Functions Functions
• It is nearly impossible to design a linear- • The above transfer function has a linear
phase IIR transfer function phase, if its impulse response h[n] is either
• It is always possible to design an FIR symmetric, i.e.,
transfer function with an exact linear-phase h[ n] = h[ N − n], 0 ≤ n ≤ N
response or is antisymmetric, i.e.,
• Consider a causal FIR transfer function H(z) h[ n] = −h[ N − n], 0 ≤ n ≤ N
of length N+1, i.e., of order N:
−n
H ( z) = ∑ N
n= 0 h[n ] z
1 2
Copyright © 2001, S. K. Mitra Copyright © 2001, S. K. Mitra
3 4 Type 3: N = 8 Type 4: N = 7
Copyright © 2001, S. K. Mitra Copyright © 2001, S. K. Mitra
1
Linear-Phase FIR Transfer Linear-Phase FIR Transfer
Functions Functions
• The corresponding frequency response is • The phase function here is given by
then given by θ(ω) = − 4 ω + β
H (e jω ) = e − j 4ω{2 h[ 0] cos( 4ω) + 2h[1] cos( 3ω) where β is either 0 or π, and hence, it is a
+ 2h[ 2] cos( 2ω) + 2h[3] cos( ω) + h[ 4]} linear function of ω in the generalized sense
• The quantity inside the braces is a real • The group delay is given by
function of ω, and can assume positive or d θ(ω )
τ(ω) = − dω = 4
negative values in the range 0 ≤ ω ≤ π indicating a constant group delay of 4 samples
7 8
Copyright © 2001, S. K. Mitra Copyright © 2001, S. K. Mitra
0.6
which is seen to be a cascade of a 2-point MA
0.4 filter with a 6-point MA filter
0.2
• Thus, H 0 (z ) has a double zero at z = −1, i.e.,
11
0
0 0.2 0.4 0.6 0.8 1
12
(ω = π)
ω/ π
Copyright © 2001, S. K. Mitra Copyright © 2001, S. K. Mitra
2
Linear-Phase FIR Transfer Linear-Phase FIR Transfer
Functions Functions
Type 2: Symmetric Impulse Response with • Making use of the symmetry of the impulse
Even Length response coefficients, the transfer function
• In this case, the degree N is odd can be written as
• Assume N = 7 for simplicity H ( z) = h[0 ](1 + z − 7 ) + h[1]( z −1 + z −6 )
• The transfer function is of the form + h[2]( z −2 + z −5 ) + h[3]( z −3 + z −4 )
H ( z) = h[ 0] + h[1]z −1 + h[2 ]z − 2 + h[3]z −3 = z −7 / 2{h[0]( z7 / 2 + z− 7 / 2 ) + h[1]( z 5 / 2 + z −5 / 2 )
−4 −5 −6 −7
+ h [4 ] z + h[5] z + h [6 ] z + h [7 ]z + h[2]( z3 / 2 + z −3 / 2 ) + h[3]( z1/ 2 + z−1/ 2 )}
13 14
Copyright © 2001, S. K. Mitra Copyright © 2001, S. K. Mitra
3
Linear-Phase FIR Transfer Linear-Phase FIR Transfer
Functions Functions
• The corresponding frequency response is • The group delay here is
given by τ( ω) = 4
H (e jω ) = e − j 4ωe − jπ / 2{2 h[ 0] sin( 4ω) + 2h[1]sin( 3ω) indicating a constant group delay of 4 samples
+ 2h[2 ]sin( 2 ω) + 2 h[3]sin( ω)} • In the general case
H (e jω ) = je − jNω / 2H (ω)
~
• It also exhibits a generalized phase response
given by where the amplitude response is of the form
θ (ω) = − 4 ω + π2 + β ~ N /2
H (ω) = 2 ∑ h[ N2 − n] sin( ω n)
where β is either 0 or π n =1
19 20
Copyright © 2001, S. K. Mitra Copyright © 2001, S. K. Mitra
4
Linear-Phase FIR Transfer Linear-Phase FIR Transfer
Functions Functions
• The magnitude and phase responses of the • Note that, even though the group delay is
linear-phase FIR are given by constant, since in general | H ( e j ω ) | is not a
constant, the output waveform is not a
| H ( e jω) | = H ( ω)
~
replica of the input waveform
− N ω + β, ~
for H (ω) ≥ 0 • An FIR filter with a frequency response that
2
θ(ω) = Nω ~ is a real function of ω is often called a zero-
− 2 + β − π, for H (ω) < 0 phase filter
• The group delay in each case is • Such a filter must have a noncausal impulse
response
25
τ(ω) = N2 26
Copyright © 2001, S. K. Mitra Copyright © 2001, S. K. Mitra
5
Zero Locations of Linear- Zero Locations of Linear-
Phase FIR Transfer Functions Phase FIR Transfer Functions
• It follows from the relation H ( z) = ± z− N H ( z−1) • Thus, a complex zero that is not on the unit
that if z = ξ o is a zero of H(z), so is z = 1/ ξ o circle is associated with a set of 4 zeros given
• Moreover, for an FIR filter with a real by
impulse response, the zeros of H(z) occur in z = re± jφ , z = 1r e ± jφ
complex conjugate pairs • A zero on the unit circle appear as a pair
• Hence, a zero at z = ξo is associated with a z = e ± jφ
zero at z = ξo* as its reciprocal is also its complex conjugate
31 32
Copyright © 2001, S. K. Mitra Copyright © 2001, S. K. Mitra
6
Zero Locations of Linear- Zero Locations of Linear-
Phase FIR Transfer Functions Phase FIR Transfer Functions
(4) Type 4 FIR filter: An odd number of • A Type 2 FIR filter cannot be used to
zeros at z = 1, and either an even number or design a highpass filter since it always has a
no zeros at z = −1 zero z = −1
• The presence of zeros at z = ±1 leads to the • A Type 3 FIR filter has zeros at both z = 1
following limitations on the use of these and z = −1, and hence cannot be used to
linear-phase transfer functions for designing design either a lowpass or a highpass or a
frequency-selective filters bandstop filter
37 38
Copyright © 2001, S. K. Mitra Copyright © 2001, S. K. Mitra
7
Bounded Real Transfer
Functions
• A causal stable real-coefficient transfer
function H(z) with | H ( e j ω ) | = 1 is thus
called a lossless bounded real (LBR)
transfer function
• The BR and LBR transfer functions are the
keys to the realization of digital filters with
low coefficient sensitivity
43
Copyright © 2001, S. K. Mitra