Structure for IlR Filter:

5.6

stands for infinite impulse response. Generally llR systems PTU- Dec. 89
We know that IIR connection is present from output side to the
are
feedback
Arecursive system means;
future samples of input and past values of
input sioe. For '
realization of IIR systems present, past, output are Tequ.
following structures are used.
For the realization of IR systems,
Direct form structure
Cascade form structure
2
Parallel form structure.
3.

5.7 Direct Form Structure for IIR Systems :

Direct form structure is again divided into two types :
A) Direct form-I structure B) Direct formn-II structure.
The general difference equation for discrete time LTÈ system is given as,
M
b,z-k
H(Z) =ka0
N
1+ ayzk
k=l

Now let H(Z) =Numerator term of Equation (5.7.1)

M
:. H, (Z) = bzk
k=0

And let, H,(Z) = N

1+ a,zk
k=l

Putting Equations (5.7.2) and (5.7.3) in Eaquation (5.7.1) we ge,

H(Z) = H,(Z) * H, (Z)
 I Processing
(PTU-Sem. 5) 5-19
Signal
Structure for Realization of
Fogial D.T.S.
e q u a t i o n
shows that H(Z)
This
represented as the cascade
be
((Z) and H, (Z) as x(n)
H{2)
ation
of H, y(n)
5.7.1(a),.

Fig.
Awn in
Fig. 5.7.1(a) : Cascade
connection of H(Z) and H, (Z)
hat H, (Z) is the transter function
of the numerator term.
system. So H,(Z) is called as all Numerator always contain
of the zero system. Now H, (Z ) is transfer
ANminator.
Denominator always contain
poles of the system. So H, (Z) is called as function of
all pole system.
A Direct form-l structure:

The direct form-I structure is obtained by

cascading
(Z) and H, (Z). First we will draw the direct form (connecting
structure for H (Z).in series)., the structure for
Drectform structure for H, (Z):
Recall the equation of H(Z) (Equation (5.7.2). It is,
M

H(Z) = b,Zk ..(5.7.5)

k=0

But we know that,

H(Z)
Output [ Y(Z)]
Input [ X(Z)]
To avoid the confusion we will
write,
W (Z)
H(Z) X(7) where W (Z) is output of first stage.

W(Z) M

X(Z) = k=0 b,Zk ..(5.7.6)

W(Z)
X(Z) = boZ°+ b, Z'+b, Z-2+...+ by Z- M

But Z = 1

:. (Z)+... +by Z M X(Z)

X(Z) + b, Z-'X(Z) +b,Z x
W(Z) =
bo ...(5.7.7)
Taking inverse Z-transform (IZT) of Equation (5.7.7) we get,
(n -Z)+... + bM x(n-M)
YI(n) = bo x(n)+b, x ( n- 1 )+b, X ...(5.7.8)
Here bo bj D2 . by are the coefficients. The direct form realization of Equation (5.7.8) is
OWn in Fig.
5.1.1(6).
 Digital Sigle bo
x(n)o
b,
b
DM 1
bM
Kie 6.7.1(b) : Direct form realization of H Ë(Z) (All zero system
H, (Z) :
Direct form structure for
is,
Recall the equation for H, (Z ). It
H, (Z) = N
1+ Xa,zk
k=1
Output [ Y(Z)]
We have, H, (Z) = Input [ X (Z)] ..50
Now H, (Z) represents second stage of Fig. 5.7.1(a). Input of second stage is the oupudn
stage. Thus input of H,(Z ) is W(Z) and output of H, (Z) is the output of overall sytemvàd
Y (Z). W(n)
Thus Equation (5.7.10) becomes,
H, (2) = W(Z)
-ag
Putting this value in Equation (5.7.9) we get,
Y(Z)
W(Z) N
-ag
1+ X 4, zk +
k=l
= W(Z) -ay-1
N
:Y(Z)+Y(Z) E4z' = W (Z)
Lk=| -ay sy
(Alpole
(Z)
Fig. 5.7.1(c): Direct form realization of H,(
 Sgnal I Processing (PTU-Sem. 5) 5-21
ota Structure for Realization of D.T.S.
. Y(Z)
...(5.7.11)
Eyanding
[Link] we get,

YiZ) = -[Z+a, '+... *a zN]

Y(Z) + W(z)
-a,ZY (Z) -
.:
Y(Z)
a,ZY(Z)... - aEZY(Z) + W(Z)
Taking ZT of EEquation (5.7.12) we get, ..5.7.12)
vn) = -a y(n-1)-a, y(n-2) ... -
an y(n -N)+ w,
..(5.7.13)
Here -a,. - å, - ay are the coefficients.
.Ainct form implementation of Equation (5.7.13) is shown in Fig.
ll R HS. terms of Equation (5.7.1l3)
5.7.1(c).
except w (n) are the delayed output terms. Thus this
connection,
ka feedback (recursive)
direct form-I structure is obtained by cascading H, (Z )
and H, (Z), Thus direct form-I
an of IIR system is obtained by connecting Fig. 5.7.1(6) and Fig.
5.7.1(c) in series. It is shown
in Fig. 5.7.1(d).

b
x(n)
y(n)

by-1 -dN-1

DM

All zero system All pole system

Fig. 5.7.1(d) : Direct form-I realization of IIR system

Computation complexity :
e
1.
computational complexity of direct form-I structure is as follows:
Number of multiplications =M+N
= +1
Number of additions = M+
.

N
Number of memory locations = M+N+1.
 Sem. 5) 5-22
Digital Signal Processing (PTU Structure for
Direct form-ll structure
:
(B)
5.7.1(d). Here, first block represents all zero system and
[Link]
ObserveThus,
Iole system. dircct form-l structure; zeros of H(Z)ure realised firstsecond block
and the
are ealizd seond.
structures for LTI (linear time invariant) systems.
Now we are studying these
positions of H, (Z ) and H, (Z). This will give Since he Sye
are linear we can interchange the us the
sinuure Thus in direct form-IIstructure; poles are realised first and zeros second direxa fgm
We have, H(Z) = H(Z)·H, (Z)
We will interchange the equations of H, (Z) and H, (Z).
: H, (Z) = N

1+ X4zk
k=1

and H, (Z) = b, zk
k=0 .51
The cascade connection of H(Z)= H,
(Z)·H, (Z)is shown in Fig. 5.7.2(a).
x(n) o H4)
Here W(Z) represents the output of w(n)
Poles Zeros
first block.

Fig. 5.7.2(a) : Decomposition of direct form-I realizaia

(1) All pole system:
Output W (Z)
We can write, H(Z) = Input ...57n
X(Z)
Comparing Equations (5.7.15) and (5.7.16) we get,
W (Z)
X(Z) =
1+ ) a, zk
k=l
N

= X(Z )
k=|
N
:.W(Z)+ = X (2)
Lk=|

W(Z) = X(Z)-V
Lk=l
 Processing
(PTTU-Sem. 5) 5-23
Signal Structure for Realization of D.T.S.
al
the.
summation
we get,
Eypanding

X(Z)-
a, Z-IW(Z)
Wn
wZ) =
-2 w(Z) ... - ay ZN W(Z) ...(5.7.18)
-ay Z

Taking
rof Equation(5.7.18) we get,
1ZT

-a, X(n -2) ....- ay X(n-N) ...(5.7.19)

dN-1

implementation of
Te direct form
-aN
(S7.19)), is shown in Fig. 5.7.2(b).
Autin

Fig. 5.7.2(b) : Direct form realization

of H,(Z) (All pole system)
AlZero system:

As shown in Fig. 5.7.2(a), the output of second stage is the overall output of the system which is
I Whileinput to this stage is the output of first stage, which is W (Z).
Thus we can write, H, (Z) =
Output Y (Z)
Input W(Z) ...(5.7.20)
Comparing Equations (5.7.16) and (5.7.20) we get,
M
Y (Z)
W (Z) ..(5.7.21)
k=0

:.Y(Z) = w(z) .k=0

b
Expanding the summation we get,
Y(Z) = W(Z)[b,Z+bz+b,Z+...* by Z M]
Here Z =1; multiplying by W
(Z) we get,
() = b, W (Z) +b, Z- w(Z) + b, ZZw(Z)+... + byZMw (Z) ...5.7.22)
laking I2T of Equation
(5.7.22) we get,
) = bo W
(n)+ b, w (n-1) + b, w(n-2)+ ... +bm w(n- M) ..(5.7.23)
The e direct
form implemcntation of Equation (5.7.23) is shown in Fig. 5.7.2(c).
 Digital Signal P0
bow(n)
W) +y(n)
w(n)
b, b,w(n 2)

b, bw(n-2)

bM-1

by w(n-M)

(All zero system)

Fig. 5.7.2(c) : Direct form realization of H, (Z )
Now irect formn-II structure is obtained by cascading H, (Z) and H (Z ). That means ts

connecting Fig. 5.7.2(b) and Fig. 5.7.2(c) inseries. This is shown in Fig. 572
obtained by
common delay elements ar
In Fig. 5.7.2(d), separate delay elements are used. Generally
structure is drawn as shons
irect fom-II structure. By using common delays, irect rom lI
Fig. 5.7.2(e).
Note that here we have considered N= M.

Do
’yn)

-aN-1 bM-1

-aN

All pole Systerm

All zero system

Fig. 5.7.2(d) : Direct form-II realization of IIR system

 P r o c e s s i n g
(PTU-Sem. 5) 5-25
i
aSignal Structure for Realization of
D.T.S.
x(n)

-aN-1 DN-1

e. 5.7.2(e): Direct form-Il realization using common delay elements

onpvtation complexity :
aoputationall complexity for direct form-II realization of IIR is as follows :
Number of multiplications = M+N+1.
Number of additions = M +N.
: The number of delay elements are reduced in direct form-II structure, compared to direct form-I
snucture. That means the menory locations are reduced. The memory locations required for
direct form-I structure are { M,N}.
hy these structure are called as direct form structures ?
The direct form-I and direct form-II structures are obtained directly from the corresponding
esler funcions without any rearrangements. So these structure is called as direct form structures.
rntage :The only advantage of direct form realization is its implementation which is easy. We
drect form-lI structure compared to direct form-I structure; because less memory locations are
quted.
wbvantage e:Both direct form structures are sensitive to the effects of quantization errors in the
wIicients. So in the practical applications ; these structures are not preferred.
Cascade Form Realization Structures :
PTU -Dec. 2005
To
HZ)is factobtoairedn the cascade form realization; the numerator and denominator of given transfer function
into the product
of second order terms.
Then the total
transfer function H(Z) is expressed as,
..(5.8.1)
Here H, (Z), H, (Z)H(Z)H(Z)
= H,are(Z)·H,(Z)...H, (Z) Then each subtransfer function
second order polynomials.
H... etc. ) can be realised using direct form-I or direct form-II structures. The total transfer
choNiS
dbained 0y connecting all second order subsystems in series as shown in Fig. 5.8.1(a).
 5-26 Structure for
Processing (PTU-Sem. 5)
Digital Signal Realization
Yk-1(n) = oDIS.
x(n) =X(n H,(Z) y,(n) = H2) | y2n) = xy(n)
HuZ) yn)
n)yn)
*g(n) Xg(n)
Fig. 5.8.1(a) :Cascade form realization

Now we have the general difference equation for discrete time LLTI system given by,
M

k=0
H(Z) = N
1+ a4,zk . 531,
k=1

Since H, (Z), H,(Z) are second Iorder polynomials ; we can write the second order differenti
(5.8.2).
equation for H, (Z) by putting M=N= 2 in Equation
2

.. H, (Z) = k=02
1+ a,Zk
k=1

by Z°+b, Z'+b,22
.:. H,(Z)
1+a,Z'+ a, Z2
But Z =1
bo+ b,2-' +b, z-2
.:. H,(Z ) = ...533)
1+a,Z'+ a,Z?
We can obtain direct form-II
structure of Equation (5.8.3), similar to
Fig. 5.7.2(e). Now, Fig. 5.7.2e) shows the
+
direct form-II realization for N= M : that x(n)
means for M stages. In this case we have b.
second order equation. So by using
Fig. 5.7.2(e) we can draw the direct form
structure for Equation (5.8.3) by putting be
M = N = 2. This
structure is shown in
Fig. 5.8.1(6). form-II realization 0
FIg. 5.8.1(b) : Direct
second order subsystem
realizationo
So all such subsystems should cascade form
IIR system. be connected in series to obtain the
F o o t a S
l i g n a l
I0r Healization of D.T.S.

Form Realization Structure :

Parallel

PTU - Dec. 2005

general difference cquation for llR systems given by,
have
lhe M
We

H(Z) = k=0N
...(5.9.))
1+ X4zk
k= |

partial fraction expansion we can express overall transfer function H(Z) as,
using
Br
H(Z) = C+ H, (Z) +H, (Z) +... +H, (Z) ...(5.9.2)
. i s constant and H, (Z ),
%2. . (2 are second order
The general block schematic of
fom
realization structure for IIR
5.9.1, x(n)
sem iSAS shoWn in Fig.
Here H, (Z), H, (Z )...etc. can be
or direct
id by using drect form-I structures
i l structures. Then all these
aanected in parallel as shown in
59.1 to obtain the parallel form HZ)
izion for IR system. y(n)
Fig. 5.9.1:Parallel form realization for IIR system
kgolications :
The parallel form realization is generally used for high speed filtering applications. Since this is a
ail connection ;the processing of filtering operation is performed parallely.
0 Representation of Structures using Signal Flow Graphs:
PTU - May 2008
baically asignal flow graph is graphical representation of the block diagram structure. Both the
gal flow
graph
and the block diagram structure provide the same information.
eAample, we will consider the second order subsystem. We know that the transfer function of
End order subsystem is
given by.
bo+b,Z'+ b, z-2
H(Z) = 1+a, Z+a, Z-2
...(5.10.1)

Wehave already drawn the direct form-II realization of second order subsystem (Fig. 5.7.2(d).
ereference
onin Fig. [Link] 1is(b)again drawn as shown in Fig. 5.10.1(a) and its signal flow graph representation is
 Processing (PTUSem. 5)
Digital Signal
Adder node Branching node
1
bo y(n) x(n) bo
A()

b
a
’
bo b

(b)
(a)
5.10.1 :Direct form-II realization and its signal flow graph
Fig.

How to draw a signal flow graph ?

given block diagram realization; it is simple to draw the signal flow grach T
From the
procedure is as follows :
1. Replace all adders by adder nodes.
Whenever there are different branches ;draw the branching
node.
2.
coefficients as it is.
3. Keep the directions of arrows and the corresponding
delay element by simple transmittance branch. For that branch, wnit Z
4 Replace every
indicate the delay operation.
correspondence between the signal flow graph and block diagram represena
There is 1:1
And the signal flow graph representation is much simple compared to block dagram repres
 (PIU 5-69
Signal|
Processing
Structure for Realization of D.T.S.
Digital
L a t t i c e
Structure for lIR Filters :
511 PTU- May 2008. May 2013
studied lattice structure for FIR filters. For the reference, recall system transfer
We h a v e stu function
filler. It is
SMorder
FIR
M

Bu( K) z--K, M>1 ...(5.11.1)

Hñ(Z)
= l+ K=! +Af(n) = y(n)
transfer function, we have x(n)
this
For
stage lattice structure as shown in
single
ahained

With the help this structure shown in ’B,(n)

how to draw lattice
we will discuss Fig. 5.11.1
fig. 5.11.1, filters.
lIR
sructure for
can be expressed as,
aneider all pole IIR system. For this system the transfer function
1 ...(5.11.2)
H (Z) N

1+ X BN (K) Z-K
K=1
Y (Z)
written as, H (Z)=y(7)
We kmow that H (Z), is a transfer function, which can be
Y (Z) 1
N
X(Z)
1+ X BN (K) Z-K
K=l

N
X (Z)
LK=1
summation.
(Z) is independent of summation. So it can be transferred inside the
Here Y

Y(Z) + ) B(K) Y(Z) Z-k = X(Z)

K=1
N ...(5.11.3)

* Y(Z) = - S By(K) Y(Z)"z-K+X(Z)

K=1

laking inverse z of both sides we gel,

N ...(5.11.4)

Y(n) = - ) Bx(K)y (n-K) +x (n )

K=1
Now we will interchange the role of x(n ) and y(n ).

** X(n) = -S B(K):x (n-K) +y (n)

K=l
 Digital Signal Processing (P ealzaOn diD.
Rearranging the terms we get,

(n)+ 2 BNK ) x(n -K)

y(n) = Kl
gei,
of both sides we
Taking Z transtorm
N

Y(Z) = X(Z)+ , BN (K)"Z- Kx(Z)

K=|

.: Y(Z) = xz1EA(K)z-* K=1

N
Y (Z)
H(Z) = X0ZÍ -1+ EAy(K )Z
:: K=1

Observe Equations (5.11.6) and

zero FIR
Forward
(5.11.1). Equation (S.11.1) is for all A,(n) = x(n)
IIR
system and Equation (5.11.6) is for all pole
means
system. Both equations are same. That
we can obtain lattice structure of IIR system by
interchanging the role of input and output.
Thus interchanging the role of input and
output; from Fig. 5.11.2, we can draw single B,(n)
stage lattice structure of IIR system as shown Feedback
in Fig. 5.11.2.
Fig. 5.11.2

Note: IR systems always require feedback which is shown in Fig. 5.11.2.

Here we have interchanged x(n ) and y(n) ; thus the directions of arow [of x(1) and y(a
are also reversed. This same structure is drawn properly as shown in Fig.
5.11.3.
Forward

-A,(n) = x(n)

y(n) k.

B,(n)
Feedback

Fig. 5.11.3
 DigitalS i g n !

Toobtaintwo,
stage
lattice structure we will put N= 2 in Equation (5.11.4).
2

= - Bu (K) y(n-K)+ x (n ) ...(5.11.7)

y( n ) K=|

difterence equation of two stage lattice structure. We know that two stage lattice
the
is iby cascading two single stage lattice structures. It is shown
Ths in Fig. 5.11.4.

Ao(0)
+y(n)

Single stage Single stage

Fig. 5.11.4
timilarly we can draWn order lattice structure as shown in Fig.5.11.5.
Ag(n)
An-1(0) Aoln) = y(n) Output

KN K4

KN

B(n)
Fig. 5.11.5

a1 Lattice-ladder Structure of lIR Filter :

OIsider all pole IIR system. For this system the transfer function can be expressed as,

H(Z) = N ...(5.11.8)
1+ XBN (K)Z1
K=1
The e
difference equation tor this IIR system is

y(n) = x(n)- X B(K)y (n-K)

K=1
ThisS equation is
realized in lattice form as shown in Fig. 5.11.6.
 (PTU-Sem. E) 5-72 Structure for Realizalr
Digital Signal Processing
function,
with system
Consider an llR tilter
M
Ay(K) 2,-K
K=0
,N>M
H(Z)=
1+ BN (K)Z-K
K= 1

we
For numerator of Equation (5.11.9) in x(n) = Pol1)
structure as shown
have constructed a ladder
Fig. 5.11.6(a).
structure for
Thus,
lattice x(n-1}
be constructed by first
Equation (5.11.9) can coefficients for the
realizing an all-pole lattice B{1)
part as
denominator and then adding a ladder Fig. 5.11.6a)
shown in Fig. 5.11.6(6).
Agln) A(n)
x(n)
Anln)
KN

-k

By(n)

oN

realizing a pole-zero IlR Mter

(b) Lattice-ladder structure for
Fig. 5.11.6
combination of B,(0
The output of the ladder part can be expressed l as a weighted linear

Thus, output is given as,

N

y(n) = CBm (n)

m =0

where Cn are called the ladder coefficients.