Introduction
Lecture 4
Solomon T.
Mekelle Institute of Technology-Mekelle
University (MIT-MU) 1/31/2019 1
Transfer function of a digital filter can be realized in different
ways.
Finite Impulse Response (FIR)-Non-recursive : the output is a
function of only past and present inputs
y(n) =b0x[n]+ b1x[n-1]+ b2x[n-2]+... bMx[n-M]
The order of the filter is M
The length of the filter M+1
There are M+1 multiplications
M additions
M memory locations
Mekelle Institute of Technology-Mekelle University (MIT-MU) 1/31/2019 2
Mekelle Institute of Technology-Mekelle University (MIT-MU) 1/31/2019 3
Mekelle Institute of Technology-Mekelle University (MIT-MU) 1/31/2019 4
Infinite Impulse Response (IIR)-Recursive : the output is a function
of past output, past and present inputs
y(n) =-a1y[n-1] -a2y[n-2]-... -aNx[n-N] +b0x[n]+ b1x[n-1]+ b2x[n-2]+...
bMx[n-M]
M+N additions
M+N+1 multiplications
M+N memory locations
Mekelle Institute of Technology-Mekelle University (MIT-MU) 1/31/2019 5
Mekelle Institute of Technology-Mekelle University (MIT-MU) 1/31/2019 6
Mekelle Institute of Technology-Mekelle University (MIT-MU) 1/31/2019 7
For FIR
H(z)=Y(z)/X(z)= b0+ b1z-1+ b2z-2+ b3z-3+... +bMz-M
For IIR
H(z)=Y(z)/X(z)= (b0+ b1z-1+ b2z-2+ b3z-3+... +bMz-
M)/(1+ a z-1+ a z-2+ a z-3+... +a z-N)
1 2 3 N
Mekelle Institute of Technology-Mekelle University (MIT-MU) 1/31/2019 8
1. Find the impulse response of the following system. Is it a FIR or IIR
filter? Stable or unstable ? y[n]=0.25y[n-1]+x[n]
2. Find the impulse response of the following system. Is it a FIR or IIR
filter? Stable or unstable ?y[n]=x[n]-2x[n-2]+ 5x[n-3]
Answer White board
Mekelle Institute of Technology-Mekelle University (MIT-MU) 1/31/2019 9
Note the fundamental building blocks:
Unit delay z−1
Multiplier/gain
Accumulator/summer/adder
Pick-off node
Mekelle Institute of Technology-Mekelle University (MIT-MU) 1/31/2019 10
Mekelle Institute of Technology-Mekelle University (MIT-MU) 1/31/2019 11
In general, we can form an equivalent structure by performing the
transpose operation, which is defined as follows:
Reverse all paths
Replace pick-off nodes with accumulators, and vice-versa
Interchange input and output
Mekelle Institute of Technology-Mekelle University (MIT-MU) 1/31/2019 12
If we perform this transpose operation on a direct form FIR filter, we get
the “direct form transposed” FIR filter structure:
Mekelle Institute of Technology-Mekelle University (MIT-MU) 1/31/2019 13
Mekelle Institute of Technology-Mekelle University (MIT-MU) 1/31/2019 14
Direct block diagram representation of an LTI recursive system described
by the difference equations
The transfer function H(z) of the IIR system is divided into two parts
connected in cascade, with the first part H1(z) containing only the zeroes,
and the second part H2(z) containing only the poles.
H(z)= H1(z) H2(z)
Realizing the two equations (H1(z) and H2(z)) using basic building blocks and
connecting them in cascade, we obtain the Direct Form – I structure as
follows
Mekelle Institute of Technology-Mekelle University (MIT-MU) 1/31/2019 15
Mekelle Institute of Technology-Mekelle University (MIT-MU) 1/31/2019 16
Mekelle Institute of Technology-Mekelle University (MIT-MU) 1/31/2019 17
Since, in a cascade arrangement, the order of the systems is not
important, the all – pole system H2(z) and the all - zero system H1(z) can
be interchanged .i.e.
Mekelle Institute of Technology-Mekelle University (MIT-MU) 1/31/2019 18
Mekelle Institute of Technology-Mekelle University (MIT-MU) 1/31/2019 19
Mekelle Institute of Technology-Mekelle University (MIT-MU) 1/31/2019 20
M+N additions
M+N+1 multiplications
Max(M,N) Memory locations
Mekelle Institute of Technology-Mekelle University (MIT-MU) 1/31/2019 21
A digital filter structure is said to be canonic if the number of delays in
the block diagram representation is equal to the order of the transfer
function
• Otherwise, it is a noncanonic structure
Example: y[n] =-d1 y[n-1] + p0x[n] +p1x[n-1]
noncanonical
Mekelle Institute of Technology-Mekelle University (MIT-MU) 1/31/2019 22
Realize the second order digital filter
y[n]=2rcos(w0)y[n-1] -r2y[n-2]+x[n]-rcos(w0) x[n-1]both in
direct form I and II
Answer White board
Mekelle Institute of Technology-Mekelle University (MIT-MU) 1/31/2019 23
A parallel form realization of an IIR system can be obtained by
performing partial – fraction expansion of H(z) .i.e.,
Where {pk}are the poles
H(z)= c +A1/(1-p1z-1)+ A2/(1-p2z-1) + A3/(1-p3z-1)+...+ AN/(1-pNz-1)
H(z) =Y(z)/X(z)= c + H1(z)+ H2(z)+.H3(z)+... HN(z)
Y(z) = cX(z) + H1(z) X(z)+ H2(z) X(z)+.H3(z) X(z)+... HN(z) X(z)
Mekelle Institute of Technology-Mekelle University (MIT-MU) 1/31/2019 24
The last equation can be realized as follows
Each of the subsystems can be realized either in Direct Form – I, or
Direct – Form II or transposed form.
Mekelle Institute of Technology-Mekelle University (MIT-MU) 1/31/2019 25
H(z)=(1+b11z-1+b21z-2)(1+b12z-1+b22z-2)(1+b13z-1+b23z-2)
H(z)=H1(z)H2(z)H3(z)
Mekelle Institute of Technology-Mekelle University (MIT-MU) 1/31/2019 26
H(z)=(1+b11z-1+b21z-2)(1+b12z-1+b22z-2)(1+b13z-1+b23z-2)/(1+a11z-1+a21z-
2)(1+a z-1+a z-2)(1+a z-1+a z-2)
12 22 13 23
H1(z) =(1+b11z-1+b21z-2)/(1+a11z-1+a21z-2)
H2(z) = (1+b12z-1+b22z-2)/(1+a12z-1+a22z-2)
H3(z) =(1+b13z-1+b23z-2)/(1+a13z-1+a23z-2)
H(z)=H1(z)H2(z)H3(z)
Mekelle Institute of Technology-Mekelle University (MIT-MU) 1/31/2019 27
Draw the cascade-form structures of the following filter
H(z)=(1- 2z-1 + 3z-2))(1 + 3z-5)
Answer (White board)
Draw the direct-form and cascade-form structures of the following filter
Mekelle Institute of Technology-Mekelle University (MIT-MU) 1/31/2019 28
