Nguyễn Công Phương

SIGNAL PROCESSING

Sampling
of Continuous – Time Signals
 Contents
I. Introduction
II. Discrete – Time Signals and Systems
III. The z – Transform
IV. Fourier Representation of Signals
V. Transform Analysis of LTI Systems
VI. Sampling of Continuous – Time Signals
VII.The Discrete Fourier Transform
VIII.Structures for Discrete – Time Systems
IX. Design of FIR Filters
X. Design of IIR Filters
XI. Random Signal Processing
s i tes.google.com/site/ncpdhbkhn 2
Analog input Sensor

Analog Pre-processing

ADC

DSP

DAC
Analog
Analog Post-processing
output
s i tes.google.com/site/ncpdhbkhn 3
 Sampling of
Continuous – Time Signals
1. Ideal Periodic Sampling of Continuous –
Time Signals
2. Reconstruction of a Bandlimited Signal from
its Samples
3. The Effect of Undersampling: Aliasing
4. Discrete – Time Processing of Continuous –
Time Signals
5. Practical Sampling and Reconstruction
6. Sampling of Bandpass Signals
s i tes.google.com/site/ncpdhbkhn 4
 Ideal Periodic Sampling of
Continuous – Time Signals (1)
x c (t ) x[n ] = xc (nT )
Ideal analog – digital converter
Fs = 1/T

x[n ] = x c (t ) t = nT = xc (nT ), −∞<n<∞

0. 8

0. 6

0. 4

0. 2

-0 .2

-0 .4

-0 .6

-0 .8

0 1 2 3 4 5 6

s i tes.google.com/site/ncpdhbkhn 5
 Ideal Periodic Sampling of
Continuous – Time Signals (2)
 X ( jΩ ) = ∞
 ←→  x (t) = 1 X ( jΩ)e
∞
Ω

 
− jΩ t CTFT jΩt
x ( t ) e dt  d 
 c −∞ c   2π
c −∞ c

∞
 − jω n   1 π 
 X ( e jω
) =  x [ n ] e  ←DTFT
→  x[ n ] =
π −π
X ( e jω
) e jω n
d ω 
 n =−∞  2
F
ω = ΩT = 2π FT = 2π = 2π f
Fs
1 ∞  2π 
X ( e ) =  X c  jΩ − j
jΩ T
k
T k =−∞  T 
1 ∞  ω 2π 
X (e ) =  X c  j − j
jω
k
T k =−∞  T T 
∞
1
X ( e j 2π FT ) =
T
 X [ j 2π ( F − kF )]
k =−∞
c s

s i tes.google.com/site/ncpdhbkhn 6
 Ideal Periodic Sampling of
Continuous – Time Signals (3)
1 ∞  2π 
X (e jΩT
) =  X c  jΩ − j k
T k =−∞  T 
X c ( jΩ )
1
Ω H = 2π FH

−Ω H 0 ΩH Ω = 2π F

1 X ( e jΩT )
T
Ω s > 2Ω H Guard band Guard band Ω s = 2π Fs

−Ω s −Ω H 0 ΩH
Ω s − ΩH Ω s Ω = 2π F

1 X ( e jΩT )
Ω s − ΩH
T
Ω s < 2Ω H Ωs Ω s = 2π Fs
−2Ω s −Ωs 0 ΩH 2Ω s Ω = 2π F
s i tes.google.com/site/ncpdhbkhn 7
 Ideal Periodic Sampling of
Continuous – Time Signals (4)
1 ∞  2π 
X (e jΩT
) =  X c  jΩ − j k
T k =−∞  T 

1 X ( e jΩT )
T Nyquist rate
Ω s > 2Ω H
2Ω H
−Ω s −ΩH 0 ΩH π 2π Ω [rad/s]
T T
Fs
Nyquist frequency FH Fs
2
Sampling frequency
Folding
frequency

s i tes.google.com/site/ncpdhbkhn 8
 Ideal Periodic Sampling of
Continuous – Time Signals (5)
1 X ( e jΩT )
T
Ω s > 2Ω H Guard band Guard band Ω s = 2π Fs

−Ω s −ΩH 0 ΩH Ω s − ΩH Ω s Ω = 2π F

X c ( jΩ )
TX ( e jΩT ), Ω ≤ Ω s / 2 1
X c ( jΩ) =  Ω H = 2π FH
0, Ω > Ωs / 2
−Ω H 0 ΩH Ω = 2π F

Let xc(t) be a continuous – time bandlimited signal with Fourier transform

X c ( jΩ) = 0 for Ω > Ω H
Then xc(t) can be uniquely determined by its samples x[n] = xc(nT), where
n = 0, ±1, ±2, …, if the sampling frequency Ωs satisfies the condition
2π
Ωs = ≥ 2Ω H
T
s i tes.google.com/site/ncpdhbkhn 9
 Sampling of
Continuous – Time Signals
1. Ideal Periodic Sampling of Continuous –
Time Signals
2. Reconstruction of a Bandlimited Signal
from its Samples
3. The Effect of Undersampling: Aliasing
4. Discrete – Time Processing of Continuous –
Time Signals
5. Practical Sampling and Reconstruction
6. Sampling of Bandpass Signals
s i tes.google.com/site/ncpdhbkhn 10
 Reconstruction of a Bandlimited
Signal from its Samples (1) gr ( t )

x[n]gr (t − nT )
x r (t )
∞
xr (t ) =  x[n] g (t − nT )
n =−∞
r

t
( n − 2 )T ( n − 1)T nT ( n + 1)T ( n + 2)T
∞
X r ( j Ω) = 
n =−∞
x[n ]Gr ( jΩ)e − jΩnT
∞
= Gr ( jΩ )  x[ n]e− jΩnT
n = −∞
∞
→ X r ( jΩ ) = Gr ( jΩ) X (e jΩT )
X ( e jω ) = 
n =−∞
x[n]e − jωn

s i tes.google.com/site/ncpdhbkhn 11
 Reconstruction of a Bandlimited
Signal from its Samples (2)
1 X ( e jΩT )
T
Ω s > 2Ω H Guard band Guard band Ω s = 2π Fs

−Ω s −ΩH 0 ΩH Ω s − ΩH Ω s Ω = 2π F

X c ( jΩ )
TX (e jΩT ), Ω ≤ Ω s / 2 1
X c ( jΩ) =  Ω H = 2π FH
0, Ω > Ωs / 2
−Ω H 0 ΩH Ω = 2π F

X r ( jΩ) = Gr ( jΩ ) X ( e jΩT )
T , Ω ≤ Ωs / 2
If Gr ( jΩ) = GBand Limited ( jΩ) = 
0, Ω > Ωs / 2

→ X r ( jΩ ) = X c ( jΩ ) → xr (t ) = xc (t )
s i tes.google.com/site/ncpdhbkhn 12
 Reconstruction of a Bandlimited
Signal from its Samples (3)
G BL ( jΩ)
T , Ω ≤ Ωs / 2 T
Gr ( jΩ) = G BL ( jΩ) = 
0, Ω > Ωs / 2
Ω
π π
sin(π t / T ) − 0
→ gr (t ) = g BL (t ) = T T
πt / T
∞ g BL (t )
xr (t ) =  x[n]g (t − nT )
n =−∞
r
1

0.8

∞
sin[π (t − nT ) / T ] 0.6

→ xr ( t ) =  x[n]
π (t − nT ) / T
0.4

n = −∞
0.2

x[n ] Ideal digital – xr (t ) - 0.2

t
analog converter - 0.4
-5 -4 -3 -2 -1 0 1 2 3 4 5

− 5T − 4T −3T − 2T −T T 2T 3T 4T 5T
Fs = 1/T
s i tes.google.com/site/ncpdhbkhn 13
Reconstruction of a Bandlimited
Signal from its Samples (4)

x[0] x[1]g r (t − T )
x[0]gr (t )
x[1]
xc (t )

0 T 2T

∞
sin[π (t − nT ) / T ]
xr (t ) =  x[n ]
π ( t − nT ) / T
n =−∞

s i tes.google.com/site/ncpdhbkhn 14
 Reconstruction of a Bandlimited
Signal from its Samples (5)
∞
X c ( j Ω) =  x c (t )e − jΩt dt
−∞
Continuous – Time

Continuous – time
xc ( t ) X c ( jΩ)

Frequency
Fourier Transform Pairs
1 ∞
x c (t ) =
2π 
−∞
X c ( j Ω)e jΩ td Ω

Lowpass – Filtering
Reconstruction
Sampling

Aliasing

Normalized Frequency
Discrete – Time

∞
X (e jΩ T ) =  x[ n]e − jΩTn

n =−∞

Discrete – time
x[n ]
Fourier Transform Pairs
X ( e jΩT )
1 π /T
x [n] =
2π π
− /T
TX (e jΩT )e jΩ Tn d Ω

s i tes.google.com/site/ncpdhbkhn 15
 Sampling of
Continuous – Time Signals
1. Ideal Periodic Sampling of Continuous –
Time Signals
2. Reconstruction of a Bandlimited Signal from
its Samples
3. The Effect of Undersampling: Aliasing
4. Discrete – Time Processing of Continuous –
Time Signals
5. Practical Sampling and Reconstruction
6. Sampling of Bandpass Signals
s i tes.google.com/site/ncpdhbkhn 16
 The Effect of Undersampling:
Aliasing (1)
x c (t ) Ideal analog – x[n ] Ideal digital – y r (t )
digital converter analog converter
X c ( j 2π F ) Fs = 1/T X (e j 2π FT ) Fs = 1/T Yc ( j2π F )

0. 8

0. 6

0. 4

0. 2

-0 .2

-0 .4

-0 .6

-0 .8

0 1 2 3 4 5 6

s i tes.google.com/site/ncpdhbkhn 17
 The Effect of Undersampling:
Ex. 1 Aliasing (2) Spectrum of x (t)
c
xc (t) = cos(2π F0t ) 1
1
2
F0 < Fs
e j 2π F0t + e− j 2π F0t 2
F
x c ( t) = − F0
2 − Fs 0 F0 Fs

j 2π F0T T GBL ( j 2π F )
X (e )
1
∞ F Fs
1 − s
 X [ j 2π ( F − kF )]
2T
= c 0 s
2 2
T F
k =−∞
− Fs − F0 0 F0 Fs
GBL ( j2π F )
Spectrum of xr(t)
T , F ≤ Fs / 2 1
= No aliasing 2
F0 <
1
Fs
0, F > Fs / 2 2
F
− Fs − F0 0 F0 Fs
X r ( j 2π F0 ) = Gr ( j 2π F ) X (e j 2π F0T )
s i tes.google.com/site/ncpdhbkhn 18
 The Effect of Undersampling:
Ex. 1 Aliasing (3) Spectrum of x (t)
c
xc (t) = cos(2π F0t ) 1
1
2
Fs < F0 < Fs
e j 2π F0t + e− j 2π F0t 2
F
x c ( t) =
2 − Fs − F0 0 F0 Fs

j 2π F0T T GBL ( j 2π F )
X (e ) 1
2T
∞ Fs Fs
1 −
=
T
 X [ j 2π ( F
k =−∞
c 0
− kFs )] 2 2
F
− Fs −F0 0 F0 Fs
GBL ( j2π F )
Spectrum of xr(t)
T , F ≤ Fs / 2 1
= Aliasing 2
F0 <
1
Fs
0, F > Fs / 2 2
F
− Fs −( Fs − F0 ) 0 Fs − F0 Fs
X r ( j 2π F0 ) = Gr ( j 2π F ) X (e j 2π F0T ) xr (t ) = cos[ 2π ( Fs − F0 )t ] ≠ xc (t)
s i tes.google.com/site/ncpdhbkhn 19
 The Effect of Undersampling:
Ex. 1 Aliasing (4)
xc (t) = cos(2π F0t )
→ xc (t) = cos[2π ( Fs / 2 + ∆F )t ]
F0 = Fs / 2 + ∆F , ∆F ≤ Fs / 2
xr (t ) = cos[ 2π ( Fs − F0 )t ]
Fapparent = Fs − F0 = Fs / 2 − ∆F
→ xr (t ) = cos 2(π Fa t ) = cos[2π ( Fs / 2 − ∆F )t ]

s i tes.google.com/site/ncpdhbkhn 20
 The Effect of Undersampling:
Ex. 2 Aliasing (5)
−At
xc (t) = e
1

xc(t)
0.5
2A
X c ( jΩ ) = 2
A + Ω2 0
-10 -8 -6 -4 -2 0 2 4 6 8 10
A>0 t
0.4

X c(jΩ)
0.2
−A n T
x[ n] = xc (nT ) = e
0
-10 -8 -6 -4 -2 0 2 4 6 8 10
− AT
= (e ) =a ,
n n
Ω

a = e− AT
6

4
X(e )
jω

0
∞ -10 -8 -6 -4 -2 0 2 4 6 8 10
jω
X (e ) =  x[n]e
n =−∞
− j ωn
1
Ω
Sum
Xc (jΩ ) shifted 2π to left & scaled 1/T

1 − a2 X (jΩ ) scaled 1/T

yr(t)

= , 0.5 c

1 − 2a cos(ω ) + a 2 Xc (jΩ ) shifted 2π to right & scaled 1/ T

Ω 0
ω= -10 -8 -6 -4 -2 0
t
2 4 6 8 10

Fs s i tes.google.com/site/ncpdhbkhn 21
 Sampling of
Continuous – Time Signals
1. Ideal Periodic Sampling of Continuous –
Time Signals
2. Reconstruction of a Bandlimited Signal from
its Samples
3. The Effect of Undersampling: Aliasing
4. Discrete – Time Processing of Continuous –
Time Signals
5. Practical Sampling and Reconstruction
6. Sampling of Bandpass Signals
s i tes.google.com/site/ncpdhbkhn 22
 Discrete – Time Processing of
Continuous – Time Signals (1)
LTI yr (t )
xc ( t ) Ideal ADC x[n ] Discrete – Time y[ n] Ideal DAC
X c ( jΩ) Fs = 1/T X ( e jω ) System Y (e jω
) Fs = 1/T Yr ( jΩ)
h[n] H(ejω)

xc ( t ) X c ( jΩ)
1

0 t −2π FH 0 2π FH Ω
x[n ] = xc (t ) t =nT = xc (nT ) 1 ∞   2π  
X (e ) =  X c  j  Ω −
jΩT
k 
1 T k =−∞   T 
T

T 2T 3T 2π −2π FH 0 2π FH 2π Ω
0 t −
T T
s i tes.google.com/site/ncpdhbkhn 23
 Discrete – Time Processing of
Continuous – Time Signals (2)
LTI yr (t )
xc ( t ) Ideal ADC x[n ] Discrete – Time y[ n] Ideal DAC
X c ( jΩ) Fs = 1/T X ( e jω ) System Y (e jω
) Fs = 1/T Yr ( jΩ)
h[n] H(ejω)

x[n] X (e jω )
1
H (e j ω )
T
⋯ ⋯ −ωH ωH
−ωc ωc ω
0 1 2 3 n −2π 0 2π

Y (e jω ) = H (e jω ) X (e jω )
y[ n] = h[n] * x[n]
1
T
⋯ ⋯
−ωc 0 ωc ω
0 1 2 3 n −2π 2π
s i tes.google.com/site/ncpdhbkhn 24
 Discrete – Time Processing of
Continuous – Time Signals (3)
LTI yr (t )
xc ( t ) Ideal ADC x[n ] Discrete – Time y[ n] Ideal DAC
X c ( jΩ) Fs = 1/T X ( e jω ) System Y (e jω
) Fs = 1/T Yr ( jΩ)
h[n] H(ejω)

T Gr ( jΩ) Y (e jΩT )
yr (nT ) = y[ n]
1
T

0 T 2T 3T t 2π π −Ω c 0 Ω c π 2π Ω
− −
T T T T
Yr ( jΩ ) = G r ( jΩ )Y (e jΩT )
yr (t )
1

0 t −Ω c 0 Ω c Ω
s i tes.google.com/site/ncpdhbkhn 25
 Discrete – Time Processing of
Continuous – Time Signals (4)
LTI yr (t )
xc ( t ) Ideal ADC x[n ] Discrete – Time y[ n] Ideal DAC
X c ( jΩ) Fs = 1/T X ( e jω ) System Y (e jω
) Fs = 1/T Yr ( jΩ)
h[n] H(ejω)
H (e jω )
1 ∞   2π 
) =  Xc  j  Ω −
jΩT T Gr ( jΩ) Y (e jΩT )
X (e k 
T k =−∞   T  1
T
Y (e jω ) = H (e jω ) X (e jω )
2π π −Ω c 0 Ω c π 2π Ω
− −
Yr ( jΩ ) = Gr ( jΩ)Y (e jΩT ) T T T T

1 ∞
  2π 
→ Yr ( j Ω) = G BL ( jΩ) H ( e j ΩT
)
T

k =−∞
X j
c  
 
Ω −
T
k 


 H (e jΩT ) X c ( jΩ), Ω ≤π /T
=
0, Ω >π /T
s i tes.google.com/site/ncpdhbkhn 26
 Discrete – Time Processing of
Continuous – Time Signals (5)
LTI yr (t )
xc ( t ) Ideal ADC x[n ] Discrete – Time y[ n] Ideal DAC
X c ( jΩ) Fs = 1/T X ( e jω ) System Y (e jω
) Fs = 1/T Yr ( jΩ)
h[n] H(ejω)

 H (e jΩT ) X c ( jΩ ), Ω ≤π /T
Yr ( jΩ) = 
0, Ω >π /T

 H (e jΩT ), Ω ≤π /T
H effective ( jΩ) = 
0, Ω >π /T

→ Yr ( jΩ) = H effective ( jΩ) X c ( jΩ)

s i tes.google.com/site/ncpdhbkhn 27
 Discrete – Time Processing of
Ex. 1 Continuous – Time Signals (6)
dxc (t ) Y ( j Ω)
yc (t ) = → Yc ( jΩ) = jΩX ( jΩ) → H c ( j Ω) = = jΩ
dt X ( j Ω)
 j Ω, Ω ≤ ΩH
H c ( j Ω) = 
0, otherwise

Ωs =2Ω H
hc (t )  → h[ n] = hc ( nT )

1 ∞  2π  π
→ H (e )jω
=  H c  jΩ − j k , T=
ω =ΩT T k =−∞  T  ΩH

1
jω jω
→ H ( e ) = H c ( jω / T ) = 2 , ω ≤ π
T T
0, n=0
1 π  jω  jωn 
→ h[ n] = − π  T 2  e d ω =  cos(π n )
2π  nT 2 , n≠0
s i tes.google.com/site/ncpdhbkhn 28
 Discrete – Time Processing of
Ex. 2 Continuous – Time Signals (7)
Yc ( s) Ω2n
H c (s) = =
X c ( s) s2 + 2ζΩ n s + Ω 2n

If 0 < ζ < 1 → hc (t) =

Ωn
1−ζ 2  (
e −ζΩnt sin  Ω n 1 − ζ 2 t  u(t)
 )
→ h[n] = hc (nT ) =
Ωn
1−ζ 2  (
e −ζΩnnT sin  Ω n 1 − ζ 2 nT  u(n)
 )
=
Ωn
1−ζ 2
( e−ζΩnT )
n

 (
sin  ΩnT 1 − ζ 2 n  u( n)
 )
Ωn
) sin (Ω T )
∞

( e 1 − ζ 2 n z − n
− ζΩ nT n
→ H ( z) =
1−ζ 2
n= 0
n


=
Ωn
×
(
e− ζΩnT sin ΩnT 1 − ζ 2 z −1 )
1−ζ 2 ( )
1 − 2e− ζΩnT cos ΩnT 1 − ζ 2 z −1 + e−2ζΩn T z −2
s i tes.google.com/site/ncpdhbkhn 29
 Discrete – Time Processing of
Ex. 2 Continuous – Time Signals (8)
Yc ( s) Ω2n
H c (s) = =
X c ( s) s2 + 2ζΩ n s + Ω 2n

If 0 < ζ < 1 → hc ( t ) =
Ωn
1−ζ 2  (
e −ζΩnt sin  Ω n 1 − ζ 2 t  u( t )
 )
Ωn
(e ) sin (Ω T )
1 − ζ 2 n  u(n )
n
− ζΩ nT
h[ n] =
1−ζ 2
n 

H ( z) =
Ωn
×
(
e− ζΩnT sin Ω nT 1 − ζ 2 z −1 )
1−ζ 2
( )
1 − 2e− ζΩn T cos Ω nT 1 − ζ 2 z −1 + e−2ζΩn T z −2

→ y[ n] =
Ωn
1−ζ 2
(
e− ζΩn T sin Ω nT 1 − ζ 2 x[n − 1] )
( )
+ 2e −ζΩnT cos ΩnT 1 − ζ 2 y[ n − 1] − e−2ζΩnT y[ n − 2]
s i tes.google.com/site/ncpdhbkhn 30
 Sampling of
Continuous – Time Signals
1. Ideal Periodic Sampling of Continuous – Time
Signals
2. Reconstruction of a Bandlimited Signal from its
Samples
3. The Effect of Undersampling: Aliasing
4. Discrete – Time Processing of Continuous –
Time Signals
5. Practical Sampling and Reconstruction
a) Analog – to – Digital Conversion
b) Digital – to – Analog Conversion
6. Sampling of Bandpass Signals

s i tes.google.com/site/ncpdhbkhn 31
 Practical Sampling
and Reconstruction
Practical approximation of ideal A/D converter

xc ( t ) Antialiasing xa ( t ) Sample A/D xq [n ]

filter and hold converter
Ha(jΩ) Fs = 1/T Fs = 1/T

Discrete – time
system

Reconstruction Sample D/A

filter and hold converter
yr (t ) Ha(jΩ) y SH ( t ) Fs = 1/T Fs = 1/T y[ n]

Practical approximation of ideal D/A converter

s i tes.google.com/site/ncpdhbkhn 32
 Analog – to – Digital Conversion
(1)
Practical approximation of ideal A/D converter

xc ( t ) Antialiasing xa ( t ) Sample A/D xq [n ]

filter and hold converter
Ha(jΩ) Fs = 1/T Fs = 1/T

xin ( t )
Hold
xin ( t ) R xout (t )
t
Sample
C 0

xout (t )
s i tes.google.com/site/ncpdhbkhn 33
 Analog – to – Digital Conversion
(2)
Practical approximation of ideal A/D converter

xc ( t ) Antialiasing xa ( t ) Sample A/D xq[n ]

filter and hold converter
Ha(jΩ) Fs = 1/T Fs = 1/T

xin ( t )
011
010
001 t
000 0
111
110
101 ∆
100 xout (t )
s i tes.google.com/site/ncpdhbkhn 34
 Sampling of
Continuous – Time Signals
1. Ideal Periodic Sampling of Continuous – Time
Signals
2. Reconstruction of a Bandlimited Signal from its
Samples
3. The Effect of Undersampling: Aliasing
4. Discrete – Time Processing of Continuous –
Time Signals
5. Practical Sampling and Reconstruction
a) Analog – to – Digital Conversion
b) Digital – to – Analog Conversion
6. Sampling of Bandpass Signals

s i tes.google.com/site/ncpdhbkhn 35
 Digital – to – Analog Conversion
(1)
Reconstruction Sample D/A
filter and hold converter
yr (t ) Ha(jΩ) y SH ( t ) Fs = 1/T Fs = 1/T y[ n]

Practical approximation of ideal D/A converter

∞
xr (t ) =  x[n] g (t − nT )
n =−∞
r

sin(π t / T )
gr (t ) = g BL (t ) =
πt /T
∞
xSH (t) =  x [n ]g
n =−∞
q SH
(t − nT )

1, 0≤t ≤T 2 sin(Ω T / 2) − jΩT / 2

gSH (t) =  ←→ GSH ( jΩ) =
CTFT
e
 0, otherwise Ω
s i tes.google.com/site/ncpdhbkhn 36
Digital – to – Analog Conversion
(2)
1, 0≤ t ≤T 2 sin(ΩT / 2) − jΩT / 2
gSH ( t ) =  ←
CTFT
→ GSH ( jΩ) = e
0, otherwise Ω
GSH ( F ) H r ( F ) = G BL ( F )

T
2
Ideal bandlimited
1.5 interpolator GBL(jΩ)
|GSH(jΩ)|

0.5

0
-5 2π π 0 π 2π 5
− − Ω
T T T T
4
π
2
∠GSH (jΩ )

-2
−π
-4
-5 2π 0 2π 5
− Ω
T s i tes.google.com/site/ncpdhbkhn T 37
 Digital – to – Analog Conversion
(3)
1, 0≤ t ≤T 2 sin(ΩT / 2) − jΩT / 2
gSH ( t ) =  ←
CTFT
→ GSH ( jΩ) = e
0, otherwise Ω
GSH ( F ) H r ( F ) = G BL ( F )

 ΩT / 2
 eΩT / 2 , Ω <π /T
H r ( jΩ) =  sin(ΩT / 2)
0,
2
 otherwise
(jΩ )|

1
SH
|G

0
-5 0 5
Ω
2
|H (jΩ )|

1
r

0
-5 0 5
Ω
|G (jΩ )H (jΩ )|

4
r

2
SH

0
-5 0 5
Ω
s i tes.google.com/site/ncpdhbkhn 38
 Sampling of
Continuous – Time Signals
1. Ideal Periodic Sampling of Continuous – Time
Signals
2. Reconstruction of a Bandlimited Signal from its
Samples
3. The Effect of Undersampling: Aliasing
4. Discrete – Time Processing of Continuous –
Time Signals
5. Practical Sampling and Reconstruction
6. Sampling of Bandpass Signals
a) Integer Band Positioning
b) Arbitrary Band Positioning

s i tes.google.com/site/ncpdhbkhn 39
 Integer Band Positioning (1)
X c ( j 2π F ) FH = 3( FH − FL )
0, Ω ≤ Ω L = 2π FL 1
X c ( jΩ) = 
0, Ω ≥ Ω H = 2π FH N P

Ω − ΩL − FH −FL 0 FL FH F
B = FH − FL = H
2π
1 X (e j 2π FT ) Fs = 2 B
T
FH = K ( FH − FL ) = KB
N N1 N2 P N3 P1
− FH −FL 0 FL FH F
1 ∞
X (e j 2π FT
) =  X c[ j 2π ( F − kFs )]
T k =−∞
Gr ( j2π F )
T
∞
xc (t) =  x (nT ) g (t − nT )
n =−∞
c r

− FH −FL 0 FL FH F
sin(π Bt ) FH − FL
g r (t ) = cos(2π Fct ), Fc =
π Bt 2
s i tes.google.com/site/ncpdhbkhn 40
Integer Band Positioning (2)
X c ( j 2π F ) FH = 4( FH − FL )
1

N P
− FH −FL 0 FL FH F

1 X (e j 2π FT ) Fs = 2B
T
N N1 N2 N3 P N4 P1
− FH −FL 0 FL FH F

Gr ( j2π F )
T

− FH −FL 0 FL FH F

s i tes.google.com/site/ncpdhbkhn 41
 Sampling of
Continuous – Time Signals
1. Ideal Periodic Sampling of Continuous – Time
Signals
2. Reconstruction of a Bandlimited Signal from its
Samples
3. The Effect of Undersampling: Aliasing
4. Discrete – Time Processing of Continuous –
Time Signals
5. Practical Sampling and Reconstruction
6. Sampling of Bandpass Signals
a) Integer Band Positioning
b) Arbitrary Band Positioning

s i tes.google.com/site/ncpdhbkhn 42
 Arbitrary Band Positioning (1)
X c ( j 2π F )
1

N P
− FH −FL 0 FL FH F
(k − 1) Fs − FL ≤ FL 1 X (e j 2π FT ) Fs ≥ 2 B

 kFs − FH ≥ FH T (k − 1)th k th
N P N P1
2F 2 FL
→ H ≤ Fs ≤
k k −1 0 FL FH F
2 FH F 1 (k − 1) Fs − FL kFs − FH
→ Fs ≥ = 2B H ×
k B k
F 
Fs ≥ 2 B → 1 ≤ k ≤  H 
 B 

2 FH Fs 2  FH  FH F  FH
× ≤ ≤  − 1  ; min Fs = 2 B / H=2
k B B k −1 B  B  B  FH / B 
s i tes.google.com/site/ncpdhbkhn 43
 Arbitrary Band Positioning (2)
2 FH Fs 2  FH 
× ≤ ≤  − 1
k B B k − 1 B 

5.5 k =1 k =2
k =3
k =4
5

4.5
Fs/B

3.5

2.5

2
1 2 3 4 5 6 7 8 9 10
F H/B

s i tes.google.com/site/ncpdhbkhn 44
 Arbitrary Band Positioning (3)
Ex.
Given a bandpass signal with FL = 1.5kHz and FH = 2.5kHz, find the appropriate Fs?
B = FH − FL
6
= 2.5 − 1.5
= 1 kHz 5.5 k =1 k =2
k =3
FH k =4
min Fs = 2 5

 FH / B 
4.5
2.5
=2
2.5 / 1
Fs/B

2.5
=2 3.5
2
= 2.5 kHz 3

FH 2.5
= = 2.5 2.5
B 1
2
FS 2. 5 1 2 3 4 5 6 7 8 9 10
= = 2.5 F H/B
B 1
s i tes.google.com/site/ncpdhbkhn 45
 Arbitrary Band Positioning (4)
Ex.
Given a bandpass signal with FL = 1.5kHz and FH = 2.5kHz, find the appropriate Fs?
FH F
= 2.5; s = 2.5 6
B B
5.5 k =1 k =2
k =3
k=2 k =4
5

4.5
2 FH 2 FL
≤ Fs ≤
k k −1
Fs/B

3.5
2 × 2. 5 2 × 1. 5
≤ Fs ≤
2 2 −1 3

2.5

2.5 kHz ≤ Fs ≤ 3 kHz

2
1 2 3 4 5 6 7 8 9 10
F H/B

s i tes.google.com/site/ncpdhbkhn 46