University of Tripoli - Faculty of Engineering

Electrical and Electronic Department

Fall 2020
EE423 Active Networks

Lecture 15 - Introduction to Digital Filters

Dr. Ali Ganoun

3/05/2021
 System Structure

• Development of software (that is, a program) that

implements the system on a general purpose
computer or a special purpose digital signal processor.
• Design and implementation of a hardware architecture
that can be used to implement the system using
discrete components or VLSI technology.

2
BASIC STRUCTURE FOR
SYSTEM/FILTER
 The Transfer Function of LTI system can be
connected in 2 ways :
a. Parallel Connection :

The overall transfer function,

H(z) = H1(z) + H2(z) + … + HL(z)

1-3
BASIC STRUCTURE FOR
SYSTEM/FILTER
 b. Cascade connection :

The overall transfer function :

H(z) = H1(z).H2(z). … HL(z)

1-4
Digital Filters

5
6
Block diagrams and signal flow graphs

first-order IIR system

Structure for the first-order IIR system

block diagram signal flow graph 7

Representing by signal-flow graph

1-8
 IIR Filter Structures
 The system of an IIR filter is given by
M

B( z ) k
b z k
b0  b1 z 1  ..  bM z  M
H ( z)   k 0
 1 N
; a0  1
N
1  a1 z  ..  aN z
 k
A( z ) k
a z
k 0

 The order of the filter is N if aN0.

 The difference equation representation is

M N
yk   bm xk m   am yk m
m0 m0

1-9
IIR Filter Structures

 Direct form
◦ Direct form I
◦ Direct form II
 Cascade form
 Parallel form

1-10
Direct form I structure for implementation of an N th order IIR system
with N =M= 2.

cascade connection of two systems.

11
 Assume M=N=4:
yk  b0 xk  b1 xk 1  b2 xk 2  b3 xk 3  b4 xk 4
a1 yk 1  a2 yk 2  a3 yk 3  a4 yk 4

xk b0
yk
z-1 z-1
b1 -a1

z-1 z-1
b2 -a2

z-1 z-1
b3 -a3

z-1 z-1
b4 -a4

Direct form I structure

1-12
Direct form II

Direct form II structure for implementation of an Nth order system.

13
Direct form II

14
Example:

15
Example:

16
Example:
1  2 z 1
H ( z) 
1  1.5 z 1  0.9 z  2

Direct form I implementation Direct form II implementation

1-17
Example:
1  2 z 1  z 2
H ( z) 
1  0.75 z 1  0.125 z 2

xk(n) yk
z-1
0.75
2

z-1
-0.125

Direct form II implementation

1-18
 Cascade form
◦ The system function H(z) is factored into smaller second
order sections with real coefficients, called biquads.
◦ The system function is then represented as a product of
these biquads.
◦ Each biquads is implemented in a direct form.
◦ The entire system function is implemented as a cascade of
biquad sections.
◦ Each numerator and denominator is factoring into their
respective roots and then combining the roots into second
order polynomials.

1-19
20
 Cascade form structure for N=4

Biquad section structure

21
 Example

1  2 z 1  3z 2 1  z 1  z 2
H ( z) 
1  0.75 z 1  0.125 z 2 1  0.5 z 1  0.1z 2

x(n) y(n)

z-1 z-1
-0.75 -2 -0.5 -1

z-1 z-1
-0.125 3 0.1

Cascade form structure

1-22
Example:

23
Cascade form structure
24
Example:
1  2 z 1  z 2 (1  z 1 )(1  z 1 )
H ( z)  1 2

1  0.75z  0.125z (1  0.5 z 1 )(1  0.25z 1 )

Cascade structure: direct form I implementation

Cascade structure: direct form II implementation

1-25
 Parallel form
◦ This is similar to the cascade form.
◦ After factorization, a partial fraction expansion is used to
represent H(z) as a sum of smaller second order sections.
◦ Each section is a gain implemented in a direct form.
◦ The entire system function is implemented as a parallel
network of sections.

1-26
27
Parallel form
structure for
M=N=4 C0

B1,0

z-1
-A1,1 B1,1

xk z-1 yk
-A1,2

B2,0

z-1
-A2,1 B2,1

z-1
-A2,2

1-28
Example:
1  2 z 1  z 2  7  8 z 1
H ( z)  1 2
 8
1  0.75z  0.125z 1  0.75z 1  0.125z  2

1-29
another alternation of the same system

1  2 z 1  z 2 18 25
H ( z)   8  
1  0.75z 1  0.125z  2 1  0.5 z 1 1  0.25z 1

1-30
 1  2 z 1  3z 2 1  z 1  z 2
Example H ( z)  1 2

1  0.75 z  0.125 z 1  0.5 z 1  0.1z 2

z-1
-0.5 -1
x(n) z-1 y(n)
0.1

z-1
-0.75 -2

z-1
-0.125 3

Parallel form structure

1-31
1-32
33
FIR Filter Structures
 A finite-duration impulse response filter has a system function
of the form
M 1
1 2
H ( z )  h0  h1 z  h2 z  ..  hM 1 z 1 M
  hk z  k
k 0

 The impulse response hk is

hk 0  k  M 1
hk  
0 else
 The difference equation representation is
yk  h0 xk  h1 xk 1  h2 xk 2  ...  hM 1uk  M 1

1-34
 The order of the filter is M-1
 The length of the filter is M (equal num of
coefficients).
 The FIR filter structures are always stable.
 They are relatively simple compared to IIR
structures.

1-35
FIR Filter Structures

 Direct form
 Cascade form

1-36
BASIC STRUCTURE FOR FIR FILTER
 An FIR filter of order 4: yk  h0 xk  h1 xk 1  h2 xk 2  h3 xk 3  h4uk 4

xk z-1 z-1 z-1 z-1

h0 h1 h2 h3 h4

+ + +

+  yk
xk z-1 z-1 z-1 z-1

h0 h1 h2 h3 h4 yk

1-37
 Cascade form
◦ The system function H(z) is converted into products of
second order sections with real coefficients.
◦ Transfer function H(z) of a causal FIR filter of order M-1
M 1


H ( z )   hk z  k  h0 1  hh10 z 1  ..  hMh01 z  M 1
k 0

◦ Factorized form
k
H ( z )  h0  (1  k ,1 z 1  k ,2 z 2 )
k 1

Where K = (M-1)/2 and βk,1 , βk,2, are real numbers

representing the coefficients of second order sections

1-38
Cascade form structure for realization of a 6th-order FIR
system.

1-39
 Example:
 An FIR is given by the system function

H ( z)  1  16 161 z 4  z 8

 Determine and draw the direct and cascade structures

1-40
 a- Direct form
 The difference equation is given by

yk  xk  16.0625 xk 4  xk 8

xk z-4 z-4

16.0625

yk

1-41
 b- Cascade form

xk yk
z-1 z-1 z-1 z-1
2.8284 0.7071 -0.7071 -2.8284

z-1 z-1 z-1 z-1

4.0 0.25 0.25 4.0

1-42
 Example
FIR direct and cascade form structures

43
 Normal direct form

The structure require nine multiplications

44
[sos,G] = tf2sos(b,a)

Normal cascade form

requires nine multiplications 45

 Structure conversion,
simulation and verification

Structure conversion

Usually, filter design packages provide the impulse

response of an FIR filter or the coefficients of second-
order sections for IIR filters. However, on several
occasions, the structure specified by the design software
is different from the structure chosen for implementation
of the filter. Hence, it is important to be able to convert
between different structures.

46
47
 Simulation and verification

After obtaining the coefficients of a given structure,

the next task is to simulate and verify that a
system with these coefficients works correctly

Verification of the converted structure can be obtained in many different

ways. The simplest is to
(a) convert the given structure to direct form I,
(b) use the filter function to implement this structure,
(c) excite the structure in (b) and the structure to be tested with the
same input and compare the corresponding outputs.

Typical test inputs include the unit step, sinusoidal sequences, or

random sequences.
48
the cascade structure

b = [1,-3,11,-27,18]; a = [16,12,2,-4,-1]
[sos,G] = tf2sos(b,a)

49
To verify that the above structure is the correct
one we will compute impulse response from
two structures and compare them

The two structures are equivalent.

50