0% found this document useful (0 votes)
566 views2 pages

Symmetric and Antisymmetric FIR Filters

This document discusses linear-phase filters and the conditions required for a filter to exhibit linear-phase characteristics. It states that for a filter to be linear-phase, its impulse response must have a finite duration and be either symmetric or antisymmetric about its midpoint. It provides examples of impulse responses that satisfy these symmetry conditions. It also discusses the Paley-Wiener theorem and its implications for ideal causal filters.

Uploaded by

Aniket Mane
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
566 views2 pages

Symmetric and Antisymmetric FIR Filters

This document discusses linear-phase filters and the conditions required for a filter to exhibit linear-phase characteristics. It states that for a filter to be linear-phase, its impulse response must have a finite duration and be either symmetric or antisymmetric about its midpoint. It provides examples of impulse responses that satisfy these symmetry conditions. It also discusses the Paley-Wiener theorem and its implications for ideal causal filters.

Uploaded by

Aniket Mane
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 2

Taling iers DTET «t heh te sies

od auatia (S4), we Neven Mn


)=ris-a)
Ts de ugt od the
S5)
Hee e kowpas ner
DOtHing but
berptase has ax atered tte
the input dclayad by an N-
a Coevesy i e pui d no been
shae of
nes, the
r{r), simply translated it by amo
an amunt
output Ho.S.3 An Impulse reiponie h ) thot h symmekc wih reapect to mdpol AvwN.
signal would have een a diston even
We wish to infam the readers
that a cusal IR
fher
pes of FIR filher that exhibit center cannot gie
symmetry in its impulse response linear-phase
a
give the
characteristic and only sp N oxdd
M
5.2 PALEY-WIENER THEOREM linear-phase.
If)s causal sequence with hAn) =
0 for n <0 and it has finite
energy, then N-I
+ Fkg. 5.4: An Impulse reiponse h{n) that h iymmec wih respect to mdpont lor N odd.
The Z-transfom ofa causal finite duration scquence Hn) that begins at n-0 and endsat N- I is givenby
wbere
Ha()= DTFTA,m)}
The imnedine H)- h(n):""
implication of this result is that an ideal filter cannot be causal,
(5.7)
interval of freauencies ( a . s shown in because it is zero within an
Fig 5.2, and therefore
integral in equation (5.6) would be infinite. Referring Fig. 5.3, the above summation can be broken into two surnmations as given below
HA
"
H) 2htni: "2htn): ($ )
Hog1H (w)|= log (0) =-» log | ()ldo =**
Letting m N - I - n in the second summation on the right-hand side of the above cquation, we get,
Fig. 5.2: Ideal owposs fer and thee Paley-Wiener condiion.
One H(:)-hHin"" HN-1-m)»N-|-m)
important inference we make from Paley-Wiener theorem is that
but t cannot be HAoN can be zero at some frequencies, -0
zero over
any finite band of frequencies, because the integral then becomes unbonded.
5.3 STMMETRIC AND ANmSYMMETRIC FIR FurERs
Tf 2 hn: 2 h(N-1- m):(N-I-m) (5.9)
Km) represents the impulse response of a discrete-time flter, a necessary and suficient condition m-0 m-0
that for lincar-phase
is Hn) must have a finite duration N and is either Since m is a dummy variable, it can be replaced by n. Accordingly cquation (5.9) becomes,
or symmetric antisymmetric about its midpoint.
Mathematically, n) = HN - 1 - n)
(Symmetric), n 0,1,.N-1
H:)= 2 hn)""h(N-1-n)x-(M-1-n) (5.10)
or
n:0
hn)--H(N-1-n) [Antisymmetric) n=0,1,,N-1 Since,
h (N- 1 - n) = h{n) for n = 0, 1,.., N- I
gives a linear-phase FIR ilter.
Let us now prove that a
linear-phase filter results if symmetry condition for hn) is met with. Fig. 5.3 and 5.4 we get, H)- hn) N
n=0
illustrate the general shapes of h(n) that give a linear-phase. he frequency response of the digital filter is obtained by letting = el in H(z).
Design of HR Fiters 335
DigitalSignal r o c e s s n

334

ein HT). A ordingly


p o s e ot the
Accordingly we get, we get, digital filter represented by Hi:) is obtained by letting =

H(el@) H(o)= =
h(ne jan +e"jo{N-+)] H(e) Ho).
n=

2hn)e
n=0

(5.11)

n=0 (N-3)
Comparing equation (5.11) with 5..16)

H(o) = HwjeMo)

N-3
where, HA») is a real function of o, we get,
Hence,

H,)- 2 2tm)cos
if H,)>0 (517)
and @) =

0(w)-
- if H,(o)>0

and
S.12)
-). if H,(o)< 0
inspection of cquations (5.12) and (5.17) reveals that ( ) is same irespective of whether
N

h{N
is even or

1 -
odd-

n) te
the symmetry condition, h{n)
=
-

uie summation in equation (5.7) can be broken in to the following forn NISO. neSe cquations reveal that the phase ( ) is linear when
Refeming Fig. 5.4,
n0,1., N - 1 ismet with.
(-3)
(S 13) be ant1symimetric. For N Odd, tne miapoin.
Hie)-2 t:". M When {n) =
-k{N- I
n--1)
-

n) the unit impulse response is said to

the antisymmetric h{n) is n=

Letting m =N-I - n in the third term on the right hand side of the equation (5.13), we get
Consequently.
(N-3) in a similar
mar
(- term of opposite sign. Proceed1ng
2 , HN-1-m):4N-1-m) But, it N is even, each term
in Hn) has a matching be expressed as
antisymmetnic {n)
can

ve did for a symmetric Hn), the frequency response of an

Since, m dummy
a varnable, it can be replaced byn and also the order ofsummation could be reversed. Accordingly. H(e)= H(o) = H,(0)e)
the above equation becomes,
(N-3) (N-3)
(N-3)
2 hN-1-nkN--a) (5.14)
where

Making use of symmetry condition: N-I -

n) h{n), we get,
N even
(N-3)
and H,o)-22 nsn
Ha- Mol".*].a).1T) (5.15)

You might also like