Digital Signal Processing Lecture Notes 2 Page |1

EXAMPLES:

a) Convert the following signals from Analogue to Digital

Consider the formula: y (t )  A sin(2ft) and that the sampling frequency is 40Hz
k (t )  2.5 cos( 400t )

b) Find the Power of: f [n]  cos(n )

c) If x[n]  cos(2Fn) Find the digital Period N if F=0.32
j 0.2 n
d) If x[n]  e  e  j 0.3n Find the digital frequencies of the signal

4. Signal Symmetry

1 Types of Symmetry Odd and Even described by:

x e [ n]  x e [  n] (Even Symmetry)
x o [n]   x o [ n] (Odd Symmetry)

 Symmetry Limits:  N  n  N (equal number of samples on either side of the origin)

 If a signal x[n] is identical to its mirror image x[n] , it is an even symmetric signal. An even
symmetric signal shows mirror symmetry about the origin n=0.

 The value of an odd symmetric signal is zero at n=0.

 The sum of an odd symmetric signal and its time reversed version is zero everywhere.


 Rules for ODD symmetry xo [0]  0 and  x [n]  0
n  
o

 To find the ODD and EVEN components of a signal the following formulae are used:

EVEN SYMMETRY : x e [ n]  0.5( x[ n]  x[ n])

ODD SYMMETRY: x o [ n]  0.5( x[ n]  x[n])


EXAMPLE: Find the ODD and EVEN Symmetry of the following signal: x[n]  {4,  2, 4,  6}

The sequence does not obey the rule of symmetry limits i.e.  N  n  N , it does not have equal
number of sample elements of either side of the origin.

This is solved by ZERO PADDING the sequence which becomes:


x[n]  {4,  2, 4,  6, 0} , by adding a zero at the beginning or the end, it does not change the value
of the sequence. The sequence obeys the limits of symmetry.

Applying the formula for even symmetry i.e. x e [ n]  0.5( x[ n]  x[ n]) 0.5x[n]  {2,  1, 2,  3, 0}

0.5x[n]  {0,  3, 2,  1, 2}

Then xe [n]  0.5( x[n]  x[n])  {2,  4, 4,  4, 2}

and xe [n]  0.5( x[n]  x[n])  {2, 2, 0,  2,  2} ,
Digital Signal Processing Lecture Notes 2 Page |2

Sketch x e [n] and x o [n]


NOTE: xo [0]  0 and  x [n]  0
n  
o

Example: If x[ n]  e Find the ODD and EVEN Symmetric Components of the Signal.
jxn

Applying the formulae:

xe [n]  0.5( x[n]  x[n])
xo [n]  0.5( x[n]  x[n])

x[n]  xe [n]  xo [n]

e jxn  e  jxn
xe [n]  0.5[e jxn  e  jxn ]   cos xn
2
e jxn  e  jxn
xo [n]  0.5[e jxn  e  jxn ]   j sin xn
2
Example: Find the ODD and EVEN component signals of:

x[n]  8(0.5) n (u[n]  u[n  3])

5. The Sampling Theorem

The sampling process can be mathematically represented where an analogue signal (a) is multiplied
by an impulse train (b) as shown below.

This can be mathematically represented by


s(t )   (t  nTs)
n  

The sampled signal is then represented by (c)

By applying convolution
Digital Signal Processing Lecture Notes 2 Page |3


y(t )  x(t ) * s(t )  x(t ) *  (t  nTs)
n  

Convolution results in: s(t )   x(nTs) (t  nTs)
n 

The location of the samples is determined by the impulse train and its weight is determined by the value
of the analogue signal at the specific instance. Convolution is multiplication in the frequency domain.

Sampling Theorem states that for a unique correspondence between and analogue signal and the
reconstructed version from its samples the rate must exceed the twice the highest frequency f m ax .

i.e. S  2 f m ax is called critical sampling rate or Nyquist Frequency.

Example:

1. For the analogue signal f (t ) find the sampling rate of the single so that it can be digitally
reconstructed.

f (t )  sin(15t ) cos(10t )
1 1
From sin  cos   (sin(   )  sin(   )) , f (t ) becomes (sin 25t  sin 5t )
2 2
Sampling rate S  2 f m ax  2  12 .5  25

2. For the analogue signal g (t ) find the sampling rate of the single so that it can be digitally
reconstructed.

g (t )  cos2 (10t )

Sampling an analogue signal without the loss of signal information is as follows:

For an analogue signal band - limited to f m ax Hz, the sampling rate S  2 f m ax must exceed 2 f m ax .

S  2 f m ax defines the Nyquist Rate.

1
ts  defines the Nyquist Interval.
2 f max

Signal Reconstruction and Aliasing

Shannon Sampling theorem states that any sinusoid can be uniquely reconstructed provided it is
sampled at least twice in one period (similar to Nyquist Frequency Theorem)

Examples of Aliasing:

Sampling causes image frequencies of the original signal. If not sampled at twice the highest frequency,
then the image frequencies will overlap and distort the original signal (as shown). Low pass filters are
used to filter out the image frequencies.
Digital Signal Processing Lecture Notes 2 Page |4

Figure: Showing Aliasing and Aliasing

Consider the signal x(t )  cos(2ft   ) and its sampled (digital) version x[n]  cos(2nF   )
f
where F …… digital frequency.
fs
If x[n] is to be a unique representation of x (t ) , we must be able to reconstruct x (t ) from x[n]

Aliasing and Sampled Sinusoids

Example:
A 100Hz sinusoid is sampled at 240Hz, 140Hz and 90Hz. In each case has aliasing occurred?

Assume the sinusoid is:

At 240Hz aliasing will not occur because the signal is sampled at a rate higher than the highest
frequency of 100Hz.

x(t )  cos(200t   )  cos(2 100t   )

Sampling at 140 Hz, f a  100 Hz  140 Hz  40 Hz (substitute above)

x a (t )  cos(80t   )  cos(80t   )  cos(2 40 t   )

Final Signal is changed and phase shifted.

Aliasing occurs if an Analogue signal x[ n]  cos(2f o t   ) is sampled below the Nyquist Rate. If
S  2 f o , the reconstructed analogue signal is aliased to a lower frequency | f a | 0.55 where
f a  f o  MS , M is an integer that places f a in the central period (0.5S  f a  0.5S )
Example:

1
sampling int erval   .......
fs
Digital Signal Processing Lecture Notes 2 Page |5

1. What is the frequency of the signal y[n] for the period y[0] to y[16] ?
2. If the entire signal y[n] appears for 1 second what is the sampling frequency?
3. What is the sampling interval of y[n] ?
Example:

What is the critical sampling rate in Hz from the following signals?

1. x(t )  cos(10t )
2. y (t )  cos(10t )  sin(15t )

6. Periodic and NON – Periodic Sequences

A discrete –time signal x[n] is said to be periodic if it satisfies the condition x[n]  x[n  N ] for all
integers of n.

The smallest value of N which satisfies the above condition is known as the fundamental period.

If the above condition is not satisfied even for one value of n, then the discrete – time signal is aperiodic.
2 2
The angular frequency is given by:  and the Fundamental Period is N 
N 
Example:
o
Show that the complex exponential sequence x[ n]  e jo n is periodic only if is a rational number.
2
j o n
Given: x[ n ]  e , x[n] will be periodic if

x[n]  x[n  N ]  x[n] i.e

e j[o ( n  N ]  e jo n
e jo n .e jo N  e jo n
e j o N  1

This is true only if k is an integer when  o N  2k

o k o
 is a rational number, then the sequence x[ n]  e jo n is periodic if is a rational number.
2 N 2
Example:

Is x[n]  cos(2Fn) periodic if ;

a) F=0.32,
b) F= 3