20
1 Finite Impulse Response (FIR) Digital Filters (I) Types of linear phase FIR filters Yogananda Isukapalli

P10 Linear phase FIR - UC Santa Barbara

  • Upload
    others

  • View
    5

  • Download
    0

Embed Size (px)

Citation preview

Page 1: P10 Linear phase FIR - UC Santa Barbara

1

Finite Impulse Response (FIR)

Digital Filters (I)

Types of linear phase FIR filters

Yogananda Isukapalli

Page 2: P10 Linear phase FIR - UC Santa Barbara

2

Key characteristic features of FIR filters

1. The basic FIR filter is characterized by the following two equations:

å-

=

-=1

0

)()()(N

k

knxkhny å-

=

-=1

0

)()(N

k

kzkhzH

where h(k), k=0,1,…,N-1, are the impulse response coefficients of the filter, H(z) is the transfer function and N the length of the filter.

(1) (2)

2. FIR filters can have an exactly linear phase response.

3. FIR filters are simple to implement with all DSP processors available having architectures that are suited to FIR filtering.

Page 3: P10 Linear phase FIR - UC Santa Barbara

3

1. The phase delay ( ) of the filter is the amount of time delay each frequency component of the signal suffers in going through the filter.

2. The group delay ( ) is the average time delay the composite signal suffers at each frequency.

3. Mathematically,

• Consider a signal that consists of several frequency components passing through a filter.

pT

gT

wwq /)(-=pT

wwq ddTg /)(-=

where q(w) is the phase angle.

Linear phase response

(4)

(3)

Page 4: P10 Linear phase FIR - UC Santa Barbara

4

a

0 wc p w

k

0 wc p w

Phase

Magnitude

Example) Ideal Lowpass filter

• A filter is said to have a linear phase response if,

awbwJawwJ-=

-=)()( (5)

(6)where a and b are constants.

Page 5: P10 Linear phase FIR - UC Santa Barbara

5

H(ejw) =k e-jwa passband0 otherwise

Magnitude response = |H(ejw)| = k Phase response (q(w)) = < H(ejw) = -wa

Follows: y[n] = kx[n-a] : Linear phase implies that the output is a replica of x[n] {LPF} with a time shift of a

- p -wu -wl 0 wl wu p w

- p -wu -wl 0 wl wu p w

Page 6: P10 Linear phase FIR - UC Santa Barbara

6

Linear phase FIR filters• Symmetric impulse response will yield linear phase FIR filters.

1. Positive symmetry of impulse response:

h(n) = h(N-n-1), n = 0,1,…, (N-1)/2 (N odd)n = 0,1,…, (N/2)-1 (N even)

a = (N-1)/2 in eqn (5)

1. Negative symmetry of impulse response:

h(n) = -h(N-n-1),

a = (N-1)/2b = p/2 in eqn (6)

n = 0,1,…, (N-1)/2 (N odd)n = 0,1,…, (N/2)-1 (N even)

Page 7: P10 Linear phase FIR - UC Santa Barbara

7

h[n] = h[N-1-n] n = 0,1,….(N-1)/2

h[0] = h[10]

h[1] = h[9]

h[2] = h[8]

h[3] = h[7]

h[4] = h[6]

h[5] = h[5]

Example) For positive symmetry and N = 11 odd length

Page 8: P10 Linear phase FIR - UC Santa Barbara

8

0 1 2 3 4 5 6 7 8 9 10

N = 11

0 1 2 3 4 5 6 7 8 9

N = 10

h[n] = h[10-1-n] = h[9-n] h[0] = h[9]

h[1] = h[8]

h[2] = h[7]

h[3] = h[6]

h[4] = h[5]

Page 9: P10 Linear phase FIR - UC Santa Barbara

9

Example : Proof of Linear Phase :

4321

4

0

]4[]3[]2[]1[]0[)(

][)(

----

=

-

++++=

=åzhzhzhzhhzH

znhzHn

n

For N = 5; Symmetric impulse Response Implies :

h[n] = h[5-1-n] = h[4-n]

h[0] = h[4]

h[1] = h[3]

h[2] = h[2]

Now : +++= --- 212 ]]3[]1[[]2[)( zzhzhzhzH222 ]]4[]0[[ --++ zzhzh

Page 10: P10 Linear phase FIR - UC Santa Barbara

10

Consider Frequency Response :

+++=

==

---

=

wjjwjwwjjw

ezjw

eeheheheH

zHeH jw

22 ]]3[]1[[]2[)(

) 1T ( )()(

wjwjwj eeheh 222 ]]4[]0[[ --+

[ ]

|)(|

))2(cos(][2]2[[

2cos]0[2cos]1[2]2[[

2

1

0

2

2

wjwj

n

wj

wj

eHe

nwnhhe

whwhhe

q-

=

-

-

=

úû

ùêë

é-+=

++=

å

Phase = -2w

( Linear Phase form)

Page 11: P10 Linear phase FIR - UC Santa Barbara

11

Group Delay :

)( phasedwdTg-

=

= 2

passband

2 w

0 wp p w

H

0 wp p w

Group delay is constant over the passband for linear phase filters.

gT

Page 12: P10 Linear phase FIR - UC Santa Barbara

12

Types of FIR linear phase systems

1. Type I FIR linear phase systemThe impulse response is positive symmetric and N an

odd integer ],1[][ nNhnh --= 2/)1(0 -££ Nn

The frequency response is

å-

=

-=2/)1(

0

][)(N

n

jwnjw enheH

å-

=

--=2/)1(

0

2/)1( )cos(][)(N

n

Njwjw wnnaeeH

],)2/)1[((2][],2/)1[(]0[

nNhnaNha

where

--=-=

.2/)1,...(2,1 -= Nn

Page 13: P10 Linear phase FIR - UC Santa Barbara

13

2. Type II FIR linear phase systemThe impulse response is positive symmetric and

N is an even integer

],1[][ nNhnh --= 1)2/(0 -££ Nn

The frequency response is

å-

=

-=1)2/(

0

][)(N

n

jwnjw enheH

,)]21

(cos[][)(2/

1

2/)1(

ïþ

ïýü

ïî

ïíì

-= å=

--N

n

Njwjw nwnbeeH

],2/[2][ nNhnbwhere

-=

.2/,...2,1 Nn =

Page 14: P10 Linear phase FIR - UC Santa Barbara

14

3. Type III FIR linear phase systemThe impulse response is negative-symmetric and

N an odd integer.],1[][ nNhnh ---= 2/)1(0 -££ Nn

The frequency response is

å-

=

-=2/)1(

0

][)(N

n

jwnjw enheH

,)sin(][)(2/)1(

1

2/)1(

ïþ

ïýü

ïî

ïíì

= å-

=

--N

n

Njwjw wnnajeeH

],)2/)1[((2][ nNhnawhere

--=

.2/)1,...(2,1 -= Nn

Page 15: P10 Linear phase FIR - UC Santa Barbara

15

4. Type IV FIR linear phase systemThe impulse response is negative-symmetric and N an even integer.

],1[][ nNhnh ---= 1)2/(0 -££ Nn

The frequency response is

å-

=

-=1)2/(

0

][)(N

n

jwnjw enheH

,)]21(sin[][)(

2/

1

2/)1(

ïþ

ïýü

ïî

ïíì

-= å=

--N

n

Njwjw nwnbjeeH

],2/[2][ nNhnbwhere

-=

.2/,...,2,1 Nn =

Page 16: P10 Linear phase FIR - UC Santa Barbara

16

Fig: A summary of four types of linear phase FIR filters1

Page 17: P10 Linear phase FIR - UC Santa Barbara

17

Fig: A comparison of the impulse of the four types of linear phase FIR filters1

Page 18: P10 Linear phase FIR - UC Santa Barbara

18

• The frequency response of a Type 2 filter is always zero at f=0.5 ( half the sampling frequency as all the frequencies are normalized to the sampling frequency) and thus is unsuitable as a highpass filter.

• The frequency response of a Type 3 filter is always zero at f=0 and 0.5, while that of Type 4 filter is zero at f=0. Thus, Type 3 filter cannot be used as either a lowpass or high pass filter whereas Type 4 cannot be used as a lowpass filter.

• Both Type 3 and 4 filters introduce a 90o phase shift and are often used to design differentiators and Hilbert transformers.

• Type 1 is the most versatile of the four.

Notes:

Page 19: P10 Linear phase FIR - UC Santa Barbara

19

• The phase delay (for type 1 and 2 filters) or group delay (for all four types) is expressible in terms of the number of coefficients of the filter.

• Thus they can be corrected to give a zero phase or group delay response.

• For example,

TNTp ÷øö

çèæ -

=21 Phase delay for

types 1 and 2

TNTg ÷øö

çèæ --

=21 p Group delay for

types 3 and 4

where T is the sampling period.

Page 20: P10 Linear phase FIR - UC Santa Barbara

20

References

1. “Digital Signal Processing – A Practical Approach” -Emmanuel C. Ifeachor and Barrie W. Jervis Second Edition