Discrete-Time Signals and Systems

Convolution and Correlation

1
Discrete-Time Linear Time Invariant System

x(n) discrete-time y(n)

LTI system
{time- {time-
varying}
input signal varying}
output signal

2
Classification of Signals

 Continuous-time vs. discrete-time

 Periodic vs. aperiodic

 Deterministic vs. random

 Energy vs. power

3
Continuous-Time vs. Discrete-Time

 A continuous -time signal x(t) is defined for all values of time, t

 x(t) need not be a continuous function of time, e.g. unit step

 A discrete-time signal x(n) = x(nT) is defined only at discrete values of time

t=nT.
u(t)
1
the unit step function u(t) is an
example of a continuous-time
0
-2 -1 0 1 2 3 t signal containing a discontinuity

4
Converting from Continuous to Discrete Time
A discrete-time signal may be formed by sampling a continuous-time signal

x(nT)
x(t) x(t)
T

0
kT (k+1)T (k+2)T t

x(t) nT x(n)

sampler quantiser

5
Periodic vs. Aperiodic Signals
A continuous -time signal x(t) is periodic if and only if

x (t )  x (t  T ) for all t.

The smallest positive value of T for which this is the case is the period of the signal

A discrete-time signal x(n) is periodic if and only if

x ( n)  x ( n  N ) for all n.

Any signal that is not periodic is aperiodic.

6
Deterministic vs. Random Signals

 Deterministic signals are described as algebraic functions of time.

 Random (stochastic) signals are described in terms of their statistical properties.

7
Power vs. Energy Signals
The energy in signal x(t) If x(t) is the current (measured in amps) flowing
T
E  Lim  x 2 (t )dt in a resistor of value 1 ohm then E is the total
T 
T
energy (measured in joules) dissipated in that

The average power in signal x(t) resistor (over all time) and P is the average

T power (measured in watts).

1
P  Lim  (t )dt
2
x
T  2T
T

If average power P is finite and non-zero then E will be infinite and x(t) is described
as a power signal.

If E is finite then x(t) is described as an energy signal.

8
Impulse (Delta or Dirac) Function
For continuous-time systems

 (t )  0, t  0


  (t )dt  1


For discrete-time systems

1, n  0
d ( n)  
0, n  0


 d ( n)  1


9
Unit Step Function
For continuous-time systems
1, t  0
u (t )  
0, t  0

u ( )    (t )dt


For discrete-time systems

1, n  0
u ( n)  
0, n  0
k
u ( k )   d ( n)


10
Sinusoid Function
For continuous-time systems

x(t )  sin( t )

For discrete-time systems

x(n)  sin( n T )

Sinusoidal signals pass through (any) linear time-invariant system with no change
to their shape.

11
Complex Exponential Function
For continuous-time systems Im
Asin(ωt)
x(t )  Ae j t rotating vector ωt
Acos(ωt) Re
For discrete-time systems

A
x(n)  Ae jn T

Rotating vector (phasor) in complex plane.

Closely related to sinusoidal signals.
Projections onto real and imaginary axes of complex plane are cosine and sine respectively.

12
Euler’s Formula

1  j t j t
cos( t )  (e e )
2
 1  j t j t
sin( t )  (e e )
2j

j t
e  cos( t )  j sin( t )

13
Impulse Response and Linear Systems

 A linear time invariant system is characterised by its impulse response h(t) or h(n).

 Its output y(t) is formed by convolution of input signal x(t) and system impulse
Response h(t).

 
y (t )   x( )h(t   )d y ( n)   x ( k ) h( n  k )
k  


14
Sampling and Reconstruction

Can we recreate from discrete-time samples the continuous-time

signal from which they were taken?

15
Digital to Analogue Conversion Using a Zero-Order Hold
A zero-order hold (ZOH) has the following impulse response

1 T is the sampling period

y ( n)  y ( nT )
0 T t
y(nT) h(t) y(t) = y(nT) * h(t)

t t t
T

digital to analogue conversion using a zero-order hold

16
Discrete Time Convolution

An arbitrary input signal y(n) may be decomposed into a sum of (delayed)

weighted impulses.

Corresponding output is formed by summing (delayed) weighted impulse responses.

d(n) h(n)
d(n) LTI y(n)
system
-1 0 1 2 3 n -1 0 1 2 3 n
Delta sequence Impulse response

17
Convolution

Consider convolution from the point of view of the input signal

x( n)  x(0) d ( n)  x(1) d ( n  1)  x ( 2) d ( n  2)  

Each weighted impulse at the system input results in a weighted impulse response
at the system output.

Each input sample contributes to a number of output sample values.

18
 output signal y(n) comprises
Convolution response to x(0) sum of responses

0 1 2 3 4 5 n

x(n) response to x(1)

x(n) LTI y(n)
system 0 1 2 3 4 5 n
0 1 2 3 4 n

h(n) response to x(2)

arbitrary input
signal (sequence)
-1 0 1 2 3 n 0 1 2 3 4 5 n
impulse response response to x(3)

0 1 2 3 4 5 n

19
Convolution
A more practical and useful approach is to consider convolution from the point of
view of the output signal

y (n)  h(0) x(n)  h(1) x(n  1)  h(2)d (n  2)  

Each output sample value can be computed based on a number of input

sample values

If impulse response is finite h(n)  0, 0n N

N
y ( n)   x ( k ) h( n  k )
k 0

20
Convolution
 Convolution is a fundamental and important building block in digital signal
processing.

 Its implementation is a sum of products.

 Single cycle MAC and Harvard architecture are suited to its efficient computation.

21
Properties of Convolution
Convolution involving the delta sequence is particularly straightforward
x ( n) * d ( n)  x ( n)
x ( n) * d ( n  s )  x ( n  s )
x ( n) * Kd ( n)  Kx ( n)
Commutative property
a ( n) * b( n)  b( n) * a ( n)
Associative property
( a ( n) * b( n)) * c ( n)  a ( n) * (b( n) * c ( n))
Distributive property
( a ( n) * b( n)  a ( n) * c ( n)  a ( n) * (b( n)  c ( n))

22
Correlation
 Correlation is concerned with determining the degree of similarity between
two signals

 Computationally it bears a resemblance to convolution


Rxy ( p )   x ( m) y ( m  p )
m  
  p  

23
Correlation
 Sum of products

 At the heart of the discrete Fourier transform (DFT)


Rxy ( p )   x ( m) y ( m  p )
m  
  p  

24
Correlation vs. Convolution
 The similarities between the computations involved in convolution and correlation
are coincidental.

 Convolution describes the relationships between input signal, output signal and
impulse response in a LTI system.

 Correlation is a method of determining the degree of similarity between two signals.

25