Input

DIGITAL FILTER
STRUCTURES

By:
Prof. Yogesh B. Patel
Assistant Professor
Output EC Dept., UVPCE
Reference book
 Digital Signal Processing by N. G. Palan (Tech-Max publication)
 Digital Signal Processing fundamentals and applications by Li Tan
(Elsevier publication)
 Discrete Time Signal Processing, Oppenheim and Schaffer (PHI
publication)
 Digital Signal Processing-Principles, algorithms and applications By
Proakis and Manolakis (PHI publication)
 IIR and FIR Filter
Input/output Relation (Time Domain) Form
Infinite Impulse Response (IIR) Filter, Finite Impulse Response (FIR) Filter,
Using convolution Using convolution
∝ M
y ( n) = ∑ h( k ) x ( n − k ) 0≤ n ≤ ∞ y ( n) = ∑ h( k ) x ( n − k ) 0≤ n ≤ N-1
k = −∝ k =0

Using Difference equation, Using Difference equation,

M
y[n] = ∑ bk x[n − k ]
N M
y[n] = −∑ ak y[n − k ] + ∑ bk x[n − k ] n≥ 0 n≥ 0
k =1 k =0 k =0

Transfer function form Transfer function form

M

∑ k
b z −k
M
H (z ) =
Y ( z)
H (z ) =
Y ( z)
= k =0 = ∑ bk z − k
N
X ( z ) k =0
1 + ∑ ak z − k
X ( z)
k =1
Here, ak and bk are the coefficients of the filter which have to be selected
suitably to make the filter to achieve the desired output.
Syllabus:
Basic IIR filter structures:
Direct forms – I
 Direct forms – II
Cascade realizations
Parallel realizations
Basic FIR filter structures
 Direct form
 Cascade realizations
 Linear phase FIR structure
 Lattice structure
Signal flow graph
Transposed structure
Block diagram representation of
y[n ] = a1 y[n − 1] + a2 y[n − 2] + b0 x[n ]
 Block Diagram Representation
• LTI systems with rational system
function can be represented as
constant-coefficient difference
equation
• The implementation of difference
equations requires delayed values of
the
– input
– output
– intermediate results
• The requirement of delayed elements
implies need for storage
• We also need means of
– addition
– multiplication
 Advantages
Observing the block diagram; computation algorithm
can be easily written
Relationship between input and output can be
determined
Memory requirement and computational complexity
can be determined
Variety of equivalent block diagram representation
can be determined
Canonic Structure :

Number of delay elements = Order of difference

in block diagram equation of digital filter

Noncanonic Structure :

Number of delay elements ≠ Order of difference

in block diagram equation of digital filter
 Difference Equation and Digital Filtering
Digital input Digital output
x(n) Digital filter
y(n)
(digital filtering)

y (=
n) b0 x (n) + b1 x (n − 1) +  + bM x (n − M )
− a1 y (n − 1) −  − aN y (n − N )
General form of difference equation
N M

∑ a y[n − k ] = ∑ b x[n − k ]
k =0
k
k =0
k
Where ak and bk are constants
Alternative equivalent form
N M
y[n] + ∑ ak y[n − k ] = ∑ bk x[n − k ]
k =1 k =0 M

Transfer function can be written as ∑k

b z −k

H (z ) =
Y ( z)
= k =0
N
1 + ∑ ak z − k
X ( z)
k =1
 Basic Structures
for
IIR Systems

Direct form structure

Cascade form structure
Parallel form structure
Direct form structure for IIR Systems

Direct form - I
Direct form - II
 Direct form - I
M

∑ k
b z −k

H (z ) = k =0
N
1 + ∑ ak z − k
k =1
M
H 1 ( z ) = ∑ bk z − k
k =0

H 2 (z ) =
1
N
1 + ∑ ak z − k
k =1

H ( z ) = H 1 ( z ).H 2 ( z )
Cont…
Direct form structure for H1(z) (All zero system):

W ( z) M
H1 (z ) = = ∑ bk z − k
X ( z ) k =0

W ( z ) = b0 X ( z ) + b1 z −1 X ( z ) + ........ + bM z − M X ( z )
Taking Inverse Z-transform
w(n) = b0 x(n) + b1 x(n − 1) + ........ + bM x(n − M )

Cont…
w(n) = b0 x(n) + b1 x(n − 1) + ........ + bM x(n − M )
Where b0, b1, b2 ….. bM are coefficients

Cont…
Direct form structure for H2(z) (All pole system):
H 2 (z ) =
Y ( z) 1
= N
1 + ∑ ak z − k
W ( z)
k =1

 N

Y ( z )1 + ∑ ak z − k  = W ( z )
 k =1 
 N −k 
Y ( z ) +  ∑ ak z Y ( z ) = W ( z )
 k =1 
 N 
Y ( z ) = − ∑ ak z − k Y ( z ) + W ( z )
 k =1 

Y ( z ) = −a1 z −1Y ( z ) − a2 z −2Y ( z )........ − a N z − N Y ( z ) + W ( z )

Taking Inverse Z-transform
y (n) = −a1 y (n − 1) − a2 y (n − 2)........ − a N y (n − N ) + w(n)
y (n) = −a1 y (n − 1) − a2 y (n − 2)........ − a N y (n − N ) + w(n)
Direct form - I of IIR System
 Computational complexity of Direct
form - I
No. of Multiplications = M+N+1
No. of additions = M+N
No. of memory locations = M+N+1
 Direct form - II
M

∑b z k
−k

H (z ) = k =0
N
1 + ∑ ak z − k
k =1

H1 (z ) =
1
N
1 + ∑ ak z −k

k =1
M
H 2 ( z ) = ∑ bk z −k

k =0

H ( z ) = H 1 (Z ).H 2 (Z )
All pole system:

H1 ( z ) =
U ( z) 1
= N
1 + ∑ ak z − k
X ( z)
k =1

 N
−k 
U ( z )1 + ∑ ak z  = X ( z )
 k =1 
 N −k 
U ( z ) = X ( z ) − U ( z ) ∑ ak z 
 k =1 
U ( z ) = X ( z ) − a1 z −1U ( z ) − a2 z − 2U ( z )........ − a N z − NU ( z )
Taking Inverse Z-transform
u (n) = x(n) − a1u (n − 1) − a2u (n − 2)........ − a N u (n − N )
u (n) = x(n) − a1u (n − 1) − a2u (n − 2)........ − a N u (n − N )
All zero system:
Y ( z) M
H 2 (z ) = = ∑ bk z − k
U ( z ) k =0
Y ( z ) = b0U ( z ) + b1 z −1U ( z ) + ........ + bM z − M U ( z )
Taking Inverse Z-transform
y (n) = b0u (n) + b1u (n − 1) + ........ + bM u (n − M )
Direct form - II of IIR System
 Direct form - II of IIR System
If we consider N=M
 Computational Complexity
No. of Multiplications = M+N+1
No. of additions = M+N
No. of memory locations reduced (delay elements)

• No need to store the same data twice in previous system

• So we can collapse the delay elements into one chain
• This is called Direct Form II or the Canonical Form
• Theoretically no difference between Direct Form I and II
• Implementation wise
– Less memory in Direct II
 Cascade Combination for IIR System

H ( z ) = H 1 ( z ).H 2 ( z ).........H k ( z )
M

∑ k
b z −k

H (z ) = k =0
N
1 + ∑ ak z − k
k =1
2

∑b k z −k

H k (z ) = k =0
2
1 + ∑ ak z − k
k =1

b0 + b1 z −1 + b2 z − 2
H k (z ) =
1 + a1 z −1 + a2 z − 2
Parallel Combination for IIR System
M

∑ k
b z −k

H (z ) = k =0
N
1 + ∑ ak z − k
k =1

H ( z ) = C + H 1 ( z ) + H 2 ( z ) + ......... + H k ( z )
Representation of structures using Signal flow Graphs
Transposed Structure
 Transposing doesn’t change the input-output relation

Step:
 Reverse directions of all branches
 Change branching nodes into adders and vice-versa
 Interchange input and output
Basic FIR filter structures

Direct form
Cascade realizations
Linear phase FIR structure
Lattice structure
FIR Systems
Using convolution
M
y ( n) = ∑ h( k ) x ( n − k ) 0 ≤ n ≤ N -1
k =0

The Difference equation of FIR System

M
y[n] = ∑ bk x[n − k ] n = 0 n≥ 0
k =0
Taking Z-transform
M
Y (Z ) = ∑ bk z − k X (Z )
k =0

System Transfer function form

M
H (z ) =
Y ( z)
= ∑ bk z − k
X ( z ) k =0
 Direct Form
M
y ( n) = ∑ h( k ) x ( n − k )
k =0

Expanding Equation
y ( n) = h(0) x ( n) + h(1) x ( n − 1) + h( 2) x ( n − 2) + .....
... + h( M − 1) x( n − M + 1) + h( M ) x( n − M )

Observation
 Delay block = M
 Addition = M
 Multiplications = M+1
Signal flow graph representation of FIR system
FIR system Direct form realization also represented in this way known as
tapped or transversal delay line

Signal flow graph representation of tapped or transversal delay line

Example:
Obtain direct form realization for the transfer function of an FIR system
given by
𝐻𝐻 𝑧𝑧 = 1 − 0.2𝑧𝑧 −1 − 0.15𝑧𝑧 −2 (1 − 0.125𝑧𝑧 −1 − 0.5𝑧𝑧 −2 )

Solution:
𝑌𝑌(𝑧𝑧)
𝐻𝐻 𝑧𝑧 = = 1 − 0.325𝑧𝑧 −1 − 0.625𝑧𝑧 −2 + 0.11875𝑧𝑧 −3 + 0.075𝑧𝑧 −4
𝑋𝑋(𝑧𝑧)
𝑌𝑌 𝑧𝑧 = 𝑋𝑋 𝑧𝑧 − 0.325𝑧𝑧 −1 𝑋𝑋 𝑧𝑧 − 0.625𝑧𝑧 −2 𝑋𝑋 𝑧𝑧 + 0.11875𝑧𝑧 −3 𝑋𝑋 𝑧𝑧 + 0.075𝑧𝑧 −4 𝑋𝑋 𝑧𝑧

Taking inverse z-transform

𝑦𝑦 𝑛𝑛 = 𝑥𝑥 𝑛𝑛 − 0.325𝑥𝑥 𝑛𝑛 − 1 − 0.625𝑥𝑥 𝑛𝑛 − 2 + 0.11875𝑥𝑥 𝑛𝑛 − 3 + 0.075𝑥𝑥 𝑛𝑛 − 4
 Cascade Form of FIR system
 Higher order FIR filter is realized by using a series connection of different FIR
section (Cascade connection)

( )
M MS
H ( z ) = ∑ h[n]z −n
= ∏ b0 k + b1k z −1 + b2 k z − 2
n =0 k =1
 Each section is characterized by the second order transfer function

H (z ) = H1 ( z ).H 2 ( z ).H 3 ( z )........H k ( z )

 Here Hk(z) is 2nd order transfer function

Yk ( z )
H k (z ) = = bk 0 + bk1 z −1 + bk 2 z − 2
X k (z )
Yk (z ) = bk 0 X k (z ) + bk1 z −1 X k (z ) + bk 2 z − 2 X k ( z )

 Taking Inverse Z-transform of both side

yk (n ) = bk 0 xk (n ) + bk1 xk (n − 1) + bk 2 xk (n − 2 )
 yk (n ) = bk 0 xk (n ) + bk1 xk (n − 1) + bk 2 xk (n − 2 )

Block diagram of Cascade structure

Cascade realization of FIR system

 Example:
Obtain cascade form realization using second order section for the transfer
function of an FIR system given by
𝐻𝐻 𝑧𝑧 = 1 − 0.2𝑧𝑧 −1 − 0.15𝑧𝑧 −2 (1 − 0.125𝑧𝑧 −1 − 0.5𝑧𝑧 −2 )
Solution:
𝐻𝐻 𝑧𝑧 = 𝐻𝐻1 𝑧𝑧 . 𝐻𝐻2 (𝑧𝑧)

𝑌𝑌1 (𝑧𝑧) 𝑌𝑌2 (𝑧𝑧)

𝐻𝐻1 𝑧𝑧 = = 1 − 0.2𝑧𝑧 −1 − 0.15𝑧𝑧 −2 𝐻𝐻2 𝑧𝑧 = = 1 − 0.125𝑧𝑧 −1 − 0.5𝑧𝑧 −2
𝑋𝑋1 (𝑧𝑧) 𝑋𝑋2 (𝑧𝑧)
𝑌𝑌2 (𝑧𝑧)
𝑌𝑌1 𝑧𝑧 = 𝑋𝑋1 (𝑧𝑧) − 0.2𝑧𝑧 −1 𝑋𝑋1 (𝑧𝑧) − 0.15𝑧𝑧 −2 𝑋𝑋1 (𝑧𝑧) 𝑌𝑌2 𝑧𝑧 = = 𝑋𝑋2 (𝑧𝑧) − 0.125𝑧𝑧 −1 𝑋𝑋2 (𝑧𝑧) − 0.5𝑧𝑧 −2 𝑋𝑋2 (𝑧𝑧)
𝑋𝑋2 (𝑧𝑧)

Taking inverse z-transform

Linear-Phase FIR Systems
Order of filter = M Length of filter = L = M+1
Type I : Symmetric impulse response Type III : Symmetric impulse response
and M is even and M is odd

h[M - n] = h[n] for all n h[M - n] = h[n] for all n

𝑀𝑀 𝑀𝑀 𝑀𝑀 𝑀𝑀 − 1 𝑀𝑀 + 1
𝑛𝑛 = 0 ⋯ ⋯ � − 1� ⋯ ⋯ ⋯⋯� + 1� ⋯ ⋯ M 𝑛𝑛 = 0 ⋯ ⋯ � �⋯⋯� �⋯⋯M
2 2 2 2 2

Type II : Antisymmetric impulse response Type IV : Antisymmetric impulse response

and M is even and M is odd

h[M - n] = - h[n] for all n h[M - n] = - h[n] for all n

𝑀𝑀 𝑀𝑀 𝑀𝑀 𝑀𝑀 − 1 𝑀𝑀 + 1
𝑛𝑛 = 0 ⋯ ⋯ � − 1� ⋯ ⋯ ⋯⋯� + 1� ⋯ ⋯ M 𝑛𝑛 = 0 ⋯ ⋯ � �⋯⋯� �⋯⋯M
2 2 2 2 2
Linear Phase FIR filter structure:

Symmetry condition of Impulse response to make FIR filter linear phase

h(n) = h(M - n) ; 0 ≤ n ≤ M (Type – 1 and 2)

Antisymmetry condition of Impulse response to make FIR filter linear phase

h(n) = - h(M - n) ; 0 ≤ n ≤ M (Type – 3 and 4 )

Difference equation of FIR filter

M

y n = � h k x(n − k)
k=0

Advantage :
Due to Symmetry and antisymmetry impulse response half the number of
multiplications require
Type-1: Symmetric impulse response and Order of filter (M) is Even
M
−1 M
2
M M
y n = � h k x n−k +h x(n − ) + � h k x n − k
2 2
k=0 M
k= +1
2
M M
−1 −1
2 2
M M
y n = � h k x n−k +h x(n − ) + � h M − k x n − M + k
2 2
k=0 k=0
Symmetry condition of Impulse response to make FIR filter linear phase h(k) = h(M-k) ; 0 ≤ k ≤ M

M M
−1 −1
2 2
M M
y n = � h k x n−k +h x(n − ) + � h k x n − M + k
2 2
k=0 k=0
M
−1
2
M M
y n = � h k [x n − k + x n − M + k ] + h x(n − )
2 2
k=0
Type-2: Antisymmetric impulse response and Order of filter (M) is Even
M
−1 M
2
M M
y n = � h k x n−k +h x(n − ) + � h k x n − k
2 2
k=0 M
k= +1
2
M
Here, h =0
2
M M
−1 −1
2 2
y n = � h k x n−k + � h M−k x n−M+k
k=0 k=0
Antisymmetry condition of Impulse response to make FIR filter linear phase h(n) = - h(M-n) ; 0 ≤ n ≤ M
M M
−1 −1
2 2
y n = � h k x n−k − � h k x n−M+k
k=0 k=0
M
−1
2
y n = � h k [x n − k − x n − M + k ]
k=0
Type-3: Symmetric impulse response and Order of filter (M) is Odd
M−1
2 M

y n = � h k x n−k + � h k x n−k
k=0 M+1
k=
2
M−1 M−1
2 2
y n = � h k x n−k + � h M−k x n−M+k
k=0 k=0
Symmetry condition of Impulse response to make FIR filter linear phase h(k) = h(M-k) ; 0 ≤ k ≤ M
M−1 M−1
2 2
y n = � h k x n−k + � h k x n−M+k
k=0 k=0
M−1
2
y n = � h k [x n − k + x n − M + k ]
k=0
Type-4: Antisymmetric impulse response and Order of filter (M) is odd and
M−1
2 M

y n = � h k x n−k + � h k x n−k
k=0 M+1
k=
2
M−1 M−1
2 2
y n = � h k x n−k + � h M−k x n−M+k
k=0 k=0
Antisymmetry condition of Impulse response to make FIR filter linear phase h(k) = - h(M-k) ; 0 ≤ k ≤ M
M−1 M−1
2 2
y n = � h k x n−k − � h k x n−M+k
k=0 k=0
M−1
2
y n = � h k [x n − k − x n − M + k ]
k=0
Lattice structure of FIR filter: