Signal Processing First

Lecture 13
Digital Filtering
of Analog Signals

02/15/21 © 2003, JH McClellan & RW Schafer 1

READING ASSIGNMENTS

 This Lecture:
 Chapter 6, Sections 6-6, 6-7 & 6-8

 Other Reading:
 Recitation: Chapter 6
 FREQUENCY RESPONSE EXAMPLES
 Next Lecture: Chapter 7

02/15/21 © 2003, JH McClellan & RW Schafer 3

LECTURE OBJECTIVES
 Two Domains: Time & Frequency
 Track the spectrum of x[n] thru an FIR
Filter: Sinusoid-IN gives Sinusoid-OUT
 UNIFICATION: How does Frequency
Response affect x(t) to produce y(t) ?

x(t) x[n] j̂ y[n] y(t)

A-to-D H (e ) D-to-A

FIR
̂ ̂
02/15/21 © 2003, JH McClellan & RW Schafer 4
 TIME & FREQUENCY
M M
y[n ]   bk x[n  k ]   h[k ] x[n  k ]
k 0 k 0
FIR DIFFERENCE EQUATION is the TIME-DOMAIN

M
H ( e )   h[k ] e
jˆ  jˆ k

k 0
jˆ  jˆ  j 2ˆ  j 3ˆ
H (e )  h[0]  h[1] e  h[2] e  h[3] e 
02/15/21 © 2003, JH McClellan & RW Schafer 5
Ex: DELAY by 2 SYSTEM
jˆ
Find h[n ] and H (e ) for y[n ]  x[n  2]
x[n] y[n] bk  { 0, 0, 1 }
h[n ]
h[n ]   [n  2]
x[n] j̂ y[n]
H (e )
̂ ̂
02/15/21 © 2003, JH McClellan & RW Schafer 6
 DELAY by 2 SYSTEM
jˆ
Find h[n ] and H (e ) for y[n ]  x[n  2]
x[n]
 [n  2] y[n]
M
H ( e )    [ k  2] e
jˆ  jˆ k

j̂ k 0
H (e )
x[n]  j 2̂ y[n] k = 2 ONLY
e
̂ ̂
02/15/21 © 2003, JH McClellan & RW Schafer 7
GENERAL DELAY PROPERTY
jˆ
Find h[n ] and H (e ) for y[n ]  x[n  nd ]
h[n ]   [n  nd ]
M
H ( e )    [k  nd ] e
jˆ  jˆ k
e  jˆ nd

k 0

ONLY ONE
non-ZERO TERM
for k at k = nd
02/15/21 © 2003, JH McClellan & RW Schafer 8
FREQ DOMAIN --> TIME ??
j̂
 START with H ( e ) and find h[n ] or bk
x[n]
h[n ] y[n]
h[n ]  ?
jˆ  j 2ˆ
H ( e )  7e cos(ˆ )
x[n] j̂ y[n]
H (e )
̂ ̂
02/15/21 © 2003, JH McClellan & RW Schafer 9
 FREQ DOMAIN --> TIME
jˆ  j 2ˆ
H ( e )  7e cos(ˆ ) EULER’s Formula

 j 2ˆ jˆ  jˆ

 7e (0.5e  0.5e )
 jˆ  j 3ˆ
 (3.5e  3.5e )

h[n ]  3.5 [n  1]  3.5 [n  3]

bk  { 0, 3.5, 0,3.5 }
02/15/21 © 2003, JH McClellan & RW Schafer 10
PREVIOUS LECTURE REVIEW

 SINUSOIDAL INPUT SIGNAL

 OUTPUT has SAME FREQUENCY
 DIFFERENT Amplitude and Phase
 FREQUENCY RESPONSE of FIR
MAG
 MAGNITUDE vs. Frequency
 PHASE vs. Freq PHASE
 PLOTTING
jˆ jˆ jH ( e jˆ )
H (e )  H (e ) e
02/15/21 © 2003, JH McClellan & RW Schafer 11
FREQ. RESPONSE PLOTS
 DENSE GRID (ww) from - to +
 ww = -pi:(pi/100):pi;
 HH = freqz(bb,1,ww)
 VECTOR bb contains Filter Coefficients
 DSP-First: HH = freekz(bb,1,ww)
M
H ( e )   bk e
jˆ  jˆ k

k 0
02/15/21 © 2003, JH McClellan & RW Schafer 12
PLOT of FREQ RESPONSE {bk }  {1,2,1}

̂
H ( e jˆ )  (2  2 cos ˆ )e  jˆ RESPONSE at /3


02/15/21 
̂ (radians)
© 2003, JH McClellan & RW Schafer 13
EXAMPLE 6.2
jˆ
Find y[n ] when H (e ) is known
j / 4 j ( / 3) n
and x[n ]  2e e
x[n] j̂ y[n]
H (e )
̂ ̂
jˆ  jˆ
H ( e )  ( 2  2 cos ˆ )e
02/15/21 © 2003, JH McClellan & RW Schafer 14
 EXAMPLE 6.2 (answer)
j / 4 j ( / 3) n
Find y[n ] when x[n ]  2e e
jˆ
One Step - evaluate H ( e ) at ˆ   / 3
jˆ  jˆ
H ( e )  ( 2  2 cos ˆ )e
jˆ  j / 3
H (e )  3e @ ˆ   / 3
y[n ]   3e  j / 3
  2e j / 4
e j ( / 3) n
 6e  j / 12 j ( / 3) n
e
02/15/21 © 2003, JH McClellan & RW Schafer 15
EXAMPLE: COSINE INPUT
jˆ
Find y[n ] when H ( e ) is known
 
and x[n ]  2 cos( 3 n  4 )
x[n] j̂ y[n]
H (e )
̂ ̂
jˆ  jˆ
H ( e )  ( 2  2 cos ˆ )e
02/15/21 © 2003, JH McClellan & RW Schafer 16
 EX: COSINE INPUT (ans-1)

Find y[n ] when x[n ]  

2 cos( 3 n )
4
j ( n / 3 / 4 )  j ( n / 3 / 4 )
2 cos( 3
n e )
4
e
 x[n ]  x1[n ]  x2 [n ]
j / 3 j ( n / 3 / 4 )
y1[n ]  H ( e )e
 j / 3  j ( n / 3 / 4 )
y2 [n ]  H ( e )e
02/15/21
 y[n ]  y1[n ]  y2 [n ]
© 2003, JH McClellan & RW Schafer 17
 EX: COSINE INPUT (ans-2)

Find y[n ] when x[n ]  

2 cos( 3 n )
4
jˆ  jˆ
H (e )  ( 2  2 cos ˆ )e
j / 3 j ( n / 3 / 4 )  j ( / 3) j ( n / 3 / 4 )
y1[n ]  H ( e )e  3e e
 j / 3  j ( n / 3 / 4 ) j ( / 3)  j ( n / 3 / 4 )
y2 [n ]  H ( e )e  3e e
j ( n / 3 / 12 )  j ( n / 3 / 12 )
y[n ]  3e  3e
 
 y[n ]  6 cos( 3 n  12 )
02/15/21 © 2003, JH McClellan & RW Schafer 18
 SINUSOID thru FIR
* jˆ  jˆ
 IF H ( e )  H ( e )
 Multiply the Magnitudes

 Add the Phases

x[n ]  A cos(ˆ1n   )
jˆ1 jˆ1
 y[ n ]  A H ( e ) cos(ˆ1n     H (e ))

02/15/21 © 2003, JH McClellan & RW Schafer 19

 LTI Demo with Sinusoids
y[n]
x[n]
FILTER

02/15/21 © 2003, JH McClellan & RW Schafer 20

DIGITAL “FILTERING”

x(t) x[n] j̂ y[n] y(t)

A-to-D H (e ) D-to-A

 ̂ ̂ 
  SPECTRUM of x(t) (SUM of SINUSOIDS)
 SPECTRUM of x[n]
̂  Is ALIASING a PROBLEM ?
 SPECTRUM y[n] (FIR Gain or Nulls)
  Then, OUTPUT y(t) = SUM of SINUSOIDS
02/15/21 © 2003, JH McClellan & RW Schafer 21
FREQUENCY SCALING

x(t) x[n] j̂ y[n] y(t)

A-to-D H (e ) D-to-A

 ˆ
 ˆ
 

 TIME SAMPLING: t  nTs

 IF NO ALIASING:
 FREQUENCY SCALING
ˆ  Ts  
fs
02/15/21 © 2003, JH McClellan & RW Schafer 22
 11-pt AVERAGER Example
x(t) x[n] j̂ y[n] y(t)
A-to-D H (e ) D-to-A

10
 ˆ ˆ 
y[ n ]    1
11
x[n  k ] 
k 0
250 Hz
sin( 2 ˆ )
?
11
jˆ  j 5ˆ
25 Hz H (e )  e
11 sin( 2 ˆ )
1

x (t )  cos(2 ( 25)t )  cos(2 ( 250)t  12  )

02/15/21 © 2003, JH McClellan & RW Schafer 23
D-A FREQUENCY SCALING

x(t) x[n] j̂ y[n] y(t)

A-to-D H (e ) D-to-A

 ˆ
 ˆ
 

 TIME SAMPLING: t  nTs  n  t f s

 RECONSTRUCT up to 0.5fs
 FREQUENCY SCALING   ˆ f s
02/15/21 © 2003, JH McClellan & RW Schafer 24
 TRACK the FREQUENCIES
x(t) x[n] j̂ y[n] y(t)
A-to-D H (e ) D-to-A

 ˆ
 ˆ
 

 250 Hz  0.5 H (e j 0.5 )  0.5  250 Hz

j 0.05
 25 Hz  .05 H (e )  .05  25 Hz

Fs = 1000 Hz NO new freqs

02/15/21 © 2003, JH McClellan & RW Schafer 25
11-pt AVERAGER

NULLS or ZEROS

ˆ  0.05 ˆ  0.5

02/15/21 © 2003, JH McClellan & RW Schafer 26

 EVALUATE Freq. Response
sin( 11  ˆ )  j 5ˆ
jˆ
H (e )  2
e
11 sin( 2 ˆ )
1
At ˆ  0.5 11 ( 0.5 ))
jˆ sin(  j 5( 0.5 )
H (e )  2
e
11 sin( 12 (0.5 ))
sin(2.75 )  j 2.5
 e
11 sin(0.25 )
 j 0.5
 0.0909e
02/15/21 © 2003, JH McClellan & RW Schafer 27
 EVALUATE Freq. Response

H ( e j 2 ( 25) /1000 )
MAG SCALE
fs = 1000
PHASE CHANGE

H ( e j 2 ( 250 ) /1000 )

02/15/21 © 2003, JH McClellan & RW Schafer 28

 DIGITAL FILTER
j̂
H (e )

EFFECTIVE RESPONSE
LOW-PASS FILTER

02/15/21 © 2003, JH McClellan & RW Schafer 29

FILTER TYPES
 LOW-PASS FILTER (LPF)
 BLURRING
 ATTENUATES HIGH FREQUENCIES
 HIGH-PASS FILTER (HPF)
 SHARPENING for IMAGES
 BOOSTS THE HIGHS
 REMOVES DC
 BAND-PASS FILTER
(BPF)
02/15/21 © 2003, JH McClellan & RW Schafer 30
B & W IMAGE

02/15/21 © 2003, JH McClellan & RW Schafer 31

B&W IMAGE with COSINE
FILTERED: 11-pt AVG

02/15/21 © 2003, JH McClellan & RW Schafer 32

FILTERED B&W IMAGE

LPF:
BLUR

02/15/21 © 2003, JH McClellan & RW Schafer 33

ROW of B&W IMAGE
BLACK = 255

WHITE = 0

02/15/21 © 2003, JH McClellan & RW Schafer 34

FILTERED ROW of IMAGE

ADJUSTED DELAY by 5 samples

02/15/21 © 2003, JH McClellan & RW Schafer 35