You are on page 1of 21

CH 7.

ELEMENTARY DISCRETE
SIGNALS AND BASIC
OPERATIONS

Universitatea “Politehnica” din Bucureşti Prof. Cristian Negrescu


7.1 Introduction
 A discrete signal f  Sd is a function f  n    defined for n 

 n is the discrete time variable n  ..., 3, 2, 1,0,1,2,3,...


 The notation f  n is also used

 Generators for the discrete signals


 By the sampling of the analog signals f  n   fa t 
t  nTs
 Direct generation (using recurrent equations, modelling a physical dicrete
process, etc.) 0 n0

f  n   1 n0
2 f n  1
   n0

Universitatea “Politehnica” din Bucureşti Prof. Cristian Negrescu


7.2 The sampling theorem
 Formulation: Let be f  t  a M band-limited signal
f  t  
F
 F   , with F    0,   M f t 


F F  

f 1 f0
f 2 f2
f1
sampled with the period Ts , where
2
s   2M 2Tt Tt Ts 2Ts t  M M 

Ts

 s   period. s  2M

Then f t   
n 
f n sinc   t  nTs  
2 
G  

M  s

Proof: We build the periodic spectrum of the signal f  t 


  2  s  M M s 

G     F   n 
FS jn

n 
s   e
n
n
s

2
Def
1 M  jn 
2
where  n   F   e s
d  f  nTs 
s  M s
SF  n  n 
G    T
n 
s fn e jnTs
 T
n 
s f n e jnTs
7.2 The sampling theorem
 Formulation: Let be f  t  a M band-limited signal
f  t  
F
 F   , with F    0,   M f t 


F F  

f 1 f0
f 2 f2
f1
sampled with the period Ts , where
2
s   2M 2Tt Tt Ts 2Ts t  M M 

Ts

 s   period. s  2M

Then f t   
n 
f n sinc   t  nTs  
2 
G  
p1  
M  s

Proof: We build the periodic spectrum of the signal f  t 


  s  M M s 

G    T s fn e  jnTs

filtering F    G   p1  
1

n  
M  1  s  M f t 
F 1 F  


How can we recover the original signal ? LPF with

F    T
n 
s fn e  jnTs
p1  
sin 1t F
 p1   
  t  M M
t
1 sin 1t F f  t    Ts 1 f nsinc 1  t  nTs  
 1t
 p1  
n  
1
 p1   
sinc1t 
1  s
F


Universitatea “Politehnica” din Bucureşti 2 Prof. Cristian Negrescu
7.2 The sampling theorem
 Formulation: Let be f  t  a M band-limited signal
f  t  
F
 F   , with F    0,   M f t 


F F  

f 1 f0
f 2 f2
f1
sampled with the period Ts , where
2
s   2M 2Tt Tt Ts 2Ts t  M M 

Ts

 s  s  period. s  2M

Then f t   
n 
f n sinc   t  nTs  
2 
1 
2
g t   T f  t  nT 
n
s n s
F 1
 G  
p1  
M  s

Proof: We build the periodic spectrum of the signal f  t  2Tt Tt


 Ts 2Ts t  s  M M s 

G    T
n 
s fn e  jnTs

filtering f  t   g  t  *

1
sinc 1t   filtering F    G   p1  
1


M  1  s  M f t 
F 1 F  


How can we recover the original signal ? LPF with

An alternate variant g  t   T
n 
s f n  t  nTs 

  t  M M

f  t    Ts 1 f nsinc 1  t  nTs  


n  
1
sinc1t 
F
 p  
 1

Universitatea “Politehnica” din Bucureşti Prof. Cristian Negrescu


7.2 The sampling theorem

Remarks about sampling frequency

 Formulation: Let be f  t  a M band-limited signal


f  t  
F
 F   , with F    0,   M f t 


F F  

f f11 f0
f 2 f 2 f2 f2
f1 f1
sampled with the period Ts , where
2
s   2M 2Tt 2T
t tTt
T TsTs 2Ts 2Ts t  M M 

Ts

 s  s  period. s 
 2M

Then f t   
n 
f n sinc   t  nTs  
2 
1 
2
g t  

f 1
f tTs f n  t  nTs 
n

f0
F 1
 G  
p1  
f2
f 2 f1 M  s 1

Remark: At the limit, to avoid the spectral aliasing s  2M 2Tt 2T tTt TsTs 2Ts 2Ts t
t T  s s  M M s s 
1

1
filtering f  t   g  t  * sinc 1t   filtering F    G   p1  

f t  F  
F 1

t  M M 

Universitatea “Politehnica” din Bucureşti Prof. Cristian Negrescu


7.2 The sampling theorem

Remarks about sampling frequency

 Formulation: Let be f  t  a M band-limited signal


f  t  
F
 F   , with F    0,   M f t 


F F  

f 1 f 00 f
f 2 f 1 f1 f1 2 f2
sampled with the period Ts , where f 2

2
s   2M 2Ts 2Tt Ts Tt Ts Ts 2Ts 2tTs t  M M 

Ts

 s  s  period. s 
 2M

Then f t   
n 
f n sinc   t  nTs  
2 
1 
2
f 1
f t 

f0
F 1
 G  
G
p1  
f
f 2 f 1 f1 f1 2 f2
f 2

Remark: What if s  2M !?!


2Ts 2Tt Ts Tt Ts Ts 2Ts 2tTs t
s s MM MMss 


1    F    G   p1  
f t   f t   g t  *
filtrare sinc 1t   Ffiltrare
1
1

f t  f t  F  
F 1

t 1 1 

Aliere spectrală !
(“reflection” of the spectrum)
Universitatea “Politehnica” din Bucureşti Prof. Cristian Negrescu
7.3 Elementary Discrete Signals and Operations
 Examples (elementary signals)  Examples (elementary ops.)
 Addition (sum)
f  n    n f  n     n  n0 
 Multiplication by a constant
f  n    n f  n     n     n   Delay
f  n     n  1 f  n   sgn  n   Mirror
 Convolution
f  n  a n
f  n  2   n
n
 Correlation
f  n   2n   n  1

Universitatea “Politehnica” din Bucureşti Prof. Cristian Negrescu


7.3 Elementary Discrete
Signals and Operations

Discrete Signals
 Examples (elementary signals)

  n   n
1 1
… …
n n
… -3 -2 -1 0 1 2 3 … -3 -2 -1 0 1 2 3

2n   n  4
  n  1
2
1 1
… … … …
n n
-3 -2 -1 0 1 2 -3 -2 -1 0 1 2 3

Universitatea “Politehnica” din Bucureşti Prof. Cristian Negrescu


7.4 Convolution and Correlation
for Discrete Signals
 Linear convolution  Properties (linearity, commutativity, conditions
of existence, support, etc…)
Def 
 x * y  n    x  k  y  n  k   Computing (def., graphic. Z transf, matrix)
k 
 Other types of convolution (periodic, circular)
Def 
 x *   n    x  k   n  k    * x  n   x  n 
k 

 Linear correlation (for real signals)


 
Def 
x *ˆ y  k    x n y n  k 
n 
 Properties (linearity, commutativity, conditions
of existence, support, etc…)
 Computing (def., graphic. Z transf, matrix)
 Connection with the convolution

Universitatea “Politehnica” din Bucureşti Prof. Cristian Negrescu


7.4 Convolution and
correlation
Linear convolution for signals with finite
support
 Convolution matrix
y3
x  n x1 y  n
x0 xN1 1 y0 y N 2 1

2 3 … 1 2 …
0 1 N1  1 n 0 3 N2  1 n
x2 y2
x3 y1
T T
x   x0 x1 x2 x3 ... xN1 1  y   y0 y1 y2 y3 ... y N2 1 
supp x  n   0, N1  1 supp y  n   0, N2 1
card  x   N1 card  y   N2

Universitatea “Politehnica” din Bucureşti Prof. Cristian Negrescu


7.4 Convolution and
correlation
Linear convolution for signals with finite
support
 Convolution matrix x
T
x   x0 x1 x2 x3 ... xN1 1 
x0 ... 0 ... 0  y
supp x  n   0, N1  1 ... ... ... ... 0 
card  x   N1    y0 
xN1 1 ... 0 ... 0  
  y

T
y   y0 y1 y2 y3 ... y N2 1 
1
... x0 ... ...
0   y2 
supp y  n   0, N2 1 h  0 ... ... ... 
0 
card  y   N2  ... xN1 1 ... 0   y3 
0   ... 
... ... 0 ... x0  
h  n    x * y  n  
0 ... ... ... ...
  yN2 1

 0 ... 0 ... xN1 1 

Universitatea “Politehnica” din Bucureşti Prof. Cristian Negrescu


7.4 Convolution and
correlation
Linear convolution for signals with finite
support
 Convolution matrix
T T
x   x0 x1 x2 x3 ... xN1 1  y   y0 y1 y2 y3 ... y N2 1 
 x0 ... 0 ... 0 
 ... ... ... ... 0 
  y 
 xN1 1 ... 0 ... 0  0 
   y1 
 0 ... x0 ... ... 
 y2 
h   0 N3 ...N2 ... ... 0   
   y3 
 0 ... xN1 1 ... 0 
 ... 
N3  N1  N2  1  ... ... 0 ... x0   
  N2 1 
y
 0 ... ... ... ... 
 
 0 ... 0 ... xN1 1 

Universitatea “Politehnica” din Bucureşti N2 Prof. Cristian Negrescu


7.4 Convolution and
correlation
Linear convolution for signals with finite
support
 x0 ... 0 ... 0 
 Convolution matrix  ...

... ... ... 0 
 y 
T  xN1 1 ... 0 ... 0  0 
x   x0 x1 x2 x3 ... xN1 1  
 0 ... x0 ...
 y
...   1 
 y 
T h 0 ... ... ... 0  2 
y   y0 y1 y2 y3 ... y N2 1  
 0 ... xN1 1 ...
 y
0  3 
 ... 
 ... ... 0 ... x0   
  yN2 1 

 x N 1 0 0   0 ... ... ... ... 
  1 
 
  0 ... 0 ... xN1 1 
h N31   x N11   y N2 1  
  



 0 0 x N11 
  x  
  
 x  y
h 
 N3 1  N3  N 2   N2 1 

x
x





   
 x 
N3  N1  N2  1  
 x
Universitatea “Politehnica” din Bucureşti Prof. Cristian Negrescu
7.4 Convolution and
correlation
Linear convolution for signals with finite
support
 
 
 Convolution matrix 
x 
  
T   
x   x0 x1 x2 x3 ... xN1 1   x  y
h x  
T   
y   y0 y1 y2 y3 ... y N2 1  

x 
 


 x 
 x N 1 0 0   
  1    x
h N31   x N11   y N2 1
 
 0 0 x N11 
 h  n h1
h0 h2 hN3 1
T
h   h0 h1 h2 h3 ... hN3 1 
3 …
supp h  n   0, N3  1 0 1 2 N3  1 n

card  h   N3  N3  N1  N2 1
h3
Universitatea “Politehnica” din Bucureşti Prof. Cristian Negrescu
7.4 Convolution and
correlation
Linear convolution for signals with finite
support
 Correlation matrix
y3
x  n x1
y  n
x0 xN1 1 y0 y N 2 1

2 3 … N1 N1  1 …
0 1 N1  1 n 0 N1  1 N1  2 N3  1 n
x2 y2
x3 y1
T T
x   x0 x1 x2 x3 ... xN1 1  y   y0 y1 y2 y3 ... y N2 1 
supp x  n   0, N1  1 supp y  n    N1  1, N3  1
card  x   N1 card  y   N2

Universitatea “Politehnica” din Bucureşti Prof. Cristian Negrescu


7.4 Convolution and
correlation
Linear convolution for signals with finite
support
 Correlation matrix x
T
x   x0 x1 x2 x3 ... xN1 1 
xN1 1 ... 0 ... 0  y
supp x  n   0, N1  1  ... ... ... 0 
card  x   N1 ...   y0 
x0 ... 0 ... 0  
  y

T
y   y0 y1 y2 y3 ... y N2 1 
1
... xN1 1 ... ...
0   y2 
supp y  n    N1  1, N3  1 h  0 ... ... ... 0  
card  y   N2  ... x0 ... 0   y3 
0   ... 
... 
 
... 0 ... xN1 1 
h  n   x *ˆ y  n  
0 ... ... ... ...
  yN2 1

 0 ... 0 ... x0 

Universitatea “Politehnica” din Bucureşti Prof. Cristian Negrescu


7.4 Convolution and
correlation
Linear convolution for signals with finite
support
 Correlation matrix
T T
x   x0 x1 x2 x3 ... xN1 1  y   y0 y1 y2 y3 ... y N2 1 
 xN1 1 ... 0 ... 0 
 
 ... ... ... ... 0 
 y0 
 x0 0  
  y1 
... 0 ...

 0 ... xN1 1 ... ... 
 y2 
h   0 N3 ...N2 ... 0  
  y3 
...

 0 ... x0 ... 0 
 ... 
N3  N1  N2  1  ... ... 0 ... 
xN1 1   
  N2 1 
y
 0 ... ... ... ... 
 
 0 ... 0 ... x0 
Universitatea “Politehnica” din Bucureşti N2 Prof. Cristian Negrescu
7.4 Convolution and
correlation
Linear convolution for signals with finite
support
 xN1 1 ... 0 ... 0 
 Correlation matrix 
 ... ... ... ... 0 

 x0  y 
T ... 0 ... 0  0 
x   x0 x1 x2 x3 ... xN1 1  
 0 ... xN1 1 ...
 y
...   1 
 y2 
h 0 0 
  y3 
T ... ... ...
y   y0 y1 y2 y3 ... y N2 1  
 0 ... x0 ... 0 
 ...   ... 
 ... 0 ... xN1 1   
 yN2 1 

 x N 1 0 0   0 ... ... ... ... 
  1 
 
  0 ... 0 ... x0 
h N31   x N11   y N2 1  
  



 0 0 x N11 
  x  
  
 x  y
 N3 1  N3  N 2   N2 1 h

x 



 x  
N3  N1  N2  1 
x   
 
 
 x
Universitatea “Politehnica” din Bucureşti Prof. Cristian Negrescu
7.4 Convolution and
correlation
Linear convolution for signals with finite
support
 
 
 Correlation matrix 
x 
  
T   
x   x0 x1 x2 x3 ... xN1 1   x  y
h x  
T   
y   y0 y1 y2 y3 ... y N2 1  

x 
 


 x 
 x N 1 0 0   
x
  1 


h N31   x N11   y N2 1
 
 0 0 x N11 
 h  n h1
T h0 h2 hN3 1
h   h0 h1 h2 h3 ... hN3 1 
supp h  n   0, N3  1 3 …
0 1 2 N3  1 n
card  h   N3  N3  N1  N2 1
h3
Universitatea “Politehnica” din Bucureşti Prof. Cristian Negrescu
7.4 Convolution and
correlation
Convolution Correlation

 
Def  Def 
 x * y  n    x  k  y  n  k  x *ˆ y  k    x n y n  k 
k  n 

 Convolution matrix  Correlation matrix


   
   
   
 x    x  
     
 x  y  x  y
h x   h x  
     
 x    x  
       
 x   x 
   
 x  x

 Theorem: xn0 * yn1   x * y n n


0 1
 Theorem: xn *ˆ yn  x *ˆ y
0 1
   n0  n1

Universitatea “Politehnica” din Bucureşti Prof. Cristian Negrescu

You might also like