresponding to the circuit shown in Figure 3.
on Pages 2–3.

17
Data truncation: Window functions
We often need to truncate data, e.g., in numerical computations we have to deal with data of finite duration.
Another example is to make the response of an ideal filter causal, as seen in Figures 28–29. In signal sampling,
to eliminate aliasing we will see later that we need to truncate the signal’s spectrum beyond the half sampling
frequency ωs /2 using an anti-aliasing filter.
Truncation can be regarded as multiplying a signal of a large width by a window function of a smaller width.

Spectral spreading
Consider a signal f (t) and a window w(t). If f (t) ⇐⇒ F (ω) and w(t) ⇐⇒ W (ω), then the frequency convolu-
tion property of the Fourier transform (see Page 14) gives:
1
fw (t) = f (t)w(t) ⇐⇒ Fw (ω) = F (ω) ∗ W (ω). (27)
2π
Hence we can see by the width property of convolution (see Page 6) that the width of Fw (ω) is the sum of the
widths of F (ω) and W (ω). Thus, truncation of a signal increases its bandwidth (spreads its spectrum).

Leakage
The window function W (ω) is really not strictly bandlimited (as we will see later in a tutorial problem, a signal
cannot be simultaneously timelimited and bandlimited) and its spectrum goes to 0 only asymptotically. This
causes the spectrum of F (ω) to leak in the band where it is supposed to be zero.

Example of filter design using windows

Recall from (22) that an ideal lowpass filter of bandwidth W rad/sec has impulse response:
W
h(t) = sinc (W t) , (28)
π
which is non causal (see Figure 26) and hence unrealisable. Truncation of h(t) by a suitable window (and
time delaying the resulting signal—see, e.g., Figure 29) makes it realisable. In this example, we will consider
truncation with a rectangular window wR (t) and a triangular window wT (t). In Figure 31 we show the effect
of multiplying h(t) of Equation (28) by a rectangular window wR (t) of width T (top part) and by a triangular

Figure 31: Filter design using windows (time-domain).

window wT (t) of width T (bottom part). (For comparison, we also show with a dashed-line in the bottom part the
effect of the rectangular window.)

18
 '(")*+#!&
'(+"#,-.%&

!"#$#%"&

'(")*+#!&
'(+"#,-.%&

Figure 32: Filter design using windows (frequency-domain).

As can be seen from (27), the effect of windowing in the frequency domain is given by the convolution of
the corresponding spectra. The result of convolving H(ω) with the Fourier transform of the rectangular window,
WR (ω), is shown in the top part of Figure 32. And the result of convolving H(ω) with the Fourier transform of
the triangular window, WT (ω), is shown in the bottom part of Figure 32.
Notice the spectral spreading at the edges of the spectra since, instead of a sudden switch, there is a gradual
transition from the passband to the stopband of the filter. The transition band is smaller (2π/T rad/sec) for the
rectangular window compared to the triangular window (4π/T rad/sec). In fact, among all the windows of a given
width, the rectangular window has the smallest spectral spread (this is because spectral spreading is determined
by the width of the mainlobe, and this width is minimal for the rectangular window).
Also notice that, although H(ω) is bandlimited, the windowed filters are not. The stopband behaviour is
superior in the triangular case than in the rectangular case. For the rectangular window, the leakage in the stopband
decreases slowly (as 1/ω) compared to that of the triangular window (1/ω 2 ). Moreover, the rectangular case has
a higher peak sidelobe amplitude compared to that of the triangular window. This is because leakage is produced
by a slow decay of the sidelobes of the window spectrum. For a signal with jump discontinuity (as the rectangular
window) the Fourier spectrum decays as 1/ω, while for a continuous signal whose derivative is discontinuous (as
the triangular window) the Fourier spectrum decays as 1/ω 2 .
In general, for a given window width, the remedies for the two effects (spectral spreading and leakage) are
incompatible; improving one deteriorates the other, and vice versa.
To reduce the spectral spread (mainlobe width) we need to increase the window’s width (which decreases the
signal’s bandwidth, as already noted from the scaling property of the Fourier transform—see Page 14).
To improve the leakage behaviour we need to select a suitably smooth window (with a faster decay of its
spectrum).
Thus, we can remedy both side effects of truncation by selecting a suitably smooth window of sufficient
width. Some other popular examples of window functions are the Hanning window (faster sidelobe decay) and
the Hamming window (smallest sidelobe magnitude for a given mainlobe width). If fact, there are hundreds of
windows, each with different characteristics, and the choice depends on the particular application. (For more
details on different window functions and their characteristics, see Section 4.9 and, in particular, Table 4.3 on
Page 305 of Lathi’s textbook.)

19
Brief review of the Laplace transform
In this course we will not give the Laplace transform a great deal of attention. We will take it as assumed
knowledge. However, most of what you will need to know is briefly reviewed here.

Let us consider again Example 1 on Page 9, where we con-

cluded that the Fourier transform of the function

f (t) = e−at u(t)

is equal to
1
F (ω) = , when a > 0.
jω + a
Figure 33: Growing exponential.
However, the Fourier transform does not converge when a < 0, in
which case the function looks like the plot of Figure 33.
But, how about if we could devise a transform with some ‘built-in
convergence’? Suppose we multiply f (t) by the factor e−σt , with
σ > −a. Now, the function fˆ(t) = e−at u(t)×e−σt = e−(a+σ)t u(t)
looks like the plot of Figure 34, and it is Fourier transformable
since a + σ > 0. Indeed,
Z ∞ Z ∞
F̂ (ω) = fˆ(t)e −jωt
dt = e−(a+σ)t u(t)e−jωt dt
−∞ −∞
1 1
Figure 34: Decaying exponential. = = , for σ > −a.
%&'()$ *+*),-$ jω +.()'()$
(a + σ) (σ + jω) +a
| {z }
s

Such a transform (with the!"#$‘built-in convergence’ factor e−σt ) is known as the Laplace transform, which is an
extension of the Fourier transform by generalising the frequency jω to the ‘complex frequency’ s = σ + jω.
For a causal signal f (t) [f (t) = 0 for t < 0], the Laplace transform is:
Z ∞
F (s) = L [f (t)] = f (t)e−st dt,
0−

!"#$
where s = σ + jω .

In the previous example, f (t) = e−at u(t), we obtained:

1
F (s) = , valid for Re{s} > −a .
s+a | {z }
ROC = Region of convergence
%&'()$

When a > 0:

The Laplace transform in this case is equal to !"#$

1 1
F (s) = = . In particular we can
s+a σ + jω + a
take σ = 0 (since it is inside the ROC) and we get
the Fourier transform:
1 Figure 35: Function f (t) = e−at u(t) when a > 0 and
F (jω) = . !"#$
jω + a ROC of its Laplace transform.

20
When a < 0:

The Laplace transform in this case is also equal to !"#$

1 1
F (s) = = . However, we cannot
s+a σ + jω + a
take σ = 0 (since it is outside the ROC).

Hence, in this case, the Fourier transform does

not exist! Figure 36: Function f (t) = e−at u(t) when a < 0 and
ROC of its Laplace transform.

As we can see, the definition of the Laplace transform is identical to the Fourier transform with jω replaced
by s. Indeed, when the region of convergence (ROC) of the Laplace transform F (s) of a function f (t) includes
the imaginary jω-axis, then the Fourier transform can be obtained simply by replacing s by jω:
F (jω) = F (s)|s=jω (29)

A word about notation: Lathi uses F (ω) and F (jω) to represent

the same thing! And we shall adopt the same convention.

When the ROC of the Laplace transform does not contain the imaginary axis, the connection between the
Fourier and the Laplace transforms is not so simple (but we do not need to worry about this technicality in this
course!).
Out of the several properties of the Laplace transform, the main property we will need is the
time differentiation property: If f (t) ⇐⇒ F (s), then,

dn f (t)
⇐⇒ sn F (s) − sn−1 f (0− ) − sn−2 f˙(0− ) − . . . − f (n−1) (0− ). (30)
dtn

Transfer function of an LTI system

Consider a linear time invariant (LTI) system given by the generic differential equation relating the input f (t) and
the output y(t):
dn y(t) dn−1 y(t) dy(t) dn f (t) dn−1 f (t) df (t)
n
+ a n−1 n−1
+ . . . + a 1 + a0 y(t) = b n n
+ b n−1 n−1
+ . . . + b1 + b0 f (t). (31)
dt dt dt dt dt dt
Assume that all the initial conditions are equal to zero. Then, applying property (30) to Equation (31) yields:
sn + an−1 sn−1 + . . . + a1 s + a0 Y (s) = bn sn + bn−1 sn−1 + . . . + b1 s + b0 F (s),

or,
bn sn + bn−1 sn−1 + . . . + b1 s + b0
Y (s) = F (s).
sn + an−1 sn−1 + . . . + a1 s + a0
| {z }
H(s)

Therefore, we conclude that

Y (s) = H(s)F (s). (32)
The block diagram representation of this rela-
tionship is shown in Figure 37. H(s) in (32) is called
the transfer function of the system, and it is given by: %&'()$ *+*),-$ .()'()$
bn sn + bn−1 sn−1 + . . . + b1 s + b0
H(s) = . (33)
sn + an−1 sn−1 + . . . + a1 s + a0 Figure 37: Block diagram representation.
!"#$
21
Frequency response of an LTI system
We will start this section with a straightforward exercise:

Exercise 4
Show that L [δ(t)] = 1.

From the above exercise and (32) we conclude that the transfer function H(s) is the Laplace transform of the
impulse response h(t) (since, when f (t) = δ(t), F (s) = 1 and, hence, Y (s) = H(s)F (s) = H(s) × 1 = H(s)).
On Page 11 we defined the frequency response of a system to be H(ω), equal to the Fourier transform of the
impulse response. (Recall that Lathi uses both H(ω) and H(jω) to represent the same entity.)
When the system is causal and asymptotically stable, all the poles of H(s) lie in the left half plane (LHP)
1
and the ROC for H(s) includes the imaginary jω-axis. For example, H = is asymptotically stable when

s+a
a > 0 and the ROC is depicted in Figure 35. As we have seen in (29), the frequency response can in this case be
obtained by replacing s = jω in the transfer function of the system H(s). We thus conclude:

The frequency response H(jω) of an asymptotically stable system can

be obtained by substituting s = jω in the system’s transfer function H(s).

For the effect of H(jω) on the amplitude and phase spectra of a signal, see (19)–(20).
In practice, the sinusoidal signals we can generate in the laboratory are of the form:

f (t) = cos (ωt + θ),

1  j(ωt+θ)
+ e−j(ωt+θ) . Recall from Page 10 that the response to ejωt is

which can be written as cos (ωt + θ) = e
2
equal to H(jω)ejωt . Similarly, the response to e−jωt = ej(−ω)t is equal to H(−jω)e−jωt . Also, recall from (16)

that, for a real function h(t), |H(−jω)| = |H(jω)| and H(−jω) = − H(jω). We thus obtain,by applying
ejθ jωt e−jθ −jωt

1  j(ωt+θ) −j(ωt+θ)

superposition, that the response to f (t) = cos (ωt + θ) = e +e = e + e
2 2 2
is given by:
 jθ   −jθ 
e e
y(t) = H(jω)e + jωt
H(−jω)e−jωt
2 2
1 h j(ωt+θ) j H(jω) −j(ωt+θ) −j H(jω)
i
= e |H(jω)|e +e |H(jω)|e
2   
1 j ωt+θ+ H(jω)
 
−j ωt+θ+ H(jω)
= |H(jω)| e +e
2
= |H(jω)| cos (ωt + θ + H(jω)).
We thus conclude:

y(t) = |H(jω)| cos (ωt + θ + H(jω)). (34)

Equation (34) says that, for a sinusoidal input of frequency ω rad/sec, the system’s response is also a sinusoid of
the same frequency ω. The amplitude of the output sinusoid is |H(jω)| times the input’s amplitude, and the phase
of the output sinusoid is shifted by H(jω) with respect to the input phase.
Clearly, the plots of |H(jω)| and H(jω) as functions of ω show at a glance how the system modifies the
amplitudes and phases of various sinusoidal inputs.

22
Filter design by placement of the poles and zeros of H(s)
If we factorise the numerator and denominator of H(s) in (33) we obtain:

(s − z1 )(s − z2 ) . . . (s − zn )
H(s) = bn , (35)
(s − p1 )(s − p2 ) . . . (s − pn )

where z1 , z2 , . . . , zn are the zeros of H(s) and p1 , p2 , . . . , pn are the poles of H(s). Hence, the frequency response
(see the previous page) is:
(jω − z1 )(jω − z2 ) . . . (jω − zn )
H(jω) = bn . (36)
(jω − p1 )(jω − p2 ) . . . (jω − pn )
Each factor (jω − zi ) in the numerator of (36) is a complex
Im number represented by a vector from zi to jω (see Figure 38)
or, in polar form, by ri ejφi (where ri is the magnitude and φi
is the angle of the vector). Similarly, each factor (jω − pi )
in the denominator of (36) is a complex number represented
by a vector from pi to jω (see Figure 38) or, in polar form,
Re by di ejθi (where di is the magnitude and θi is the angle of
the vector). Hence we can rewrite (36) as:

r1 ejφ1 r2 ejφ2 . . . rn ejφn

H(jω) = bn
Im representation of the factors of
Figure 38: Vector (d1 ejθ1 ) (d2 ejθ2 ) . . . (dn ejθn )
r1 r2 . . . rn j[(φ1 +φ2 +...+φn )−(θ1 +θ2 +...+θn )]
H(s). = bn e .
d1 d2 . . . dn
So, we conclude that the amplitude of the frequency response of the filter is given by:
r1 r2 . . . rn product of the distances of the zeros to jω
|H(jω)| = bn = bn ,
Re d1 d2 . . . dn product of the distances of the poles to jω
Im
and the phase of the frequency response of the filter is given by:

H(jω) = (φ1 + φ2 + . . . + φn ) − (θ1 + θ2 + . . . + θn )

= sum of the zero angles to jω − sum of the pole angles to jω.
Re

Gain enhancement by a pole

Consider the situation represented in Fig- Im
ure 39, where a pole has been placed op-
posite the frequency ω0 . By using the
above expressions, we can conclude that
the amplitude of the frequency response
as ω passes close to ω0 varies proportion- Re
ately to 1/d; i.e.,

K
|H(jω)| = ,
d
Figure 39: Gain enhancement by a pole.
(since d0 is relatively bigger than d and,
hence, its variations are not so significant). The frequency response is represented in the middle and right plots of
Figure 39. Therefore, we conclude that we can enhance the gain at a frequency ω0 by placing a pole opposite the
point jω0 .

23
Gain suppression by a zero
Im
Consider the situation represented in Fig-
ure 40, where a zero has been placed op-
Im
posite the frequency ω0 . By a similar
analysis to the previous one we can con-
clude that a zero has the opposite effect.
Re
If the zero in placed at −α ± jω0 , it will
suppress the gain in the vicinity of ω0 .
The frequency response is represented in
the middle and rightReplots of Figure 40.
Im Figure 40: Gain suppression by a zero.
Im
Lowpass filters

Im Consider a first order filter normalised to have DC gain H(0) = 1 (see

Re
Figure 41):
Re ωc
H(s) = .
s + ωc
By a similar analysis to the one carried on previously, |H(jω)| = ωdc
Re
and the frequency response clearly has a lowpass characteristic with
gain enhanced in the vicinity of ω = 0 (see the plot in Figure 42 cor-
Figure 41: Lowpass filter (n = 1). responding to n = 1).

Imgain
An ideal lowpass filter has a constant
of unity up to frequency ωc and the gain
drops suddenly to 0 for ω > ωc (see Fig-
ure 42). We already noticed previously that
such an amplitude response is not physi-
cally realisable since it is zero over a band
of frequencies (in this case, over an infinite Re
band of frequencies) and, hence, it does not
satisfy the Paley-Wiener criterion (see (25)
above). Figure 42: Amplitude response of lowpass (Butterworth) filters.

Im Im Im Im

Re Re Re Re

Figure 43: Wall of poles. Figure 44: Butterworth filter (n = 5).

Im
Im
To achieve an approximation of the ideal lowpass characteristic we need enhanced gain over the entire fre-
quency band from 0 to ωc . That is, we need a continuous wall of poles facing the imaginary axis opposite the

24
Re
Re
 frequency band from 0 to ωc (and from 0 to −ωc for the conjugate poles) as shown in Figure 43. It can be shown
that for a maximally flat response over the frequency range 0 to ωc the wall is a semicircle with poles uniformly
distributed, as shown in Figure 44 for n = 5. Figure 42 above shows the amplitude responses for various values
of n when the Im poles are placed on a semicircle as in Figure 44. As n → ∞, the filter response approaches the
ideal lowpass response. This family of filters is known as Butterworth filters. Another family of commonly used
filters is that of the Chebyshev filters, in which the wall shape is a semi-ellipse. The characteristics of Chebyshev
filters are inferior to those of Butterworth filters over the passband 0 to ωc (they have a rippling effect) but in the
stopband, ωc to ∞, their behaviour is superior (the gain drops faster than in the Butterworth filters).
Re Re

Bandpass filters
In the case of a bandpass filter, we need enhanced gain over the entire passband, as shown in Figure 45.

Im Im

Re
Re

Figure 45: Bandpass filter. Figure 46: Bandstop filter.

Notch (bandstop) filters

A practical second-order notch filter (or bandstop filter) to obtain zero gain at a frequency ω = ω0 can be realised
with two poles and two zeros, as shown in Figure 46.
A bandstop filter is the complement of a bandpass filter, thus:
HBS (s) = 1 − HBP (s).
Similarly, a highpass filter is the complement of a lowpass filter:
HHP (s) = 1 − HLP (s).

Practical filter specifications

An ideal filter has a passband (with unity gain) and a stopband (with zero gain) with sudden transitions between
them (see Figure 27). For practical (or realisable) filters, the transitions have to be gradual. Moreover, for
realisable filters the gain cannot be zero over a finite band (due to Paley-Wiener’s condition, see Equation (25)
and the related discussion on Page 16). We thus need to define:

Gp = minimum passband gain,

Gs = maximum stopband gain,

and the corresponding frequencies, ωp , ωs , etc., as indicated in Figure 47.

Usually the gains are specified in terms of decibels:
Ĝ(dB) = 20 log10 G. (37)
√ √
A gain of 1 is equal to 0 dB, a gain of 2 is (approximately) equal to 3 dB, a gain of 1/ 2 is (approximately)
equal to −3 dB (also called an attenuation of 3 dB), etc.

25
 Lowpass filter Bandpass filter

Highpass filter Bandstop filter

Figure 47: Practical filter specifications.

Butterworth filters
The amplitude response |H(jω)| of an nth order Butterworth lowpass filter is:
1
|H(jω)| = r  2n . (38)
ω
1+ ωc

At ω = 0 the DC gain is |H(j0)| = 1. √

At ω = ωc the gain drops by a factor of 2 (3 dB). The frequency ωc is called the 3 dB-cutoff frequency.
It is convenient to design first a normalised filter H(s) with ωc = 1, and the desired transfer function H(s) is
then obtained by replacing s in H(s) by s/ωc (see frequency scaling on Page 28).
The amplitude characteristic of a normalised lowpass
Butterworth filter is:
1
|H(jω)| = √ , (39)
1 + ω 2n
and it is shown in Figure 48 for various values of n. In the
amplitude response we note:
1. The amplitude response decreases monotonically;
2. It is ‘maximally flat’ at ω = 0 (i.e., the first 2n − 1
derivatives are zero at ω = 0);
3. The 3 dB bandwidth is 1 rad/sec; Figure 48: Amplitude response of a normalised
4. For large n, the amplitude response approaches the lowpass Butterworth filter.
ideal characteristic.
To determine the transfer function, recall from (16) that H(jω) and H(−jω) are complex conjugate, thus,
1
H(jω)H(−jω) = |H(jω)|2 = .
1 + ω 2n
Substituting s = jω we obtain:
1
H(s)H(−s) =  2n .
s
1+ j

26
The poles of H(s)H(−s) satisfy
Im
Im
2n 2n
s = −j .

We have, −1 = ejπ(2k−1) , with k an integer, and j = ejπ/2 .

Thus,

2n Re
s2n = −1 × ejπ/2 = ejπ(2k−1) ejπn = ejπ(2k−1+n) ,

with k an integer. Hence, the poles are:

π
sk = ej 2n (2k+n−1) , k = 1, 2, . . . , 2n.
Figure 49: Poles of a normalised lowpass Butter-
The poles sk are located on the unit circle, separated by an worth filter transfer function and its conjugate.
angle of π/n, as shown in Figure 49.
Since H(s) is required to be stable, the poles must lie on the left-half plane (LHP). The poles of H(s) are thus
obtained by setting k = 1, 2, . . . , n (the other poles, for k = n + 1, n + 2, . . . , 2n, correspond to H(−s) and are
the reflections about the vertical imaginary axis).
Thus, the poles of H(s) are:
π π π
sk = ej 2n (2k+n−1) = cos (2k + n − 1) + j sin (2k + n − 1), k = 1, 2, . . . , n,
2n 2n
and H(s) is given by:
1
H(s) = .
(s − s1 ) . . . (s − sn )

For example, for n = 4 we have the poles separated by an angle of

Im π π 5π 7π 9π 11π
= 2 × , and they are located at angles , , and (see Fig-
4 8 8 8 8 8
ure 50), and given by −0.3827 ± j0.9239 and −0.9239 ± j0.3827. The
transfer function of the normalised filter is thus given by:
1
Re Re H(s) =
(s + 0.3827 − j0.9239)(s + 0.3827 + j0.9239)
1
×
(s + 0.9239 − j0.3827)(s + 0.9239 + j0.3827)
1
= 2
Figure 50: Poles of normalised low- (s + 0.7654s + 1)(s2 + 1.8478s + 1)
pass Butterworth filter, n = 4. 1
= 4 .
s + 2.6131s + 3.4142s2 + 2.6131s + 1
3

We can proceed in this way to find H(s) for any value of n. In general,
1 1
H(s) = = n , (40)
Bn (s) s + an−1 sn−1 + . . . + a1 s + 1

where Bn (s) is the Butterworth polynomial of nth order. The design of Butterworth filters is facilitated by
ready-made tables of the coefficients of Bn (s) for various values of n. For example, see Tables 7.1 and 7.2 on
Page 509 of Lathi’s textbook. There are also Matlab c
functions that give the Butterworth filter transfer function,
for example the function buttap.

27
Frequency scaling
The procedure explained above, and the tables with coefficients of Butterworth filters, e.g., Tables 7.1 and 7.2
on Page 509 of Lathi’s textbook, are for normalised Butterworth filters with 3 dB bandwidth ωc = 1. The
results can be extended to any value of ωc by replacing s by s/ωc (this implies replacing ω by ω/ωc in (39), thus
obtaining (38)). For instance, from the previous example  = 4) we can obtain a fourth-order Butterworth filter
 s (n
with ωc = 10 by replacing s by s/10, i.e., H(s) = H , yielding,
10
1
H(s) =
s 4 s 3
s 2 s





10
+ 2.6131+ 3.4142 10
10
+ 2.6131 10 +1
10, 000
= 4 .
s + 26.131s + 341.42s2 + 2, 613.1s + 10, 000
3

The amplitude response |H(jω)| of this filter is identical to that of the normalised filter |H(jω)| in Equation (39)
(shown in Figure 48), expanded by a factor of 10 along the horizontal ω-axis (frequency scaling).

Determination of the filter order n

Suppose a gain Ĝp in dB (recall the definition of dB in (37)) is specified at frequency ωp , and a gain Ĝs in dB is
specified at frequency ωs for a lowpass Butterworth filter, as shown on the top-left plot of Figure 47. From (37)
and (38) we obtain: "  2n #
ωp
Ĝp = 20 log10 |H(jωp )| = −10 log10 1 + ,
ωc
"  2n #
ωs
Ĝs = 20 log10 |H(jωs )| = −10 log10 1 + .
ωc
Rearranging,
 2n
ωp
= 10−Ĝp /10 − 1, (41)
ωc
 2n
ωs
= 10−Ĝs /10 − 1. (42)
ωc

Dividing (42) by (41) we obtain:

2n
10−Ĝs /10 − 1

ωs
= ,
ωp 10−Ĝp /10 − 1
and, finally, h   i
−Ĝs /10 −Ĝp /10
log10 10 − 1 / 10 −1
n= . (43)
2 log10 (ωs /ωp )
From (41) we obtain,
ωp
ωc = h i1/2n . (44)
10−Ĝp /10 − 1
Alternatively, from (42) we obtain,
ωs
ωc = h i1/2n . (45)
10−Ĝs /10 −1
In general, when computing n from (43) we obtain a non integer number. To be able to realise the filter, we
then have to approximate the order of the filter by the next larger integer. This implies that we will obtain two
different values for ωc from (44) and (45). With a larger n, the roll-off (i.e., the steepness of the transition from the

28
passband to the stopband) will be increased (notice how the roll-off increases with n in Figure 48). This means
that, if we compute ωc from (44) then the response will satisfy exactly the requirement Gp over the passband
0 ≤ ω ≤ ωp and will surpass the requirement Gs on the stopband ω ≥ ωs (see the top-left plot of Figure 47). On
the other hand, the use of (45) to compute ωc will exactly satisfy the requirement on Gs but will oversatisfy the
requirement for Gp . (The choice is yours!)

Chebyshev filters
The amplitude response of a normalised Chebyshev lowpass filter is:
1
|H(jω)| = p , (46)
1 + ε2 Cn 2 (ω)

where Cn (ω) is the nth-order Chebyshev polynomial given by:

Cn (ω) = cos n cos−1 ω .

(47)

An alternative expression for Cn (ω) is:

Cn (ω) = cosh n cosh−1 ω .

(48)

Form (47) is most convenient for |ω| < 1 and form (48) is convenient for |ω| > 1. The Chebyshev polynomial
has the property
Cn (ω) = 2ωCn−1 (ω) − Cn−2 (ω), n > 2.
Thus, we can find the polynomials recur-
sively as follows:

C0 (ω) = cos 0 = 1,
C1 (ω) = cos 1 × cos−1 ω = ω,

C2 (ω) = 2ωC1 (ω) − C0 (ω) = 2ω 2 − 1,

C3 (ω) = 2ωC2 (ω) − C1 (ω)
= 4ω 3 − 2ω − ω = 4ω 3 − 3ω,

etc. (see Table 7.3 on Page 514 of Lathi’s

textbook).
The normalised Chebyshev lowpass Figure 51: Amplitude response of normalised sixth- and seventh-
amplitude response for n = 6 and n = 7 order lowpass Chebyshev filters.
is shown in Figure 51.
In general, we have the following observations on the amplitude response of Chebyshev filters:
1. The amplitude response has ripples in the passband 0 ≤ ω ≤ 1, and there are a total of n maxima and
minima;
2. The amplitude response is smooth and decreases monotonically in the stopband ω > 1;
3. From the expressions for Cn (ω) given above (see also Table 7.3 on Page 514 of Lathi’s book) we have:
(
2 0 when n is odd,
Cn (0) =
1 when n is even.

Therefore, the DC gain of the filter is


1 when n is odd,
|H(0)| = 1 (49)
√ when n is even;
1 + ε2

29
 4. The Chebyshev polynomials are equal-ripple functions, hence the ripples in the passband are of equal
height. The parameter ε controls the height of the ripple, the ratio of the maximum gain to the minimum
gain in the passband is:
√ √
r = 1 + ε2 , or, in dB: r̂ = 20 log10 1 + ε2 = 10 log10 1 + ε2 ,

from where we obtain:

ε2 = 10r̂/10 − 1. (50)
1 1
5. At ω = 1 the amplitude response is |H(j1)| = √ = . Hence, r̂ in dB (with an opposite sign) takes
1+ε 2 r
the place of Ĝp on the top-left plot of Figure 47. This means that in Chebyshev filter design we do not need
to compute the frequency ωc (as in Butterworth filters, cf. (44)–(45)). For Chebyshev filters, the critical
frequency happens to be the frequency where the gain is Gp . This frequency is ωp = 1 for the normalised
filters (see Figure 51).
6. If we reduce the ripple by decreasing ε the passband behaviour improves, but at the cost of an increased
gain in the stopband. In the extreme case, there is no ripple but the filter becomes an allpass filter, as seen
from (46) when ε = 0.
7. The Chebyshev filter has a sharper cutoff (smaller transmission band) than the same order Butterworth filter
(at the expense of an inferior passband behaviour caused by rippling).

Transfer function and pole locations

The transfer function H(s) of the normalised nth-order Chebyshev filter is given by:

Kn Kn
H(s) = 0
= n n−1
, (51)
Cn (s) s + an−1 s + . . . + a1 s + a0

where the polynomial Cn0 (s) can be found in ready-made tables (e.g., see Table 7.4 on Page 518 of Lathi’s
textbook). There are also Matlab c
functions that give the Chebyshev filter transfer function, for example the
function cheb1ap.
Im
The constant Kn in (51) is selected to have proper DC gain
|H(0)| according to (49). Thus,

 a0 when n is odd,
Kn = a0 a0 (52) Re
√ = r̂/20 when n is even.
1+ε 2 10

It can be shown that the poles of an nth order Chebyshev

filter lie on a semi-ellipse, as shown in Figure 52 for n = 3. Figure 52: Poles of Chebyshev filter, n = 3.

Frequency scaling
The procedure explained above, and the tables with coefficients of Chebyshev filters, e.g., Table 7.4 on Page 518
of Lathi’s textbook, are for normalised Chebyshev filters with ωp = 1. The results can be extended
  to any value
s
of ωp by replacing s by s/ωp in the normalised transfer function H(s). That is, H(s) = H . The resulting
ωp
amplitude response is then obtained from (46),
1
|H(jω)| = r  . (53)
2
1 + ε Cn ωωp
2

30
Determination of the filter order n
Suppose a gain Ĝs in dB (recall the definition of dB in (37)) is specified at frequency ωs for a lowpass Chebyshev
filter, as shown on the top-left plot of Figure 47. From (37) and (53) we obtain:
  
2 2 ωs
Ĝs = 20 log10 |H(jωs )| = −10 log10 1 + ε Cn ,
ωp
or,  
ωs
2
ε Cn 2
= 10−Ĝs /10 − 1.
ωp
Use of (48) and (50) in the above expression yields
  " −Ĝs /10 #1/2
−

ωs 10 1
cosh n cosh−1 = .
ωp 10r̂/10 − 1

Finally, h i1/2 
−1 10−Ĝs /10 −1
cosh 10r̂/10 −1
n= .
cosh−1 (ωs /ωp )

Frequency transformations
We can obtain transfer functions of highpass, bandpass and bandstop filters using frequency transformations on
a basic lowpass filter, called the prototype filter. The prototype lowpass filter, denoted Hp (s), can be, e.g., a
Butterworth or a Chebyshev filter. We then replace s in the prototype filter with a proper transformation T (s), as
explained below.

Highpass filters

Highpass filter Prototype lowpass filter

Figure 53: Frequency transformation for highpass filters.

Given the highpass filter specifications shown on the left plot of Figure 53, we first design a prototype lowpass
filter Hp (s) with the specifications shown on the right plot of Figure 53, and then we replace s with T (s) in Hp (s),
where
ωp
T (s) = . (54)
s
To see how this transformation works, note that when s = jω we have:
Bandpass filter ωp ωp Prototype lowpass filter
T (jω) = = −j ,
jω ω
and, hence, when ω → 0 we have that T (jω) → −j∞. We can then see that the resulting transformed filter
has, at low frequencies (ω → 0), the characteristics of the prototype lowpass filter at high (negative) frequencies

31
(T (jω) → −j∞). Recalling, from (16), that the amplitude response of the filter has conjugate symmetry, that
is |Hp (−jω)| = |Hp (jω)|, we can then see that the transformed filter has, at low frequencies, the characteristics
of the prototype lowpass filter at high positive frequencies as well and, thus, it attenuates the low frequencies as
required on the left plot of Figure 53.
ωp
Performing a similar analysis, and noticing that, when ω = ωs , T (jωs ) = −j ; when ω = ωp , T (jωp ) =
ωs
−j1; and, when ω → ∞, T (jω) → 0, we can see that when the prototype lowpass filter satisfies the specifications
on the right plot of Figure 53, the transformed filter satisfies the original highpass filter specifications given on
the left plot of the figure.

Bandpass filters Highpass filter Prototype lowpass filter

Bandpass filter Prototype lowpass filter

Figure 54: Frequency transformation for bandpass filters.

Given the bandpass filter specifications shown on the left plot of Figure 54, we first design a prototype lowpass
filter Hp (s) with the specifications shown on the right plot of Figure 54, where ωs is given by:

ωp1 ωp2 − ωs1 2 ωs2 2 − ωp1 ωp2

 
ωs = min , . (55)
(ωp2 − ωp1 ) ωs1 (ωp2 − ωp1 ) ωs2

We then replace s with T (s) in Hp (s), where

s2 + ωp1 ωp2
T (s) = . (56)
(ωp2 − ωp1 ) s
To see how this transformation works, note that when s = jω we have:
−ω 2 + ωp1 ωp2 ω 2 − ωp1 ωp2
T (jω) = =j .
(ωp2 − ωp1 ) jω (ωp2 − ωp1 ) ω
Noticing that:
ω → 0 ⇒ T (jω) → −j∞,
ωs1 2 − ωp1 ωp2 ωp1 ωp2 − ωs1 2
 
ω = ωs1 ⇒ Im {T (jωs1 )} = Im j =− ≤ −ωs ,
(ωp2 − ωp1 ) ωs1 (ωp2 − ωp1 ) ωs1
ωp1 2 − ωp1 ωp2
ω = ωp1 ⇒ T (jωp1 ) = j = −j1,
(ωp2 − ωp1 ) ωp1
ωp2 2 − ωp1 ωp2
ω = ωp2 ⇒ T (jωp2 ) = j = j1,
(ωp2 − ωp1 ) ωp2
2
− ωp1 ωp2 ωs2 2 − ωp1 ωp2
 
ωs2
ω = ωs2 ⇒ Im {T (jωs2 )} = Im j = ≥ ωs ,
(ωp2 − ωp1 ) ωs2 (ωp2 − ωp1 ) ωs2
ω → ∞ ⇒ T (jω) → j∞,

32
and performing a similar analysis to the one above, for the case of a highpass filter, we can see that when the
prototype lowpass filter satisfies the specifications on the right plot of Figure 54, the transformed filter satisfies
the original bandpass filter specifications given on the left plot of the figure.

Bandstop filters

Bandstop filter Prototype lowpass filter

Figure 55: Frequency transformation for bandstop filters.

Given the bandstop filter specifications shown on the left plot of Figure 55, we first design a prototype lowpass
filter Hp (s) with the specifications shown on the right plot of Figure 55, where ωs is given by:

(ωp2 − ωp1 ) ωs1 (ωp2 − ωp1 ) ωs2

 
ωs = min , . (57)
−ωs1 2 + ωp1 ωp2 ωs2 2 − ωp1 ωp2

We then replace s with T (s) in Hp (s), where

(ωp2 − ωp1 ) s
T (s) = . (58)
s2 + ωp1 ωp2

To see how this transformation works, note that when s = jω we have:

(ωp2 − ωp1 ) jω (ωp2 − ωp1 ) ω

T (jω) = 2
=j .
−ω + ωp1 ωp2 −ω 2 + ωp1 ωp2

Noticing that:
ω → 0 ⇒ T (jω) → 0,
(ωp2 − ωp1 ) ωp1
ω = ωp1 ⇒ T (jωp1 ) = j = j1,
−ωp1 2 + ωp1 ωp2
(ωp2 − ωp1 ) ωs1 (ωp2 − ωp1 ) ωs1
 
ω = ωs1 ⇒ Im {T (jωs1 )} = Im j 2
= ≥ ωs ,
−ωs1 + ωp1 ωp2 −ωs1 2 + ωp1 ωp2
(ωp2 − ωp1 ) ωs2 (ωp2 − ωp1 ) ωs2
 
ω = ωs2 ⇒ Im {T (jωs2 )} = Im j 2
=− ≤ −ωs ,
−ωs2 + ωp1 ωp2 ωs2 2 − ωp1 ωp2
(ωp2 − ωp1 ) ωp2
ω = ωp2 ⇒ T (jωp2 ) = j = −j1,
−ωp2 2 + ωp1 ωp2
ω → ∞ ⇒ T (jω) → 0,
and performing a similar analysis to the ones above, for the cases of highpass and bandpass filters, we can see
that when the prototype lowpass filter satisfies the specifications on the right plot of Figure 55, the transformed
filter satisfies the original bandstop filter specifications given on the left plot of the figure.

33
Sampling
Brief review of Fourier series
Recall that a signal f (t) is periodic with period T0 if

f (t) = f (t + T0 ), for all t, (59)

where T0 is the smallest value such that (59) is satisfied. Recall also that a periodic signal with period T0 can be
expressed as an exponential Fourier series:
∞
X
f (t) = Dn ejnω0 t , (60)
n=−∞

2π
where ω0 = is the fundamental frequency. Equation (60) expresses f (t) as a (possibly infinite) sum of
T0
exponential functions of frequencies nω0 (i.e., integer multiples of the fundamental frequency), called the
nth harmonics. Finding the coefficients Dn of the series (60) is quite simple, we just have to compute the follow-
ing integral over an interval of duration T0 :
1
Z
Dn = f (t)e−jnω0 t dt. (61)
T0 T0

(Where the interval of duration T0 is located on the real axis does not matter since f (t)e−jnω0 t is periodic with
period T0 .)
What about the Fourier
R∞ transform of such a signal?

Since a periodic signal never extinguishes, the first
Dirichlet condition −∞ |f (t)|dt < ∞, see Page 10 is not satisfied! However, notice that those conditions
are only sufficient (but not necessary). Anyway, you would expect something remarkable to happen with such
a signal. Intuitively, one would expect such a signal to have its spectrum concentrated at the fundamental and
harmonic frequencies. Effectively, applying the result of Exercise 2 on Page 12 to (60) we obtain
∞
X
F (ω) = 2π Dn δ(ω − nω0 ), (62)
n=−∞

that is, not only is the spectrum concentrated at the harmonic frequencies, it actually shoots to infinity at those
frequencies!
As an example, let us consider a remarkable signal, the
impulse train (also known as the ‘Dirac comb’) with pe-
riod T , shown in Figure 56,
∞
X
δT (t) = δ(t − nT ). (63)
n=−∞ Figure 56: Impulse train with period T .
2π
The frequency of this signal is ωs = = 2πFs (where Fs is the frequency expressed in hertz = cycles/sec).
T
We obtain Dn from (61),
1 T /2
Z
Dn = δT (t)e−jnωs t dt.
T −T /2
In the impulse train (see Figure 56) there is only one impulse in the interval [−T /2, T /2], that is, δ(t). Hence,

1 T /2 1 1
Z
Dn = δ(t)e−jnωs t dt = e−jnωs ×0 = ,
T −T /2 T T

where we have used the sampling property of the impulse function (see (9)).

34
 Hence, δT (t) can be expressed as a Fourier series (60), i.e.,
∞
1 X jnωs t
δT (t) = e , ωs = 2π/T.
T n=−∞

Applying again the result of Exercise 2 on Page 12 we get

∞ ∞
2π X
F [δT (t)] =
X
δ(ω − nωs ) = ωs δ(ω − nωs ).
T n=−∞ n=−∞

Thus, we have derived Fourier pair 21 in Table 4.1 of Lathi’s textbook, Page 252,
∞
X ∞
X
δT (t) = δ(t − nT ) ⇐⇒ ωs δ(ω − nωs ), ωs = 2π/T. (64)
n=−∞ n=−∞

The sampling theorem

Consider a real signal f (t) whose spectrum is bandlimited to B hertz, as shown on the top plots of Figure 57.
Sampling f (t) at a rate of Fs Hz (i.e., at Fs samples per second) can be accomplished by multiplying f (t) by an

Lowpass
filter

Figure 57: Sampled signal and its Fourier spectrum.

impulse train δT (t) consisting of unit impulses repeated periodically every T seconds (T = 1/Fs ), as expressed
by (63) and shown in Figure 56. The result is the sampled signal f¯(t) shown on the bottom-left plot of Figure 57,
consisting of impulses spaced every T seconds, each weighted by the value of the function at that instant,
∞
f¯(t) = f (t)δT (t) =
X
f (nT )δ(t − nT ).
n=−∞

By the frequency convolution property (see Page 14) we have:

1
f¯(t) = f (t)δT (t) ⇐⇒ F̄ (ω) = F (ω) ∗ F [δT (t)] .
2π
Therefore, using (64), and recalling the shift property and the convolution with an impulse property (see Proper-
ties 4 and 5 on Page 6), we obtain
∞ ∞
1 2π X 1 X
F̄ (ω) = F (ω) ∗ δ(ω − nωs ) = F (ω − nωs ).
2π T n=−∞ T n=−∞

35
Therefore,
∞
1 X
F̄ (ω) = F (ω − nωs ). (65)
T n=−∞

2π 1
Thus, the spectrum of f¯(t) is the spectrum of f (t) repeated periodically with period ωs = rad/sec, or Fs =
T T
Hz, and divided by T , as shown on the bottom-right plot of Figure 57.
If we want to reconstruct f (t) from f¯(t) we should be able to recover F (ω) from F̄ (ω). As can be seen from
Figure 57, this recovery is possible if there is no overlap between successive cycles of F̄ (ω). That is, we require:


 Fs ≥ 2B,

Nyquist-Shannon sampling theorem: or, equivalently, (66)
1

T ≤ 2B .


As long as the sampling frequency Fs is greater than twice the signal bandwidth B (in hertz), F̄ (ω) will
consist of non overlapping repetitions of F (ω) and f (t) can be recovered from its samples f¯(t) by passing the
sampled signal f¯(t) though an ideal lowpass filter of bandwidth B Hz. The frequency Fs = 2B is called the
1
Nyquist rate for f (t). The sampling interval T = is called the Nyquist interval for f (t).
2B

Signal reconstruction: The interpolation formula

As we have seen, if the signal f (t) is bandlimited to B Hz it can be reconstructed (interpolated) exactly
from its samples by passing the sampled signal through an ideal lowpass filter of bandwidth B and gain T .
The frequency response of such filter is
 ω 
H(ω) = T rect ,
4πB
and it is shown in Figure 58. From (22) we can obtain the impulse
response [inverse Fourier transform of H(ω)] of the filter as,

Figure 58: FrequencySampled

response of ideal 2πB
signal h(t) =Reconstructed
T sinc (2πBt) = |2BT
signal {z } sinc (2πBt) = sinc (2πBt) ,
lowpass filter for signal reconstruction. π
=1

where we have assumed the Nyquist sampling rate:

1
T = ⇔ 2BT = 1. The impulse response h(t)
2B
is shown in Figure 59. Observe from the figure that
n
h(t) = 0 at each sampling instant t = ± , ex-
2B
cept
P∞ at t = 0. When the sampled signal f¯(t) =
n=−∞ f (nT )δ(t − nT ) is applied to this filter, each
impulse generates a sinc function of height equal to the Figure 59: Impulse response of ideal lowpass filter for
strength of the sample, that is, Sampled signal Reconstructed signal
signal reconstruction.

∞
X ∞
X
f (t) = f (nT )h(t − nT ) = f (nT ) sinc [2πB(t − nT )] . (67)
n=−∞ n=−∞

Equation (67) is the interpolation formula that yields the values of f (t) between samples as a weighted sum
of all the sample values. The process of signal reconstruction by interpolation is illustrated in Figure 60.

36
 Sampled signal Reconstructed signal

Figure 60: Ideal interpolation.

Practical difficulties in signal reconstruction

1- Need for ideal filters

Sampling at the Nyquist rate Fs = 2B Hz re-

sults in repetitions of F (ω) without any gaps,
as represented in Figure 61. To recover f (t)
we need to pass f¯(t) though an ideal lowpass
filter, which is unrealisable (non-causal). Figure 61: Spectrum of a signal sampled at the Nyquist rate.

A practical solution is to sample at a ratio

higher than the Nyquist rate Fs > 2B Hz, as
shown in Figure 62. Now F (ω) can be recov-
ered from F̄ (ω) by using a lowpass filter with a
gradual cutoff characteristic. But, even in this
case, the filter gain must ideally be zero be-
Recovered spectrum
yond the first cycle of F̄ (ω) (something impos-
Figure 62: Spectrum of a signal sampled above the Nyquist rate.
sible, as we have seen from the Paley-Wiener
Recovered spectrum
criterion—see Equation (25) and the related
discussion on Page 16).
In conclusion, it is impossible in practice to recover a bandlimited signal f (t) exactly from its samples, even
Lost tail gets
if the sampling rate is higher than the Nyquist rate. However, as the samplingfolded
rate backincreases, the recovered
Lost tail. signal
approaches the desired signal more closely. Lost tail gets
folded back Lost tail.

2- Aliasing

Recovered spectrum

Lost tail gets

folded back Lost tail.

Figure 63: Aliasing effect.

All practical signals are timelimited, and a signal cannot be simultaneously timelimited and bandlimited (see
the tutorial problems). As a result, there will always be an amount of overlap between the repetitions of the
spectrum of F (ω) and, hence, parts of the spectrum (the tail beyond the half sampling frequency, Fs /2) get

37
folded back producing an effect known as aliasing, illustrated in Figure 63. A practical solution is to use an
antialiasing filter (a lowpass filter) “before the signal is sampled”, so as to suppress the frequency components
beyond the folding frequency Fs /2.

Practical sampling
We assumed before that ideal samples are obtained by multiplying a signal f (t) by an impulse train (as illustrated
in Figure 57), which is physically nonexistent. In practice, we multiply a signal by a train of pulses of finite width,
as shown on the left part of Figure 64.

Lowpass
filter

Figure 64: Practical sampling.

Since the pulse train pT (t) is periodic, we can express it as a Fourier series (60), that is,
∞
X 2π
pT (t) = Dn ejnωs t , ωs = .
n=−∞
T

Applying the result of Exercise 2 on Page 12 to the above expression we obtain,

∞
F [pT (t)] = 2π
X
Dn δ(ω − nωs ).
n=−∞

Hence, by a similar analysis to the one performed on Page 35 for the case of ideal sampling we obtain:
∞ ∞
1 1
F (ω) ∗ F [pT (t)] =
X X
F̄ (ω) = F (ω) ∗ 2π Dn δ(ω − nωs ) = Dn F (ω − nωs ).
2π 2π n=−∞ n=−∞

The spectrum F̄ (ω) of the sampled signal is shown on the bottom-right plot of Figure 64. Clearly, the signal f (t)
can be recovered by lowpass filtering f¯(t), provided ωs > 4πB (i.e., Fs > 2B).

Dual of time-sampling: The spectral sampling theorem

On Page 36 we stated the sampling theorem [see (66)], which establishes that a signal bandlimited to B hertz
(i.e., with spectrum over the frequency range −B to B; or, in other words, with spectral width equal 2B) can be
reconstructed from samples taken at a rate:
samples
Fs ≥ 2B , (2B = spectral width). (68)
second

38
 The dual of the above theorem is the frequency-sampling theorem, which states that the spectrum F (ω) of
a signal timelimited to τ seconds (signal width) can be reconstructed from the samples of F (ω) taken at a rate
R ≥ τ samples per hertz (see (71) below).
Consider a timelimited signal f (t) as shown in Figure 65.

Figure 65: Time limited signal and spectrum.

The Fourier transform [recall (13)] of f (t) is:

Z ∞ Z τ
−jωt
F (ω) = f (t)e dt = f (t)e−jωt dt. (69)
−∞ 0

We now construct a periodic signal fT0 (t), formed by repeating f (t) every T0 seconds (with T0 > τ ), as shown in
Figure 66.

Figure 66: Periodic repetition of a signal amounts to sampling its spectrum.

The periodic signal fT0 (t) can be expressed by an exponential Fourier series (60):
∞
X 2π
fT0 (t) = Dn ejnω0 t , ω0 = ,
n=−∞
T0

where, assuming that τ < T0 , the coefficients are computed using (61):
Z T0 Z τ
1 −jnω0 t 1
Dn = f (t)e dt = f (t)e−jnω0 t dt.
T0 0 T0 0
We can see from (69) that
1
Dn = F (nω0 ),
T0
 
1
that is, the coefficients of the Fourier series of fT0 (t) are times the samples of F (ω) taken at intervals of
T0
ω0 . As long as τ ≤ T0 the successive cycles of f (t) do not overlap, so that f (t) can be recovered from fT0 (t).
ω0 1
Thus, the condition for recovery is T0 ≥ τ . Equivalently, since F0 = = ,
2π T0
1
F0 ≤ Hz, (70)
τ

39
or, in terms of the sampling rate R (samples/Hz),
1 samples
R= ≥τ , (τ = signal width). (71)
F0 hertz
Condition (71) on the sampling of the spectrum for recovery of a signal is the dual of condition (68) on the
sampling of the time-signal for recovery.

The discrete Fourier transform (DFT)

The numerical computation of the Fourier transform of f (t) requires sample values of f (t) because a digital
computer can only work with sequences of numbers. Also, a digital computer can compute F (ω) at only discrete
values of ω (samples of F (ω)). We therefore need to relate the samples of F (ω) to the samples of f (t).

(a) (b)

(c) (d)

(e) (f)

Figure 67: Sampling and periodic repetition of a signal results in sampling and periodic repetition of its spectrum.

Consider a timelimited signal f (t) as shown in Figure 67(a) and its spectrum shown in Figure 67(b). The
spectrum of the sampled signal f¯(t) consists of F (ω) repeated every Fs = 1/T (Figure 67(d)). Then, the sampled
signal f¯(t) is repeated periodically every T0 seconds (Figure 67(e)). According to the spectral sampling theorem
(explained in the preceding section), such an operation results in sampling the spectrum at a rate of T0 samples
per hertz (Figure 67(f)).
In conclusion, when a signal f (t) is sampled and periodically repeated, the corresponding spectrum is also
sampled and periodically repeated. The discrete Fourier transform (DFT) relates the samples of f (t) to the
samples of F (ω).
Number of samples: Let N0 be the number of samples of f (t) in one period T0 . We can see from Figure 67(e)
T0
that N0 = .
T
Let N0 0 be the number of samples of the spectrum in one period Fs . We can see from Figure 67(f) that
Fs
N0 0 = .
F0
1 1 T0 Fs
Since Fs = and F0 = , we have that N0 = = = N0 0 . That is,
T T0 T F0
N0 = N0 0 .

40
So, we conclude that, interestingly, the number N0 of samples of the signal in Figure 67(e) in one period T0 is
identical to the number N0 0 of samples of the spectrum in Figure 67(f) in one period Fs .
The sampled signal f¯(t) in Figure 67(c) can be expressed as:
N 0 −1

f¯(t) =
X
f (kT )δ(t − kT ).
k=0

Since δ(k − kT ) ⇐⇒ e−jkωT (prove it!), the Fourier transform of f¯(t) is:
0 −1
 NX
F̄ (ω) = F f (t) =
¯ f (kT )e−jkωT .

k=0

F (ω)
From Figure 57 and Equation (65) we have (assuming negligible aliasing) that F̄ (ω) is in the interval
T
ωs ωs
− ≤ ω ≤ . Thus,
2 2
N0 −1
X ωs
F (ω) = T F̄ (ω) = T f (kT )e−jkωT , |ω| ≤ .
k=0
2
2π
The samples of F (ω) at multiples of ω0 = are:
T0
N
X 0 −1

Fr = F (rω0 ) = T f (kT )e−jkrω0 T .

k=0

2π 2π
If we call ω0 T = Ω0 = T = , and fk = T f (kT ), then
T0 N0
N 0 −1
X 2π
Fr = fk e−jrΩ0 k , Ω0 = ω0 T = .
k=0
N0

Now, if we change the index k by m we have,

N
X 0 −1

Fr = fm e−jrΩ0 m .
m=0

Multiplying the last equation by ejrΩ0 k and summing it over r we obtain,

0 −1 0 −1
"N −1
0 −1
N N
# N
"N −1 #
X X X0 X X0

Fr ejrΩ0 k = fm e−jrΩ0 m ejrΩ0 k = fm ej(k−m)Ω0 r .

r=0 r=0 m=0 m=0 r=0

Noticing that
N 0 −1
(
X N0 when m = k,
ej(k−m)Ω0 r =
r=0
0 when m 6= k,
N
X0 −1

we thus obtain Fr ejrΩ0 k = N0 fk . Therefore,

r=0

N0 −1
1 X 2π
fk = Fr ejrΩ0 k , Ω0 = ω0 T = .
N0 r=0 N0

41
 We have thus found a relationship between the samples of f (t) (Figure 67(e)) and the samples of F (ω)
(Figure 67(f)):

Sample f (t) (and scale by T ): fk = T f (kT )

2π
Sample F (ω): Fr = F (rω0 ), ω0 =
T0

The relationship between these sampled signals is given by the equations derived in the previous page, namely:
N 0 −1
X 2π
Fr = fk e−jrΩ0 k , Ω0 = ω0 T = (72)
k=0
N0

N0 −1
1 X
fk = Fr ejrΩ0 k (73)
N0 r=0

Equation (72) defines the direct discrete Fourier transform (DFT), and equation (73) defines the
inverse discrete Fourier transform (IDFT):

fk ⇐⇒ Fr

The sequences fk and Fr are N0 –periodic (see Figures 67(e) and 67(f)), so it only makes sense to find their
values at k = 0, 1, . . . , N0 − 1 and r = 0, 1, . . . , N0 − 1.
The DFT (and IDFT) relationships are transforms in their own right and are exact. However, when we identify
fk and Fr as the samples of a continuous-time signal f (t) and its spectrum F (ω), then the DFT relationships are
approximations, because of the aliasing and leakage effects (compare Figures 67(b) and 67(f)).

Properties of the DFT

The DFT is basically the Fourier transform of a sampled signal repeated periodically. Hence, the properties
mentioned earlier for the Fourier transform apply to the DFT as well.

1. Linearity
If fk ⇐⇒ Fr and gk ⇐⇒ Gr , then,
a1 fk + a2 gk ⇐⇒ a1 Fr + a2 Gr .

2. Conjugate symmetry
From the conjugate symmetry property of the Fourier transform [recall (16)] we have that, for a real signal fk ,

F−r = Fr ∗ .

Also, since Fr is N0 –periodic, F−r = F−r+N0 = FN0 −r . Therefore,

FN0 −r = Fr ∗ . (74)

Because of this property we need to only compute half the DFT for real signals fk .

3. Time-shifting (circular shifting)

If fk ⇐⇒ Fr , then,
fk−n ⇐⇒ Fr e−jrΩ0 n .

42
4. Frequency-shifting
If fk ⇐⇒ Fr , then,
fk ejkΩ0 m ⇐⇒ Fr−m .

5. Circular convolution

For two N0 –periodic sequences fk

and gk , the circular convolution is
defined by:
N
X 0 −1

fk ~ gk = fn gk−n
n=0
N
X 0 −1

= gn fk−n .
n=0

In Figure 68, an example of cir-

cular convolution between two se- Figure 68: Example of circular convolution.
quences of 4 elements (N0 = 4) is
shown for k = 0 and k = 1. It can be observed that, at k = 0,

[fk ~ gk ]|k=0 = f0 g0 + f1 g3 + f2 g2 + f3 g1 ,

and, at k = 1,
[fk ~ gk ]|k=1 = f0 g1 + f1 g0 + f2 g3 + f3 g2 ,
and so on.

The convolution fk ~gk can be conveniently visualised

as shown in Figure 69. The inner sequence fk is clock-
wise and fixed. The outer sequence gk is inverted so
that it becomes counterclockwise. This sequence is
rotated clockwise and the overlapping terms are mul-
tiplied and added. It can be easily verified (see Fig-
ure 69) that this operation yields the exact same ex-
pressions for [fk ~ gk ]|k=0 and [fk ~ gk ]|k=1 obtained
Figure 69: Graphical picture of circular convolution.
above.
In connection with circular convolution, the DFT has the property that, if fk ⇐⇒ Fr and gk ⇐⇒ Gr , then,

fk ~ gk ⇐⇒ Fr Gr , (75)

and,
1
fk gk ⇐⇒ Fr ~ Gr . (76)
N0

The fast Fourier transform (FFT)

To compute one sample Fr from (72) we need N0 complex multiplications and N0 − 1 complex additions. To
compute the N0 values (Fr for r = 0, 1, . . . , N0 − 1) we need N0 2 complex multiplications and N0 (N0 − 1)
complex additions. The FFT is a procedure (algorithm) that reduces this number of computations. The FFT
algorithm is simplified if we choose N0 as a power of 2.

43
Define: 2π
−j N
WN0 = e 0 = e−jΩ0 . (77)
Note that,  2
2π 2π
−j N ×2 −j N
W N0 = e 0 = e 0 = WN0 2 , (78)
2

and,
N0 N0
WN0 r+( 2 ) = W ( N20 ) W r = e−j N2π0 × 2
WN0 r = e−jπ WN0 r = −WN0 r . (79)
N0 N0

Note from (72) and (77) that,

N
X 0 −1

Fr = fk WN0 kr , 0 ≤ r ≤ N0 − 1. (80)
k=0
 
N0
Now, divide the N0 –point sequence fk into two –point subsequences:
2
f0 , f2 , f4 , . . . , fN0 −2 , f1 , f3 , f5 , . . . , fN0 −1 (81)
| {z } | {z }
subsequence gk subsequence hk

Note that we can also split the sum in (80) into two sub-sums, as follows,
N0 N0
2
−1 2
−1
X X
Fr = f2k WN0 2kr + f2k+1 WN0 (2k+1)r ,
k=0 k=0

and, using (78) and (81), we can write,

N0 N0
2
−1 2
−1
X X
Fr = gk W N0 kr + WN0 r hk W N0 kr = Gr + WN0 r Hr , 0 ≤ r ≤ N0 − 1, (82)
2 2
k=0 k=0
 
N0
where Gr and Hr are the –point DFT of gk and hk , respectively.
2 
N0
Also, since Gr and Hr are –periodic we have,
2
Gr+( N0 ) = Gr , Hr+( N0 ) = Hr ,
2 2

and, using (79), we obtain

N0
Fr+( N0 ) = Gr+( N0 ) + WN0 r+( 2 )H
r+(
N0
r
) = Gr − WN0 Hr . (83)
2 2 2

N0 N0
Thus, we can compute the first points of Fr using (82) and the last points using (83). That is,
2 2
N0
Fr = Gr + WN0 r Hr , 0≤r≤ − 1,
2 (84)
r N0
Fr+( N0 ) = Gr − WN0 Hr , 0≤r≤ − 1.
2 2

In
 conclusion,
 an N0 –point DFT can be computed by combining two
N0 Figure 70: Butterfly structure.
–point DFTs as in (84). Equations (84) can be conveniently rep-
2
resented with a butterfly structure, as shown in Figure 70.
For example, for N0 = 8 (8-point DFT), the first step of the FFT algorithm consists in computing two 4-point
DFTs and then combining them according to (84) and Figure 70. This is shown in Figure 71. The next step is to
compute the 4-point DFTs, Gr and Hr , and, for this, we repeat the same procedure by dividing gk and hk into two
2-point sequences corresponding to even- and odd-numbered samples. This is shown in Figure 72.

44
 Figure 71: 8-point DFT computed from two 4-point DFTs. Figure 72: 4-point DFTs computed from
two 2-point DFTs. (Note, from (78), that
W4 = W8 2 .)

The next, and final, step is to compute the 2-point DFTs. Note
from (72) that the 2-point DFT (i.e., for N0 = 2) is
F 0 = f0 + f1 ,
F1 = f0 + f1 e−jπ = f0 − f1 ,
and, thus, multiplication in this case is not required. The compu-
tation of the 2-point DFTs is illustrated in Figure 73. Note that,
at this point we have reached the 1-point DFT (i.e., the original
sequence of time-data itself).
To compute all the N0 points of Fr (see (84)) from Gr and Hr
we require N0 complex additions and N0 /2 complex multiplica-
r
tions 
  W
(corresponding to the products N0 Hr ). To compute the
N0 N0
–point DFT Gr from the –point DFTs we require
Figure 73: 2-point DFTs computed from 1- 2 4
point DFTs (the original time-data sequence). N0 /2 complex additions and N0 /4 complex multiplications, and
the same for Hr . Hence, in the second step there are N0 complex
additions and N0 /2 complex multiplications.
Therefore, the number of computations required remains the
same at each step. Since a total of log2 N0 steps is needed to
arrive at the 1-point DFT (i.e., the original sequence), we re-
quire, conservatively, a total of N0 log2 N0 complex additions
N0
and log2 N0 complex multiplications to compute the N0 –
2
point DFT.
Recall (Page 43) that to compute the DFT from (72) we
need of the order of N0 2 computations. With the FFT we in-
stead need of the order of N0 log2 N0 computations (log2 N0 =
(log10 N0 ) / (log10 2)). The order of the number of computa-
tions required by both methods is illustrated in Figure 74, where
the advantages of the FFT algorithm can be clearly appreciated.
The procedure to obtain the IDFT is identical with Figure 74: Order of number of computations
j 2π 2
WN0 = e N0 and the additional multiplication by 1/N0 required to compute the DFT using (72) (N0 )
(see (73)). and the FFT using (84) (N0 log2 N0 ).

45
Discrete-time signals and systems
When we sample a continuous-time signal (we consider uniformly
spaced discrete instants, . . . , −2T , −T , 0, T , 2T , . . ., kT , . . .,
where T is the sampling interval) we obtain a discrete-time signal:

continuous-time signal: f (t),

discrete-time (sampled) signal: f (kT ) = f [k],

where k is an integer number. The discrete-time signal is typically

Figure 75: Discrete-time signal.
indexed by the sample number k. This is illustrated in Figure 75
where both scales are shown, the time t scale and the index k scale.
Discrete-time systems afford a convenient way of processing continuous-time signals, !"#$$ as illustrated in#,-%*)+)./0
Fig-
%&'()*+)*$
ure 76. The continuous-time signal f (t) is first sampled by an analog-to-digital converter (A/D), then the discrete--1-+)0$
time signal f [k] is processed by a discrete-time system to produce the processed discrete-time signal y[k] (for ex-
ample, a filtered version of f [k]), and finally the signal y[k] is converted to continuos-time by a digital-to-analog
converter (D/A). This constitutes a Data Acquisition System, usually abbreviated as DAQ.

!"#$$ #,-%*)+)./0)$ #"!$$

%&'()*+)*$ -1-+)0$ %&'()*+)*$

Figure 76: Processing a continuous-time signal with a discrete-time system.

Discrete-time systems have several advantages over continuous-time systems (precision, stability, duplication,
flexibility, easy to alter, use of IC technology resulting in low power consumption, etc.) and there is, thus, a trend
nowadays towards processing continuous-time signals with discrete-time systems.

Some examples of discrete-time signals !"#$$ #,-%*)+)./0)$ #"!$$

%&'()*+)*$ -1-+)0$ %&'()*+
Discrete-time impulse function δ[k]
The discrete-time impulse function (also called
unit impulse) is defined as:
(
1 for k = 0,
δ[k] = (85)
0 for k 6= 0.

The discrete-time impulse δ[k] and a delayed ver- Figure 77: Discrete-time impulse.
sion δ[k − m] are shown in Figure 77.

Discrete-time unit step function u[k]

The discrete-time unit step function u[k] is defined as:

(
1 for k ≥ 0,
u[k] = (86)
0 for k < 0,

Figure 78: Discrete-time unit step. and it is shown in Figure 78.

46
Discrete-time sinusoid
A continuous-time sinusoid cos (ωt) sampled every T seconds yields a discrete-time sinusoid,
f [k] = cos (ωkT ) = cos (Ωk),
where Ω = ωT is the frequency in radians per sample.

Peculiarities of discrete-time sinusoids

1. Not all discrete-time sinusoids are periodic
A signal cos (Ωk) is N0 –periodic if:
cos (Ωk) = cos [Ω(k + N0 )] = cos (Ωk + ΩN0 ),
which is only possible if ΩN0 = 2πm, with m integer. That is, if,
Ω m
= .
2π N0
Ω
So, cos (Ωk) is periodic only if is a rational number. See Lathi, Page 549, Figure 8.10, for interesting
2π
examples of periodic and non-periodic discrete-time sinusoids. For example (see Lathi, Figure 8.10), cos (0.8 k)
 Ω 0.8 
is not periodic note that = is irrational .
2π 2π

2. Nonuniqueness of discrete-time sinusoids

Note that
cos [(Ω ± 2πm)k] = cos (Ωk ± 2πmk) = cos (Ωk), (87)
when m is an integer. Thus, a discrete-time sinusoid has a unique waveform only for values of Ω over a range
of 2π (since a frequency Ω is undistinguishable from Ω ± 2πm). For this reason, the range −π to π is called the
fundamental range of frequencies. Furthermore, since
cos (−Ωk + θ) = cos (Ωk − θ),
we have that frequencies in the range −π to 0 can be expressed as frequencies in the range 0 to π with opposite
phase.
Consider, for example, a discrete-time phasor
ejΩk rotating counterclockwise at angular velocity Ω
rad/sample. Suppose Ω = π + x (with x < π). The pha-
sor progresses counterclockwise at velocity π +x radians
per sample as shown by the black directed arcs of Fig-
ure 79. But, we may also interpret this motion as moving
“clockwise” at the lower speed of π−x radians/sample as
shown by the red directed arcs of Figure 79 (stroboscopic
effect). [Note that a clockwise speed of π−x corresponds
Figure 79: Nonuniqueness of discrete-time sinusoids.
to a counterclockwise speed of Ω̃ = −(π − x) = −π + x
and, hence (see (87)), Ω̃ differs from Ω by −2π.]
This also explains the aliasing effect when sampling sinusoidal signals. Suppose we sample f (t) = cos (ωt)
every T seconds, resulting in the discrete-time sinusoid f [k] = cos (ωkT ) = cos (Ωk), with Ω = ωT . Discrete-
time sinusoids cos (Ωk) have unique waveforms only for values of frequencies in the range 0 ≤ Ω ≤ π, or
0 ≤ ωT ≤ π (since a sinusoid of frequency Ω > π appears as a sinusoid of a lower frequency Ω ≤ π). This
π
means that the samples of a sinusoid of frequency ω > appear as samples of a sinusoid of a lower frequency
T
π ωs π
ω ≤ . This phenomenon is called aliasing (folding back of frequencies about the folding frequency = ,
T 2 T
see Figure 63).

47
An example of a discrete-time system
Consider a regular deposit made in a bank account every month. Denote:

f [k] = deposit made at the kth discrete instant,

r = interest rate per dollar per period T (= one month),
y[k] = balance at the kth instant, computed immediately after the kth deposit is received.

We then have:
y[k] = y[k − 1] + r y[k − 1] + f [k] ,
| {z } | {z } | {z } | {z }
current balance previous balance interest deposit

or, more succinctly,

y[k] = (1 + r)y[k − 1] + f [k]. (88)
Since the choice of the index k is arbitrary, Equation (88) can also be written as,

y[k + 1] − (1 + r)y[k] = f [k + 1]. (89)

Difference equations
Equation (89) is an example of a difference equation. In general,

y[k + n] + an−1 y[k + n − 1] + . . . + a1 y[k + 1] + a0 y[k]

(90)
= bm f [k + m] + bm−1 f [k + m − 1] + . . . + b1 f [k + 1] + b0 f [k].

Causality condition
For a causal system, the output cannot depend on future input values. So, in Equation (90) we require m ≤ n .
In general, we can write a causal system as:

y[k + n] + an−1 y[k + n − 1] + . . . + a1 y[k + 1] + a0 y[k]

(91)
= bn f [k + n] + bn−1 f [k + n − 1] + . . . + b1 f [k + 1] + b0 f [k].

Iterative solution of difference equations

Contrary to continuous-time systems, discrete-time systems are very easy to solve iteratively. We can express (91)
as,
y[k] + an−1 y[k − 1] + . . . + a0 y[k − n] = bn f [k] + bn−1 f [k − 1] + . . . + b0 f [k − n],
and then as,

y[k] = −an−1 y[k − 1] − . . . − a0 y[k − n] + bn f [k] + bn−1 f [k − 1] + . . . + b0 f [k − n]. (92)

We can then solve Equation (92) recursively. For example, to determine y[0] we need the values of y[−1],
y[−2], . . . , y[−n] (the initial conditions) and the values of the input f [0], f [−1], f [−2], . . . , f [−n]. We then store
y[0] and, when the next value of the input f [1] becomes available, we can compute y[1] (from the values of y[0],
y[−1], . . . , y[−n + 1] and f [1], f [0], . . . , f [−n + 1]), and so on.

The unit impulse response h[k]

Consider a system defined by the difference equation (91). The unit impulse response h[k] is the solution of this
equation for the input δ[k] defined in (85) when all the initial conditions are equal to zero (i.e., y[−1] = y[−2] =
. . . = y[−n] = 0). Equation (91) can be solved iteratively to determine h[k], as illustrated in the following
example.

48
Example 3
Find the unit impulse response
( of the system: y[k] − 0.6 y[k − 1] − 0.16 y[k − 2] = 5 f [k].
1 for k = 0,
We let f [k] = δ[k] = and y[k] = h[k]. The difference equation then becomes:
0 for k 6= 0,
h[k] − 0.6 h[k − 1] − 0.16 h[k − 2] = 5 δ[k],
subject to zero initial conditions h[−1] = h[−2] = 0. We thus obtain:
For k = 0: h[0] = 5 δ[0] + 0.6 h[−1] + 0.16 h[−2] = 5 × 1 + 0.6 × 0 + 0.16 × 0 = 5,
For k = 1: h[1] = 5 δ[1] + 0.6 h[0] + 0.16 h[−1] = 5 × 0 + 0.6 × 5 + 0.16 × 0 = 3,
For k = 2: h[2] = 5 δ[2] + 0.6 h[1] + 0.16 h[0] = 5 × 0 + 0.6 × 3 + 0.16 × 5 = 2.6,
For k = 3: h[3] = 5 δ[3] + 0.6 h[2] + 0.16 h[1] = 5 × 0 + 0.6 × 2.6 + 0.16 × 3 = 2.04,
..
.
etc.
The following Matlab
c
script computes the first 11 points of the previous iteration.

f=[0 0 1 zeros(1,10)]; % Generate unit impulse function

H=[0 0]; % Initial conditions equal zero
for i=1:11; % Begin for loop to perform iterations
h=0.6*H(i+1)+0.16*H(i)+5*f(i+2); % Difference equation
H=[H h]; % Store impulse response
end % End for loop
stem(0:10,H(3,13)) % Plot

The values obtained with the above script are (the first two
values of H are not displayed because they correspond to the
initial conditions—equal to 0—at times k = −2 and k = −1;
so we start from the third value corresponding to time k = 0):
H(3:13)=[5 3 2.6 2.04 1.64
1.3104 1.0486 0.8388]
0.6711 0.5369 0.4295]
The plot of the impulse response is shown in Figure 80.
Figure 80: Example of impulse response.
Operational notation
In difference equations it is convenient to use operational notation. We use the operator E to denote the operation
of advancing a sequence by one time unit:
E f [k] = f [k + 1],
E 2 f [k] = f [k + 2],
..
.
n
E f [k] = f [k + n].
Using this notation, the difference equation (91) can be written as
E n + an−1 E n−1 + . . . + a1 E + a0 y[k] = bn E n + bn−1 E n−1 + . . . + b1 E + b0 f [k],



(93)
| {z } | {z }
Q[E] P [E]

49
or,
Q[E] y[k] = P [E] f [k], (94)
where Q[E] and P [E] are nth order polynomials of the operator E.

System response to initial conditions (zero-input response)

The zero-input response y0 [k] is the solution of (94) with f [k] = 0, that is,

Q[E] y0 [k] = 0, (95a)

or,
y0 [k + n] + an−1 y0 [k + n − 1] + . . . + a1 y0 [k + 1] + a0 y0 [k] = 0. (95b)
To make this equation equal to zero for all values of k, the sequence y0 [k] and its advanced versions have to
have the same form. The function that has this property is the exponential function γ k , since γ k+m = γ m γ k , so
that, advanced γ k is equal to γ k scaled by a constant γ m . Hence, the solution must be of the form y0 [k] = c γ k .
Substituting in (95) we obtain:

c γ n + an−1 γ n−1 + . . . + a1 γ + a0 γ k = 0.


| {z }
Q[γ]

Thus, for a nontrivial solution we must have:

Q[γ] = γ n + an−1 γ n−1 + . . . + a1 γ + a0 = 0. (96)

There are three possible scenarios when computing the solutions of Equation (96) i.e., the roots of Q[γ] . They
are covered by the following three cases.

Case 1: Distinct real roots

Factorising Q[γ] when it has distinct roots:

Q[γ] = (γ − γ1 ) (γ − γ2 ) . . . (γ − γn ) = 0.

Therefore, (95) has n possible solutions: c1 γ1 k , c2 γ2 k , . . . , cn γn k , and the general solution is:

y0 [k] = c1 γ1 k + c2 γ2 k + . . . + cn γn k , (97)

where γ1 , γ2 , . . . , γn are the solutions of (96) and c1 , c2 , . . . , cn are arbitrary constants obtained from n auxiliary
conditions (usually, the initial conditions). We have the following commonly used terminology:

Q[γ] is called the characteristic polynomial of the system,

Q[γ] = 0 is called the characteristic equation of the system,
γ1 , γ2 , . . . , γn are called the characteristic values (or characteristic roots) of the system,
k k k
γ1 , γ2 , . . . , γn are called the characteristic modes of the system.

Case 2: Repeated roots

If two or more roots coincide, the form of the characteristic modes is modified. Consider, for example,
Q[E] = (E − γ1 )3 = E 3 − 3 γ1 E 2 + 3 γ1 2 E − γ1 3 . Let us consider first the mode y0 [k] = γ1 k in (95):

Q[E] y0 [k] = E 3 − 3 γ1 E 2 + 3 γ1 2 E − γ1 3 γ1 k = γ::::::

3 k 3 k 3 k 3 k


1 γ1 − 3γ1 γ1 + 3γ1 γ1 − γ 1 γ1 = 0,
::::::

50
so, γ1 k is one solution. Consider now the mode y0 [k] = k γ1 k in (95):
Q[E] y0 [k] = E 3 − 3 γ1 E 2 + 3 γ1 2 E − γ1 3 k γ1 k

= (k + 3) γ1 k+3 − 3 γ1 (k + 2) γ1 k+2 + 3 γ1 2 (k + 1) γ1 k+1 − γ1 3 k γ1 k

 
= k + 3 − 3k ::
− 6 + 3k
::
+ 3 − k γ1 k+3 = 0,

so, k γ1 k is also a solution! Let us try next (feeling lucky . . . ) with y0 [k] = k 2 γ1 k in (95):
Q[E] y0 [k] = E 3 − 3 γ1 E 2 + 3 γ1 2 E − γ1 3 k 2 γ1 k

= (k + 3)2 γ1 k+3 − 3 γ1 (k + 2)2 γ1 k+2 + 3 γ1 2 (k + 1)2 γ1 k+1 − γ1 3 k 2 γ1 k

 
= k 2 + 6k
::
+ 9 − 3k 2
− 12k
:::
− 12 + 3k 2
+ 6k
::
+ 3 − k 2
γ1 k+3 = 0.

It works too! Let us then try with y0 [k] = k 3 γ1 k in (95):

Q[E] y0 [k] = E 3 − 3 γ1 E 2 + 3 γ1 2 E − γ1 3 k 3 γ1 k

= (k + 3)3 γ1 k+3 − 3 γ1 (k + 2)3 γ1 k+2 + 3 γ1 2 (k + 1)3 γ1 k+1 − γ1 3 k 3 γ1 k

 
= k 3 + 9k 2 + 27k + 27 − 3k 3
:::
− 18k 2
− 36K − 24 + 3k
:::
3
+ 9k 2
+ 9k + 3 − k 3
γ1 k+3
= 6 γ1 k+3 6= 0,
so, k 3 γ1 k is not a solution. (Phew! The computations were becoming quite messy . . . )
Thus, in general, for a root γ1 repeated r times, the characteristic modes are γ1 k , k γ1 k , . . . , k r−1 γ1 k . There-
fore, if the characteristic polynomial of a system is
Q[γ] = (γ − γ1 )r (γ − γr+1 ) . . . (γ − γn ),
the zero-input response is given by:

y0 [k] = c1 + c2 k + c3 k 2 + . . . + cr k r−1 γ1 k + cr+1 γr+1 k + cr+2 γr+2 k + . . . + cn γn k .

(98)

Case 3: Complex roots

Complex roots occur as pairs of conjugates, γ and γ ∗ , such that,
γ = |γ|ejβ , γ ∗ = |γ|e−jβ .
The zero-input response is: y0 [k] = c1 γ k + c2 (γ ∗ )k . For a real system, c1 and c2 are also conjugate, so that y0 [k]
is a real function:
c c
c1 = ejθ , c2 = c1 ∗ = e−jθ .
2 2
Thus, we have:
c
y0 [k] = |γ|k ej(βk+θ) + e−j(βk+θ) ,
 
2
that is,
y0 [k] = c|γ|k cos (βk + θ), (99)
where c and θ are arbitrary constants determined from the auxiliary conditions.

System stability
A system is asymptotically stable if the zero-input response approaches zero as k → ∞.
A system is marginally stable if the zero-input response neither approaches zero nor grows without bound as
k → ∞.
A system is unstable if the zero-input response grows without bound as k → ∞.

51
Consider the following facts: Im
k
if |γ| < 1, γ → 0 as k → ∞,
Marginally stable Unstable
if |γ| > 1, |γ|k → ∞ as k → ∞,
if |γ| = 1, |γ|k = 1 for all k. Stable

It is then clear from (97), (98) and (99) that a system Re

is asymptotically stable if and only if

|γi | < 1, i = 1, 2, . . . , n. (100)

If one (or more) characteristic value lies outside the

unit circle, then the system is unstable. In general, we Figure 81: Stability according to the location of the char-
can characterise the stability region of the characteris- acteristic roots.
tic values on the complex plane as shown in Figure 81.
The stability of a discrete-time system can be determined as follows:
1. The system is asymptotically stable if and only if all the characteristic roots are inside the unit circle. The
roots may be simple or repeated.
2. The system is unstable if and only if either one or both of the following conditions exist: (a) At least one
root is outside the unit circle; (b) There are repeated roots on the unit circle.
3. The system is marginally stable if and only if there are no roots outside the unit circle and there are some
unrepeated roots on the unit circle.

System response to an external input (zero-initial-condition response)

We shall assume that the initial conditions of the system are zero
and evaluate the system response to an external input f [k]. A sig-
nal f [k] can be expressed as a sum of impulse components [see the
definition of the unit impulse in (85)], as illustrated in Figure 82:
f [k] = . . . + f [−2]δ[k + 2] + f [−1]δ[k + 1]
+ f [0]δ[k] + f [1]δ[k − 1]
+ f [2]δ[k − 2] + . . . (101)
X∞
= f [m]δ[k − m].
m=−∞

If we know the unit impulse response h[k] (see Page 48), i.e.,

δ[k] −→ h[k],

then, the response to a delayed impulse is:

δ[k − m] −→ h[k − m].

Hence, using the principle of superposition on (101) we obtain:

∞
X ∞
X
f [k] = f [m]δ[k − m] −→ y[k] = f [m]h[k − m].
m=−∞ m=−∞

Therefore,
∞
X
Figure 82: Signal f [k] represented in terms y[k] = f [m]h[k − m]. (102)
m=−∞
of unit impulse components.

52
The convolution sum
The summation in (102) is the convolution sum of f [k] and h[k], denoted as:
X∞
f [k] ∗ h[k] = f [m]h[k − m].
m=−∞

Properties of the convolution sum

The properties are similar to those of the convolution integral (Page 6). In particular, the commutative, distributive,
associative, and the shift properties are identical (with the obvious change of (t) by [k], etc.). Also, we have these
two important properties:
1. Convolution with an impulse: f [k] ∗ δ[k] = f [k] (see Equation (101) above).
2. Width property: The width of a signal is one less than the number of its elements (its length). For example,
the signal g[k] shown in Figure 83 has 3 elements (length = 3), but has a width of only 2. We then have
the following property: if f1 [k] has m elements (width: W1 = m − 1) and f2 [k] has n elements (width:
W2 = n − 1), then f1 [k] ∗ f2 [k] has m + n − 1 elements (width: W1 + W2 = m + n − 2). Thus, the width
of the convolution sum is the sum of the widths of each of the individual signals.

Causality
In general, we consider inputs to a system that start at k = 0 and are zero before that,
f [k] = 0 for k < 0.
If the system is causal, then h[k] = 0 for k < 0. In this case, for such f [k] and h[k], Equation (102) reduces to:
k
X
y[k] = f [m]h[k − m].
m=0

Computation of the convolution sum

The convolution sum can be computed graphically or using the sliding tape method (they are basically the same).
To compute the convolution of f [k] and g[k], given by,
∞
X
c[k] = f [k] ∗ g[k] = f [m]g[k − m],
m=−∞

we follow these steps:

1. Plot f [m] and g[m] as functions of m;
2. Flip g[m] about the vertical axis (about m = 0) to obtain g[−m];
3. Time shift (slide) g[−m] by k units to obtain g[k − m] (for k > 0 the shift is to the right , for k < 0 it is to
the left);
4. Multiply (element by element) f [m] and g[k − m] and add all the products to obtain c[k];
5. Repeat for each k in the range −∞ to ∞.
In the sliding tape method, instead of plotting the functions, they are displayed as a sequence of numbers.

Example 4
In Figure 83, two discrete-time functions f [k] and g[k] are shown together with the steps of the sliding tape
method and the final result c[k] = f [k] ∗ g[k]. The same convolution of these two sequences can be computed in
Matlab c
with the following command:
» conv([0 1 2 3 2 1],[1 1 1])
resulting in:
» ans = 0 1 3 6 7 6 3 1

53
 0 1 2 3 2 1

1 1 1

0 1 2 3 2 1

1 1 1

0 1 2 3 2 1

0 1 2 3 2 1 1 1 1

1 1 1
0 1 2 3 2 1

Rotate g about the vertical axis 1 1 1

0 1 2 3 2 1

1 1 1

0 1 2 3 2 1

1 1 1

0 1 2 3 2 1

1 1 1

0 1 2 3 2 1

1 1 1

0 1 2 3 2 1

1 1 1

Figure 83: Example of discrete-time convolution using the sliding tape method.

Discrete-time filtering (convolution) using the DFT

Because of the computational efficiency of the FFT algorithm explained on Page 45 (see Figure 74), the DFT is
particularly appealing for applications such as filtering, convolution and correlation. For example, as we have
seen in (75), the circular convolution of two sequences f [k] and g[k] (fk and gk in (75) denote the same thing as
f [k] and g[k], respectively) corresponds to the product of the two DFTs:
f [k] ~ g[k] ⇐⇒ Fr Gr . (103)
Recall that the circular convolution (see Property 5 on Page 43) is defined as:
N
X0 −1

f [k] ~ g[k] = f [m] g[k − m],

m=0

where the sequences f [k] and g[k] are N0 –periodic. Note that the circular convolution differs from the regular
(linear) convolution (102) by the facts that the summation is over one period and both sequences are N0 –periodic
(see, e.g., Figure 68), whereas in the linear convolution the summation is from −∞ to ∞ and the sequences are
not periodic (see, e.g., Figure 83). Fortunately, linear convolution can be made equivalent to circular convolution
by padding both sequences with zeros.
Suppose that we want to compute the linear convolution (102) of two finite length sequences f [k] and h[k] of
lengths (= numbers of elements) Nf and Nh , respectively. The linear convolution of these two sequences,
y[k] = f [k] ∗ h[k],
has length N0 = Nf + Nh − 1 (see the width property of the convolution sum on the previous page). Note that if
we pad N0 − Nf (= Nh − 1) zeros at the end of f [k] and N0 − Nh (= Nf − 1) zeros at the end of h[k], then both

54
sequences have now N0 elements and the first period of the circular convolution of the zero-padded f [k] and h[k]
is identical to their linear convolution (since the products of the parts that overlap due to the periodic repetitions
of f [k] and h[k] in the circular convolution are always equal to zero—for example, this can be seen in Figure 68,
if we pad 3 zeros at the end of both, f [k] and g[k], and now perform the circular convolution of the resulting
7-point sequences). So, with the sequences f [k] and h[k] zero-padded conveniently (so that both have now N0
elements) we have that y[k] = f [k] ~ h[k] is an N0 –periodic sequence (since it is the circular convolution of
two N0 –periodic sequences) whose first period is the linear convolution of (unpadded) f [k] and h[k]. According
to (103), the DFT of y[k] = f [k] ~ h[k] is given by:

Yr = Fr Hr .

In conclusion, the procedure to compute the linear convolution (102) of two finite length sequences f [k] and h[k]
of lengths Nf and Nh is:
1. Pad Nh − 1 zeros to f [k] and Nf − 1 zeros to h[k].
2. Find Fr and Hr , the DFTs of the zero-padded sequences f [k] and h[k].
3. Multiply Fr by Hr to obtain Yr .
4. The desired convolution y[k] is the IDFT of Yr .

Example 4 revisited
The following Matlab
c
script computes the convolution calculated in Example 4 above (see Figure 83) using the
DFT computed with the FFT algorithm and their inverse counterparts.

f=[0 1 2 3 2 1]; % Form sequence f[k]

g=[1 1 1]; % Form sequence g[k]
N0 = length(f) + length(g) -1; % Compute N0 = Nf + Ng - 1
F=fft(f,N0); % N0-point FFT of f[k] padded with zeros to make its length equal to N0
G=fft(g,N0); % N0-point FFT of g[k] padded with zeros to make its length equal to N0
C=F.*G; % DFT of c[k] is the componentwise product of the DFTs of f[k] and g[k]
c=ifft(C); % Signal c[k] is the IDFT of Cr

The result obtained with the previous Matlab

c
script is:
» c = 0 1.0000 3.0000 6.0000 7.0000 6.0000 3.0000 1.0000

For small length sequences, the direct convolution method (such as the sliding tape method) is faster than the
DFT method. However, for long sequences the DFT method using the FFT algorithm is much faster and far more
efficient. This is due to the fact that the use of the FFT algorithm to compute the DFT reduces the number of
computations dramatically, especially for large N0 (see Figure 74). The method of computing the convolution
using the FFT is known as fast convolution.

Total response
To find the total response of a linear time invariant (LTI) system, we can exploit the linearity of the system and use
superposition. The zero-input response y0 [k], due to the initial conditions of the system, is computed from (97),
(98) and (99), and satisfies Equation (95a) above, that is,

Q[E] y0 [k] = 0. (95a)

The zero-initial-condition response, due to the external input f [k] let us denote it here as yf [k] , is computed
∞
X
from (102), yf [k] = f [k] ∗ h[k] = f [m]h[k − m], and satisfies (94) above, that is,
m=−∞

Q[E] yf [k] = P [E] f [k]. (94b)

55
 We can then see that the total response is the sum of the zero-input and zero-initial-condition responses,
y[k] = y0 [k] + yf [k], since adding (94b) and (95a) we obtain:

Q[E] y[k] = Q[E] (y0 [k] + yf [k]) = P [E] f [k],

and, hence, y[k] satisfies the difference equation (94). In conclusion,

Total response y[k] = Expressions (97), (98) and (99) + f [k] ∗ h[k]
| {z } | {z }
zero-input component y0 [k] zero-initial-condition component yf [k]

Overview of the Z-transform

Similarly as with the Laplace transform before, we will not cover the Z-transform in detail in this course (there is
an entire chapter in Lathi’s textbook, Chapter 11, dedicated to the Z-transform). We will only cover the essentials,
so as to be able to write and understand the transfer functions of digital filters.
Formally, the Z-transform of a discrete-time signal f [k] is defined as:

∞
F [z] = Z {f [k]} =
X
f [k]z −k , (104)
k=−∞

and we write f [k] ⇐⇒ F [z].

Key properties
Linearity
If f1 [k] ⇐⇒ F1 [z] and f2 [k] ⇐⇒ F2 [z] then, for any constants a1 and a2 ,

a1 f1 [k] + a2 f2 [k] ⇐⇒ a1 F1 [z] + a2 F2 [z].

Right shift (delay)

If f [k] u[k] ⇐⇒ F [z], then
1
f [k − 1] u[k] ⇐⇒ F [z] + f [−1],
z
where u[k] is the discrete-time unit step function defined in (86) and f [k − 1] is the sequence f [k] delayed
(right-shifted) by one sample.
In general,
m
1 1 X
f [k − m] u[k] ⇐⇒ m F [z] + m f [−k]z k . (105)
z z k=1

Transfer function of a discrete-time LTI system

A discrete-time linear time invariant (LTI) system (91) can be expressed equivalently in the delay operator form
(see Page 48):

y[k] + an−1 y[k − 1] + . . . + a0 y[k − n] = bn f [k] + bn−1 f [k − 1] + . . . + b0 f [k − n]. (106)

We assume that all the initial conditions are zero

y[−1] = y[−2] = . . . = y[−n] = 0, (107)

56
and that the input starts at time k = 0 and is zero before that. In particular,
f [−1] = f [−2] = . . . = f [−n] = 0. (108)
From (105), (107) and (108) we have, for m = 1, 2, . . . , n,
1
y[k − m] u[k] ⇐⇒ Y [z],
zm
1
f [k − m] u[k] ⇐⇒ m F [z],
z
and, using the linearity property of the Z-transform (see the previous page), we have that the Z-transform of
Equation (106) is given by
 
 an−1 a0  bn−1 b0
1+ + . . . + n Y [z] = bn + + . . . + n F [z].
z z z z
Rearranging the previous equation we obtain,
bn z n + bn−1 z n−1 + . . . + b0
Y [z] = F [z].
z n + an−1 z n−1 + . . . + a0
| {z }
H[z]

Therefore, we conclude that

Y [z] = H[z]F [z]. (109)

H[z] in (109) is the system’s transfer function, given by:

!"#$%& '('%)*& +$%#$%&
P [z] bn z n + bn−1 z n−1 + . . . + b1 z + b0
H[z] = = n . (110) Figure 84: Block diagram representation.
Q[z] z + an−1 z n−1 + . . . + a1 z + a0
The block diagram representation of relationship (109) for a discrete-time linear system is shown in Figure 84.
,-.&& .!'/1)%)23*)&
Note that P [z] and Q[z] in (110) are the same polynomials that we defined in (93)–(94) and, in particular,
/+"0)1%)1& &4('%)*&&&&&&&&&&
the denominator of the transfer function H[z] in (110), Q[z], is the characteristic polynomial of the system (see


Page 50). Therefore, the poles of H[z] i.e., the roots of the denominator Q[z] are the characteristic roots of the
system. Hence, the stability of the system can be determined by analysing the location of the poles of the system
in exactly the same way as it was done for the characteristic roots on Page 52 (in particular, the stability region is
the interior of the unit circle, see Figure 81).

Exercise 5
Show, using (104), that the Z-transform of the discrete-time impulse function δ[k] defined in (85) is
Z {δ[k]} = 1.
The result of Exercise 5 and Equation (109) tell us that the transfer function H[z] is the Z-transform of the
impulse response h[k] of the system since, when f [k] = δ[k], F [k] = 1 and, hence, Y [z] = H[z]F [z] = H[z] ×


1 = H[z] .

Frequency response of a discrete-time system

If we apply an exponential input f [k] = z k to a system with impulse response h[k] the output of the system,
according to (102), is
X∞ X∞
k
y[k] = h[k] ∗ z = h[m]z k−m
=z k
h[m]z −m .
m=−∞ m=−∞
| {z }
H[z]

57
According to (104), the summation on the right-hand term above is the Z-transform of the impulse response
h[k], i.e., the transfer function H[z] of the system (see the previous conclusion, drawn after Exercise 5). Hence,
y[k] = H[k]z k , which can be denoted by a directed arrow representation:
k k
| z{z } −→ | H[z]z
{z }
.
input output

In the particular case when z = e±jΩ we obtain

ejΩk −→ H ejΩ ejΩk ,

 

e−jΩk −→ H e−jΩ e−jΩk ,

 

 jΩ    jΩ  j H ejΩ   −jΩ    jΩ  −j H ejΩ 

where H e = H e
  e and H e = H e  e .
If we apply to the input of the system a discrete-time sinusoidal signal,

f [k] = cos (Ωk + θ),

 jθ   −jθ 
1  j(Ωk+θ) −j(Ωk+θ) e e
jΩk
e−jΩk , then ap-

which can be written as f [k] = cos (Ωk + θ) = e +e = e +
2 2 2
plying the principle of superposition we have that the output is:
 jθ   −jθ 
e e
H e−jΩ e−jΩk
 jΩ  jΩk  
y[k] = H e e +
2 2
 
1 j(Ωk+θ)   jΩ  j H ejΩ   jΩ  −j H ejΩ 
 
−j(Ωk+θ) 
= e H e e +e H e e
2
"       #
  jΩ  1 j Ωk+θ+ H ejΩ −j Ωk+θ+ H ejΩ
= H e  e +e
2
= H ejΩ  cos Ωk + θ + H ejΩ ,
    

and we conclude that the output to a discrete-time sinusoidal input f [k] = cos (Ωk + θ) is given by:

y[k] = H ejΩ  cos Ωk + θ + H ejΩ .

    

(111)

This result only applies to asymptotically stable systems since the Z-transform we used in its derivation,
X∞
H[z] = h[m]z −m , is valid only for values of z lying in the region of convergence of H[z]. For z = ejΩ , z
m=−∞
lies on the unit circle (|z| = 1) and thus it is not included in the region of convergence for unstable and marginally
stable systems.
Equation (111) says that the response of an asymptotically stable linear discrete-time system to a discrete-
time sinusoidal input of frequency
 jΩ  Ω is also a discrete-time sinusoid of the same frequency. The amplitude of the
 jΩ
output sinusoid is H e
  times the amplitude of the input, and the phase of the output is shifted by H e
 jΩ
with respect
  jΩ  to the input’s phase. Therefore, H e encompassing the information of both, amplitude gain
H e  and phase shift H ejΩ is the frequency response of the system.
 

The periodic nature of the frequency response of discrete-time systems

Unlike the frequency response of a continuous-time system, the frequency response of every discrete-time system
is a periodic function of Ω with period 2π. This simply results from the fact that the argument ejΩ of H ejΩ
 

is 2π–periodic. The physical reason for this periodicity is that, as explained on Page 47, discrete-time sinusoids
separated by values of Ω in integral multiples of 2π are identical. Therefore, the system response to such sinusoids
(or exponentials) is also identical. Thus, for discrete-time systems, we need to only plot the frequency response
over the frequency range from −π to π (or from 0 to 2π).

58
Digital filters
Digital filters can be classified as either recursive (or IIR) filters or nonrecursive (or FIR) filters.

Recursive (IIR) filters

Consider for example the transfer function

b 3 z 3 + b2 z 2 + b1 z + b0
H[z] = . (112)
z 3 + a2 z 2 + a1 z + a0
Working backwards from the transfer function (110) to the difference equation (106) we can see that the input
sequence f [k] and the corresponding output sequence y[k] of this system are related by

y[k] = −a2 y[k − 1] − a1 y[k − 2] − a0 y[k − 3] + b3 f [k] + b2 f [k − 1] + b1 f [k − 2] + b0 f [k − 3] . (113)

| {z } | {z }
output terms input terms

The output is therefore determined iteratively (or recursively) from its past values. If we apply an impulse input
δ[k], the impulse response h[k] will continue forever (it propagates itself because of the recursive nature of the
filter) up to k → ∞. For this reason, these filters are also called infinite impulse response (IIR) filters.

Nonrecursive (FIR) filters

If a2 = a1 = a0 = 0 in (112), the transfer function reduces to

b 3 z 3 + b2 z 2 + b1 z + b0 b2 b1 b0
H[z] = 3
= b3 + + 2 + 3 , (114)
z z z z
and the difference equation (113) reduces to

y[k] = b3 f [k] + b2 f [k − 1] + b1 f [k − 2] + b0 f [k − 3] .
| {z }
input terms

Note that y[k] is now computed only from the (present and past) values of the input f [k] (i.e., there is no recur-
sion). If we apply an impulse input f [k] = δ[k] to this system, the impulse response will be

h[0] = b3 δ[0] + b2 δ[−1] + b1 δ[−2] + b0 δ[−3] = b3 × 1 + b2 × 0 + b1 × 0 + b0 × 0 = b3 ,

h[1] = b3 δ[1] + b2 δ[0] + b1 δ[−1] + b0 δ[−2] = b3 × 0 + b2 × 1 + b1 × 0 + b0 × 0 = b2 ,
h[2] = b3 δ[2] + b2 δ[1] + b1 δ[0] + b0 δ[−1] = b3 × 0 + b2 × 0 + b1 × 1 + b0 × 0 = b1 ,
h[3] = b3 δ[3] + b2 δ[2] + b1 δ[1] + b0 δ[0] = b3 × 0 + b2 × 0 + b1 × 0 + b0 × 1 = b0 ,
h[4] = b3 δ[4] + b2 δ[3] + b1 δ[2] + b0 δ[1] = b3 × 0 + b2 × 0 + b1 × 0 + b0 × 0 = 0,
..
.
h[k] = 0, for k ≥ 4.

We can see that the impulse input will “pass through” the system and will be “completely out of the system” by
time instant k = 4. Therefore, the duration of of the impulse response h[k] of the filter is finite. For this reason,
these filters are also known as finite impulse response (FIR) filters.
In general, we can identify the coefficients

of the filter with the impulse response values as done before, e.g.,
b3 = h[0], b2 = h[1], . . . , b0 = h[3] and a generic nth order FIR filter impulse response can be expressed as

h[k] = h[0]δ[k] + h[1]δ[k − 1] + h[2]δ[k − 2] + . . . + h[n]δ[k − n]. (115)

The transfer function, H[z], is the Z-transform of h[k] (see Page 57). Hence, applying (104) to (115) we obtain

59
 ∞
h[1] h[2] h[n]
H[z] = Z {h[k]} =
X
h[k]z −k = h[0] + + 2 + ... + n
z z z
k=−∞ (116)
h[0]z n + h[1]z n−1 + h[2]z n−2 + . . . + h[n]
= ,
zn
and the frequency response (see Page 58), H ejΩ = H ejωT , is
   

H ejωT = h[0] + h[1]e−jωT + h[2]e−j2ωT + . . . + h[n]e−jnωT .

 
(117)
A note about frequencies: Recall that on Page 47 we considered a continuous-time sinusoid cos (ωt) with fre-
quency ω given in radians per second. When we sampled this sinusoidal signal every T seconds we obtained
a discrete-time sinusoid f [k] = cos (ωkT ) = cos (Ωk), where the frequency is now Ω = ωT and is measured
in radians per sample. When we deal with purely discrete-time signals we normally use Ω as the frequency,
however when we want to relate to the “outside-world” continuous-time signals we usually are interested in the
frequency ω. (See, e.g., Figure 87, where the “outside-world” continuous-time signals there would be f (t) and
y(t), whereas the discrete-time signals “inside our digital processor” are f [k], y[k], etc.) As we have seen, these
two frequencies are simply related by Ω = ωT (i.e., a simple rescaling of frequencies). This relationship was
also present in the DFT, as can, e.g., be seen in Equation (72).
Nonrecursive filters are a special case of recursive filters (e.g., the nonrecursive filter (114) given above is
a particular case of the recursive filter (112) with a2 = a1 = a0 = 0). Hence, we expect the performance
of recursive filters to be superior. This expectation is true in the sense that a given amplitude response can be
achieved by a recursive filter of an order smaller than that required for the corresponding nonrecursive filter.
Thus, recursive filters are in general of lower order (faster) and require less memory. Nonrecursive filters, on
the other hand, can be designed to have an arbitrarily shaped frequency response and also have the important
advantage that they can be designed (under some symmetry conditions, as we will see next) to achieve a linear
phase characteristic. Recursive filters can realise linear phase only approximately. So, if processing delay is not
critical, nonrecursive filters are the obvious choice.

Symmetry conditions for linear phase response in FIR filters

Consider, for example, n = 4 in the frequency response (117),
Prototype lowpass filter
H ejωT = h[0] +Bandpass
h[1]e−jωTfilter
+ h[2]e−j2ωT + h[3]e−j3ωT + h[4]e−j4ωT
 
Bandpass filter Prototype lowpass filter
= e−j2ωT h[0]ej2ωT + h[1]ejωT + h[2] + h[3]e−jωT + h[4]e−j2ωT .

Figure 85: Symmetry condition for linear Figure 86: Antisymmetry condition for linear
phase response. phase response.

Suppose now that the impulse response h[k] is symmetric about its center point. That is, h[0] = h[4] and
h[1] = h[3] (see Figure 85). We then have,
H ejωT = e−j2ωT h[0] ej2ωT + e−j2ωT + h[2] + h[1] ejωT + e−jωT
     

= e−j2ωT (h[2] + 2h[1] cos ωT + 2h[0] cos 2ωT ) .

60
The quantity inside the parentheses
 jωT  in the last expression is real
(there is no j term whatsoever), and represents
−j2ωT
the amplitude response H e
  since e has amplitude 1 . The phase response is:

H ejωT = −2ωT,
 

and, hence, it is a linear function of ω.

If h[k] is antisymmetric about its center point, i.e., h[0] = −h[4], h[1] = −h[3] and h[2] = 0 (see Figure 86),
then
H ejωT = e−j2ωT h[0] ej2ωT − e−j2ωT + h[1] ejωT − e−jωT
     

= 2je−j2ωT (h[1] sin ωT + h[0] sin 2ωT )

= 2ej ( 2 −2ωT ) (h[1] sin ωT + h[0] sin 2ωT ) .
π

Hence, the phase response is:

 π
H ejωT = − 2ωT.

2
We thus conclude that if the impulse response h[k] of an FIR filter is either symmetric or antisymmetric about its
center point, the filter phase response is a linear function of ω. (Recall the discussion on Page 15 on the advan-
tages of a linear phase response in relation to distortionless transmission through a channel.)

Digital filter design

!"#$%&
Consider the analog filter realisation '('%)*&
with a digital+$%#$%&
filter represented in Figure 87. Our objective is to determine

,-.&& .!'/1)%)23*)& .-,&&

/+"0)1%)1& &4('%)*&&&&&&&&&& /+"0)1%)1&

Figure 87: Analog filter realisation with a digital filter.

the digital processor H[z] that will make the system on the top part of Figure 87 “equivalent” to a desired analog
system with transfer function Ha (s). We can aim at making the two systems behave similarly in the time domain
or in the frequency domain.

The time-domain equivalence criterion

In this criterion we want that, for the same input f (t), the outputs y(t) of the two systemsZof Figure 87 be equal
∞
(or as close as possible). The output of the system Ha (s) [see Equation (12)] is y(t) = f (τ )ha (t − τ )dτ ,
−∞
where ha (t) is the impulse response of the system (i.e., the inverse Laplace transform of the transfer function
Ha (s), see Page 22). The integral can be expressed as the limit of a sum:
∞
X
y(t) = lim f (m∆τ )ha (t − m∆τ )∆τ.
∆τ →0
m=−∞

It is convenient here to use the notation T for ∆τ , thus we obtain,

61
 ∞
X
y(t) = lim T f (mT )ha (t − mT ),
T →0
m=−∞

and the response at the kth sampling instant is then:

∞
X
y(kT ) = lim T f (mT )ha ((k − m)T ). (118)
T →0
m=−∞

The output of the system H[z] [see Equation (102)] is:

∞
X
y[k] = f [m]h[k − m], (119)
m=−∞

where h[k] is the impulse response of the digital filter (i.e., the inverse Z-transform of the transfer function H[z],
see Page 57).
For the two systems to be equivalent we require y(kT ) in (118) to be equal to y[k] in (119). Therefore, the
time-domain criterion for equivalence of the two systems is:

h[k] = lim T ha (kT ), (120)

T →0

that is, h[k], the unit impulse response of system H[z] on the top of Figure 87 must be equal to T times the
samples of ha (t), the unit impulse response of the system Ha (s) on the bottom part of Figure 87, assuming that
T → 0. For this reason, this method is known as the impulse invariance criterion of filter design.

The frequency-domain equivalence criterion

On Page 10 we obtained that the response of a system Ha (s) to an exponential function ejωt is equal to y(t) =
Ha (ω)ejωt . By an identical derivation the response to an exponential function f (t) = est can be found to be equal
to y(t) = Ha (s)est , and hence the response at the kth sampling instant is
y(kT ) = Ha (s)eskT . (121)
On Page 58 we obtained that the response of a discrete-time system H[z] to an exponential input f [k] = z k is
given by y[k] = H[z]z k . If the input to this system are the samples of the exponential function f (t) = est , then

k
f [k] = eskT = esT = z k , with z = esT , and the response of the system is:

k
y[k] = H[z]z k  = H esT esT = H esT eskT .
   
(122)

z=esT

For the two systems of Figure 87 to be equivalent we require y(kT ) in (121) to be equal to y[k] in (122). Thus,
the frequency-domain criterion for equivalence of the two systems is that H esT = Ha (s).
With this criterion we only ensure that the digital filter’s response matches exactly that of the desired analog
filter at the sampling instants. If we want the two responses to match at every value of t we must have T → 0.
Therefore,
lim H esT = Ha (s).
 
(123)
T →0

Recursive filter design by the time-domain criterion: The impulse invari-

ance method
In practice, T in (120) cannot be made T → 0 (that would require an infinite sampling rate). Provided T is chosen
small enough to make aliasing not important (see Page 37) we can ignore the condition T → 0 and Equation (120)
can be expressed as:
h[k] = T ha (kT ). (124)

62
We then take the Z-transform of this equation:
H[z] = T Z {ha (kT )} , (125)
and this yields the desired discrete-time transfer function H[z].
A systematic procedure to find H[z] given an analog filter desired transfer function Ha (s) consists in ex-
panding Ha (s) in partial fractions, finding the individual impulse responses of each term in the partial fractions
expansion by the inverse Laplace transform (since the Laplace transform of the impulse response is equal to the
transfer function, see Page 22) and then compute the Z-transform of each term sampled (t = kT ) and add all
of them together multiplied by T . The procedure can be simplified by the use of tables (see, e.g., Table 12.1 on
Page 736 of Lathi’s textbook). Also, the Matlab c
function impinvar solves this problem. The input data are
the coefficients of the numerator and denominator polynomials of Ha s and the sampling interval T . Matlab c

function impinvar returns the numerator and the denominator polynomial coefficients of the desired digital fil-
ter H[z] (beware though that there is a scaling factor T discrepancy between Matlab c
’s solution and that of (125),
see Lathi’s book for the details).

Limitations of the impulse invariance method

Since the impulse invariance method involves sampling the impulse response of an analog filter [see (124)] the
spectrum will consist of periodic repetitions of Ha (jω) with period equal to the sampling frequency ωs = 2π/T
(see Figure 57). Since Ha (jω) cannot be made bandlimited (by the Paley-Wiener criterion, see Page 16), aliasing
among the various repeating cycles cannot be prevented. Consequently, the method can only be used to design
filters where Ha (jω) becomes negligible beyond some frequency B Hz. The impulse invariance method cannot
be used for highpass or bandstop filters. In general, the frequency-domain method, discussed next, is superior.

Recursive filter design by the frequency-domain criterion: The bilinear

transformation method
This method requires a lower sampling rate compared to the impulse invariance method because of the absence of
aliasing and the filter rolloff
 characteristics obtained are sharper. The frequency-domain design criterion is given
by (123), that is, lim H esT = Ha (s). Consider the following approximation, as T → 0,
T →0
  sT
2 e −1
→ s,
T esT + 1
which can be verified by applying L’Hôpital’s rule, that is, differentiating numerator and denominator:
  sT
2 e −1 2sesT 2s
lim sT
= lim sT sT
= = s.
T →0 T e + 1 T →0 e + 1 + T se 2
Equation (123) can then be expressed (for small enough T ) as:
 sT
2 e −1

 sT 
H e = Ha . (126)
T esT + 1
In deriving the frequency-domain criterion (123) we substituted z = esT [see Equation (122)], hence we have
2 z−1
  
H [z] = Ha = Ha (s) 2 z−1 . (127)

T z+1 s= T z+1

Therefore, we can obtain H[z] from Ha (s) by using the bilinear transformation:
2 z−1
 
s= . (128)
T z+1
A Matlab c
function to find digital filters by the bilinear transformation method is bilinear. The input data
are the coefficients of the numerator and denominator polynomials of Ha (s) and the sampling frequency in Hz.
Matlab c
returns the numerator and denominator polynomial coefficients of the desired digital filter H[z].

63
Choice of T in the bilinear transformation method
Since there is no aliasing (of the kind obtained with the impulse invariance method), the only consideration in the
choice of the sampling interval is the maximum frequency of the signals to be processed. If the highest frequency
to be processed is Fh Hz (ωh = 2πFh rad/sec) then, to avoid signal aliasing, we must use [see (66)],
1 π
T ≤ = .
2Fh ωh

Frequency warping in the bilinear transformation method

Consider Equation (126) with s = jω (frequency response),
 jωT jωT −jωT
!
−1 2 e 2 −e 2
  
 jωT  2e 2 ωT
H e = Ha = Ha −jωT = Ha j tan .
T ejωT + 1 T e jωT
2 +e 2 T 2
Therefore, the response of the resulting digital filter at some
particular frequency ωd is,
 
 jω T  2 ωd T
H e d
= Ha j tan = Ha (jωa ) ,
T 2
where,
2 ωd T
ωa = tan .
T 2
Thus, in the resulting digital filter, the behaviour of the de-
sired response Ha (jω) at some frequency ωa appears, not at
ωa but at frequency ωd , where
2 ωa T Figure 88: Mapping between analog and digital
ωd = tan−1 . (129) frequencies.
T 2
As can be seen from Figure 88, for small values of ωa the curve is practically linear so that ωd ≈ ωa and, thus, the
digital filter imitates the desired
analog filter at low frequencies.
At higher frequencies, however,
there is considerable distortion.
If we try to synthesise an ana-
log filter Ha (jω) (top plot of Fig-
ure 89), the resulting digital fil-
ter frequency response will be as
shown on the bottom plot of Fig-
ure 89. The analog filter be-
haviour in the frequency range ωa
from 0 to ∞ is compressed in the
digital filter in the range of ωd
from 0 to π/T .

Figure 89: Analog filter and corresponding digital filter frequency responses.

Bilinear transformation method with prewarping

The idea is to begin with a distorted analog filter (prewarping) so that the distortion caused by the linear trans-
formation is cancelled by the built-in (prewarping) distortion. Usually the prewarping is done at certain critical
frequencies. If we require a filter to have gains g1 , g2 , . . . , gm at frequencies ω1 , ω2 , . . . , ωm , respectively, then we

64
 0
must start with an analog filter H 0 (jω) which has gains g1 , g2 , . . . , gm at frequencies ω10 , ω20 , . . . , ωm , respectively,
where:
2 ωi T
ωi0 = tan , i = 1, 2, . . . , m. (130)
T 2
Application of the bilinear transformation (128) to this filter yields the desired digital filter with gains g1 , g2 , . . . ,
gm at frequencies ω1 , ω2 , . . . , ωm , respectively, since from (129) and (130) we have that the behaviour of the
analog filter at a frequency ωi0 appears in the digital filter at frequency:
0
 
2 −1 ωi T 2 −1 ωi T
tan = tan tan = ωi .
T 2 T 2

A simplified procedure
The procedure of prewarping followed by the bilinear transformation can be simplified if, instead of using (128)
we use
z−1
s= , (131)
z+1
and, instead of using (130) we use
ωi T
ωi0 = tan , i = 1, 2, . . . , m. (132)
2
The reason this simplification works just as well is that the factors 2/T cancel each other. If we use (131) instead
2
of (128) we obtain ωd = tan−1 ωa instead of (129), hence if we prewarp ωi according to (132) we obtain:
T
 
2 −1 0 2 −1 ωi T
tan ωi = tan tan = ωi .
T T 2

Bandpass filter design by the bilinear transformation method with prewarping

(Although we will only cover bandpass filters here, the procedure is similar for highpass and bandstop filters. See
the details in Lathi’s textbook.)

Bandpass filter Prototype lowpass filter

Figure 90: Frequency transformation for bandpass filters (with prewarping).

Suppose we want to design a bandpass filter to satisfy the specifications given on the left plot of Figure 90
(cf. Figure 54).
All critical frequencies (ωs1 , ωp1 , ωp2 , ωs2 ) are first prewarped using (132), thus obtaining
0 0 0 0
ωs1 , ωp1 , ωp2 , ωs2 . Next, we design (except for the prewarping of frequencies, the steps to design the analog
filter are identical to those presented on Page 32) the prototype lowpass filter Hp (s) with the specifications shown
on the right plot of Figure 90, where ωs0 is given by:
( )
0 0 0 2 0 2 0 0
ω p1 ωp2 − ωs1 ωs2 − ω p1 ωp2
ωs0 = min 0 0

0 , 0 0

0 . (133)
ωp2 − ωp1 ωs1 ωp2 − ωp1 ωs2

65
We then convert the lowpass prototype filter to the desired analog bandpass filter by replacing s with T (s) in
Hp (s), where
0 0
s2 + ωp1 ωp2
T (s) = 0 0

. (134)
ωp2 − ωp1 s
z−1
Finally, using (131), s is replaced with . The two transformations can be combined into a single one,
z+1
z−1 2 0 0

0 0

z+1
+ ωp1 ωp2 (z − 1)2 + ωp1 ωp2 (z + 1)2
Tbp [z] = T (s) = =

0 0 z−1 0 0




s= z−1
z+1 ωp2 − ωp1 z+1
ωp2 − ωp1 (z + 1)(z − 1)
0 0 0 0 0 0
+ 1 z 2 + 2 ωp1




ωp1 ωp2 ωp2 − 1 z + ωp1 ωp2 +1
= 0 0

.
ωp2 − ωp1 (z 2 − 1)
Therefore,
0 0
ωp1 ωp2 −1

a= 0 0 ,

z 2 + 2az + 1

ωp1 ωp2 + 1
Tbp [z] = , 0
ωp2 0
− ωp1
b (z 2 − 1)
b= 0 0 .


ωp1 ωp2 + 1
Thus, the digital bandpass filter transfer function H[z] can be obtained from the prewarped analog prototype
lowpass filter transfer function Hp (s) by directly replacing s with Tbp [z].

Time-domain equivalence method of FIR filter design

Given an analog filter transfer function Ha (s), we first find the impulse response ha (t) by computing the inverse
Laplace transform of Ha (s) or, equivalently, the inverse Fourier transform of Ha (jω) using expression (14).
However, since a digital filter has a frequency response that is periodic (see Page 58) with the first period in
π π
− ≤ ω ≤ , the best we can hope for is to realise the equivalence of Ha (jω) within this range. Therefore, the
T T
π π
limits of integration in (14) are taken from − to :
T T
Z π
1 T
ha (t) = Ha (jω)ejωt dω.
2π − Tπ

According to the time-domain equivalence criterion (120), for T small enough, we must have h[k] = T ha (kT ),
hence, Z π
T T
h[k] = Ha (jω)ejωkT dω. (135)
2π − Tπ

Windowing
The impulse response found in (135) has in general infinite duration. But, for an FIR filter, h[k] must have a finite
duration and must start at k = 0 for the filter to be causal. Consequently, the h[k] found in (135) needs to be
N0 − 1
truncated using an N0 –point (N0 = n + 1, where n is the order of the filter) window and then delayed by
2
to make it causal (recall that delay introduces linear phase, see Page 15). Straight truncation of data amounts
to using a rectangular window. Although such a window gives the smallest transition band (minimal spectral
spreading, see Figure 32 and the discussions on Page 19), it results in a slowly decaying oscillatory frequency
response in the stopband (due to leakage, see Figure 32 and the discussions on Page 19). The behaviour can be
corrected by using a tapered window (see Table 12.2 on Page 762 of Lathi’s textbook for some window functions
and their characteristics).
Once we know h[0], h[1], h[2], . . . , h[n] (obtained from (135) followed by windowing and  delaying) we can
find the transfer function H[z] of the digital filter from (116) and the frequency response, H ejωT , from (117).


66
Nonrecursive filter design by the frequency-domain criterion: The fre-
quency sampling method
The frequency-domain criterion [see (123)], with T small enough, is

Ha (s) = H esT .
 

We shall realise this equality for real frequencies, that is, for s = jω, so:

Ha (jω) = H ejωT .
 
(136)

For an nth order filter there are N0 = n+1 elements in h[k] [cf. (115)] and we can hope to force the two frequency
spectra in (136) to be equal only at N0 points. Because the spectral width is 2π/T (see Page 58), we choose these
2π/T
frequencies equally spaced, ω0 = rad/sec apart. That is, the sampling interval of the spectrum is:
N0
2π
ω0 = , (137)
N0 T
and we require
Ha (jrω0 ) = H ejrω0 T ,
 
r = 0, 1, 2, . . . , N0 − 1. (138)
The problem is now to determine the filter impulse response h[k] from the knowledge of the N0 uniform samples
of Ha (jω), that we can denote Hr :

Hr = Ha (jrω0 ), r = 0, 1, 2, . . . , N0 − 1. (139)

Recall that a digital filter’s transfer function is the Z-transform of its impulse response. For the finite impulse
response sequence h[0], h[1], . . . , h[N0 − 1], we have,
∞
X N
X 0 −1
−k
H[z] = h[k]z = h[k]z −k .
k=−∞ k=0

In particular, when z = ejrω0 T , we have,

N 0 −1 N0 −1
jrω0 T −k
X X
jrω0 T
h[k]e−jrω0 T k .
 

Hr = H e = h[k] e =
k=0 k=0

Recalling (72) we conclude thath[k] and Hr are a DFT pair with Ω0 = ω0 T . Hence, the desired h[k] is the IDFT
of Hr = Ha (jrω0 ) = H e 0 T given by [see (73)]:
 jrω

N0 −1 N0 −1
1 X 1 X j 2πrk
h[k] = Hr ejrkω0 T = Hr e N0 , k = 0, 1, . . . , N0 − 1. (140)
N0 r=0 N0 r=0

Thus, we can use the powerful IFFT routine (recall the computational efficiency of the FFT and IFFT algorithms
explained on Page 45 and illustrated in Figure 74) to compute the N0 values of h[k], as in (140), from the problem
data given by the samples of the desired spectrum (139).

Linear phase filters

If we want a linear phase characteristic we start with a zero phase frequency response, in which case the resulting
h[k] is an even (or odd) function of k. Once we obtain the impulse response h[k] via the IDFT (140) (which will
N0 − 1
be centred at k = 0), we make the filter causal by delaying h[k] by sampling units. Such a delay amounts
2

67
 N0 −1 N0 −1
to multiplying Hr = H ejrω0 T by e−jrΩ0 2 = e−jr 2 ω0 T (see the time-shifting property number 3 of the
 

DFT on Page 42).

Thus, for a desired frequency response H ejωT = Ha (jω) [see (136)], we begin our computations with
 

N0 −1
h 2π i N0 −1
j r −jrπ N
Hr = H ejrω0 T e−jr 2 ω0 T = H e N0 e
 
0 , r = 0, 1, 2, . . . , N0 − 1. (141)

The desired impulse response h[k], k = 0, 1, . . . , N0 − 1 is obtained with the IDFT (or IFFT) and is causal and
has a linear phase response.
In obtaining the samples of (141) we can do it for r = 0, 1, . . . , N02−1 , and the ones for r =
N0 +1 N0 +3
2
, 2 , . . . , N0 − 1, can be obtained from the conjugate symmetry property [see (74)]:

Hr = HN∗ 0 −r .

68