DIGITAL SIGNAL PROCESSING

Chapter 9a:
DIGITAL FILTER REALIZATIONS
Block Diagram of Digital Filters
Reference:
S J.Orfanidis, ”Introduction to Signal Processing”, Prentice –Hall , 1996,ISBN 0-13-209172-0
M. D. Lutovac, D. V. Tošić, B. L. Evans, “Filter Design for Signal Processing Using MATLAB
and Mathematica”, Prentice Hall, 2001

Lectured by Prof. Dr. Thuong Le-Tien

National Distinguished Lecturer
Tel0903 787 989
Email: [email protected],

Dated on April 2025

1

1
Digital filter realizations
◼ 1. Direct form realization

N ( z ) b0 + b1 z −1 + b2 z −2
H (z) = =
D ( z ) a 0 + a1 z −1 + a 2 z − 2

y n = − a1 y n −1 − a 2 ùy n − 2 + b 0 x n + b1 x n −1 + b 2 x n − 2

• b multiplier terms are feeding forward and x

dependent. (non-recursive terms)
• a multiplier terms are feeding back and y
dependent (recursive terms)
2
 Second order IIR digital filter
x(n) b0 y(n)

v0(n) w0(n)
-1 -1
z z
b1 -a1
v1(n) w1(n)

-1 -1
z z
b2 -a2
v2(n) w2(n)

** General case of a IIR digital filter

−1 −2 −L
N ( z ) b0 + b1 z + b2 z +  + bL z
H (z) = =
D ( z ) a 0 + a1 z −1 + a 2 z − 2 +  + a M z − M
3
y n = − a 1 y n −1 − a 2 y n − 2 −  − a M y n − M + b 0 x n + b1 x n −1 + b 2 x n − 2 +  + b L x n − L
x(n) b0 y(n)

v0 w0
-1 -1
z z
b1 -a1
v1 w1

-1 -1
z z
b2 -a2
v2 w2

-1 -1
z z
bL -aM
vL wM

4
Example: 2 − 3 z −1 + 4 z −3
H (z) =
1 + 0 . 2 z −1 − 0 . 3 z − 2 + 0 . 5 z − 4
y n = − 0 . 2 y n −1 + 0 . 3 y n − 2 − 0 . 5 y n − 4 + 2 x n − 3 x n −1 + 4 x n − 3
x(n) 2 y(n)

v0 w0
-1 -1
z z
-3 -0,2
v1 w1

-1 -1
z z
0,3
v2 w2

-1
z
4
v3
w3

-1
z
-0,5
w4 5
2. Canonical or Direct Form 2 of IIR digital filters

y n = ( b 0 x n + b1 x n −1 + b 2 x n − 2 ) + ( − a1 y n −1 − a 2 y n − 2 )
1
H ( z) = N ( z)
D ( z)
x(n) y(n)

v0(n) b0 w0(n)
-1 -1
z z

v1(n) w1(n)
b1 -a1
-1 -1
z z

v2(n) w2(n)
b2 -a2

N(z) 1/D(z)

6
 * Interchanging N(z) and 1/D(z)

x(n) w(n) w(n) y(n)

w0(n) w0(n) b0
-1 -1
z z

w1(n) w1(n)
-a1 b1
-1 -1
z z

w2(n) w2(n)
-a2 b2

1/D(z) N(z)

7
 •Merging blocks to build the
canonical form of the second order
section (SOS) digital filter
x(n) w(n) = wo(n) y(n)

wo(n) bo
-1
z

-a1 w1(n) b1

-1
z

-a2 w2(n) b2

8
* M order IIR filter realization
x(n) w(n) y(n)

wo bo
-1
z

-a1 w1 b1

-1
z

-a2 w2 b2

-1
z

-aM wM bM
9
3. Cascade Form
The cascade realization form of a general
transfer function assumes that the transfer
function is the product of such second
order sections

K −1
bi 0 + bi1 z −1 + bi 2 z −2
K −1
H (z) =  H i (z) = 
i=0 i=0 1 + a i1 z −1
+ a i2 z −2

10
 H0(z) H1(z)

x0 = x w0(n) x1 = y0 w1(n)

x(n) w00 b00 w10 b10

-1 -1
z z

-a01 w01 b01 -a11 w11 b11

-1 -1
z z

-a02 w02 b02 -a12 w12 b12

x2 = y1 w2(n) x3 = y2 w3(n) y = y3

w20 b20 w30 b30 y(n)

-1 -1
z z

-a21 w21 b21 -a31 w31 b31

-1 -1
z z

-a22 w22 b22 -a32 w32 b32

11
H2(z) H3(z)
 Example
 3 − 4 z −1 + 2 z −2   3 + 4 z −1 + 2 z −2 
H ( z) =  −1 −2   −1 −2 
= H 0 ( z) H1 ( z)
 1 − 0 .4 z + 0 .5 z   1 + 0 .4 z + 0 .5 z 
9 − 4 z −2 + 4 z −4
=
1 + 0.84 z − 2 + 0.25 z − 4
x(n) w(n) y(n)

w0(n) 9
-1
z

-1
z
-0,84 w2 -4

-1
z
w3
-1
z
-0,25 w4 4
12
 Example:
1 − 1 .5 z − 1 + 0 .48 z − 2 − 0 .33 z − 3 + 0 .9376 z − 4 − 0 .5328 z − 5
H (z) =
1 + 2 .2 z −1 + 1 .77 z − 2 + 0 .52 z − 3
The roots of Numerator: z = 0.9, z = -0.5 ± 0.7j,
z = 0.8 ± 0.4j
The roots of Denominator: p = -0.8, p = - 0.7 ± 0.4j
(1 − 0.9 z−1 )
(1 − (−0.5 + 0.7 j ) z )(1 − (−0.5 − 0.7 j ) z ) = (1 + z
−1 −1 −1
+ 0.74 z− 2 )
(1 − (0.8 + 0.4 j ) z )(1 − (0.5 − 0.4 j ) z ) = (1 − 1.6 z
−1 −1 −1
+ 0.8 z− 2 )
(1 + 0 .8 z−1 )
(1 − (−0.7 + 0.4 j ) z )(1 − (−0.7 − 0.4 j ) z ) = (1 + 1.4 z
−1 −1 −1
+ 0 .65 z− 2 )
1 − 0 .9 z − 1 1 + z − 1 + 0 .74 z − 2
H ( z) = .
1 + 0 .8 z −1 1 + 1 .4 z −1 + 0 .65 z − 2
. 1 − 1 . 6 z (
−1
+ 0 . 8 z −2
)
13
 Example: 1 + z −8
H (z) =
1 − 0.0625 z −8
−8 −8 j j 2j k j ( 2 k +1 )
1+ z =0 z = −1 = e =e e =e
j ( 2 k +1 ) / 8
zk = e , k = 0 ,1 , ,7
Complex conjugate pairs z0 , z7 , z1 , z6 , z2 , z5 , z3 , z4 
(1 − z z )(1 − z z ) = 1 − 2 cos 8 z
0
−1
7
−1 −1
+ z− 2 = 1 − 1.8478z−1 + z− 2
 
 3 
(1 − z z )(1 − z z )
1
−1
6
−1
= 1 − 2 cos  z−1 + z− 2 = 1 − 0.7654 z−1 + z− 2
 8 

(1 − z z )(1 − z z ) = 1 − 2 cos 58 z

2
−1
5
−1 −1
+ z− 2 = 1 + 0.7654 z−1 + z− 2
 
 7  −1
(1 − z z )(1 − z z )
3
−1
4
−1
= 1 − 2 cos  z + z− 2 = 1 + 1.8478z−1 + z− 2
 8  14
 1 − 0 . 0625 z − 8 = 0  z 8 = 0 . 0625 = 0 . 0625 e 2  jk = ( 0 . 5 ) 4 e j ( 2 k + 1 ) 
p k = 0 .5 e 2  j k / 8 , k = 0 ,1 , ,7
 p 0 , p 7 ,  p1 , p 6 ,  p 2 , p 5 ,  p 3 , p 4 
(1 − p z
0
−1
)(1 − p z ) = (1 − 0.5z )(1 − 0.5z ) = 1 − 0.5z
7
−1 −1 −1 −2

 2  −1
(1 − p z )(1 − p z ) = 1 −
1
−1
6
−1
2 cos −2 −1
  z + 0.5z = 1 − z + 0.5z
−2

 8 
 4  −1
( )( )
1 − p2 z 1 − p5 z = 1 − 2 cos  z + 0.5z− 2 = 1 + 0.5z− 2
−1 −1

 8 
 6  −1
( )( )
1 − p3 z 1 − p4 z = 1 − 2 cos  z + 0.5z− 2 = 1 + z−1 + 0.5z− 2
−1 −1

 8 
1 − 1.8478 z−1 + z−2  1 − 0.7654 z−1 + z−2 
H ( z) =  −2  . −1 −2 .
 1 − 0.5 z   1 − z + 0.5 z 
1 + 0.7654 z−1 + z− 2  1 + 1.8478 z−1 + z− 2 
. −2  . −1 −2 .
 1 + 0.5 z   1 + z + 0.5 z  15
 Magnitude response
z2 p2
z1 3
p3 p1 Pole frequencie s = (2k)/8
zero frequencie s = (2k+1)/8
z3 z0
2
p4 /8 p0

|H()|
z4 z7
1
Unit circle p5 p7
z5 p6 z6
0
= poles 0 1 2 3 4 5 6 7 8
= zeros  in units of /8
x y

w0
For each input
-8 w0 = x + 0,0625
z
y = w0 + w8
0,0625
w8 delay (8,w)

16