You are on page 1of 14

Shift invariant system

A relaxed system is shift invariant if and only if x(n)y(n)


implies that x(n-k) y(n-k)
for every i/p signal x(n) and every time shift k.

causal system
A system is said to be causal if the output of the system at any time n depends only
on present and past but does not depend on future inputs. In mathematical terms,
the output of a causal system satisfies an equation of the form, y(n)=F[x(n),x(n1),x(n-2).]
Example : y(n)=x(n)-x(n-1)

Static system
A discrete time system is called static or memory less if its output at any instant n
depends at most on the input sample at the same time, but not on past or future
samples of the input.

Dynamic system
A discrete time system is called dynamic if its o/p at any instant n depends on the
present & past samples of i/p, & the dynamic system is said to have memory.

Stability
A linear time invariant system is said to be stable, if its impulse response is absolutely

sum able. (ie).,

h( n)

Recursive and non-recursive system


The recursive system is described by the difference equation of the form
N

k 1

k 0

y (n) a k y(n-k) + bk x(n-k)


The output at any instant depends upon the present and past values of input and
past values of output.

The non-recursive system is described by the difference equation of the form


M

y (n) bk x(n-k)
k 0

The output at any instant depends upon the present and past values of input alone.
Determine the range of values of the parameter a for which the LTI system with impulse
response h(n)=an u(n) is stable.

For stability,

h( n)

Hence,

a
n0

1 a

The above geometric series converges, if |a|<1.


Therefore, the system is stable if |a|<1.
Check whether the following systems for linearity, time invariance, causality and
stability.
i) y(n) e

x( n)

ii) y(n) x(n 2).


Examine the system defined by y(n) x(n 2) with respect to the following
properties:
i)Linear or Non-linear.
ii)Static or dynamic.
iii)Time invariant or time varying.
iv)Causal or Non-causal.
v)Stable or unstable.
Test the following properties:
i)Linearity of y(n) bx(n) ne x ( n ) .
ii)Time invariance property of y(n) (n 1) x 2 (n) c.
iii)Causality and stability condition of

y(n) x(n) x(n 1) x(n 2).


y(n) sin x(n).

Determine whether the following is causal, linear and time invariant.

y(n) x(n 2).


For each of the following discrete time system determine whether or not the
system is linear, time variant, causal and stable.

y(n) x(n 7)
y(n) nx(n)

y(n) x3 (n).

Check whether the system y(n) ax(n 1) x(n) is linear, casual, shift invariant
and stable.
Determine whether the following signals are Linear, Time variant, causal and
stable.

y(n) cos[ x(n)]


y(n) x(n 2)
y(n) x(2n)
y(n) x(n) nx(n 1) .

Determine whether the system is causal and stable.


n
i) h(n) 2 u(n)

n
ii) h(n) sin 2

iii) h(n) sin n (n)


2n
iv) h(n) e u(n 1)

Solution:
n
i) h(n) 2 u(n)

For stability,

h( n)

Here,

2 0.5
n

n 0

1
2
1 0.5

So the system is stable.

h(n) 0, for n 0, so the system is non causal.


n
ii) h(n) sin 2

h(0) 0, h(1) 0.707, h(2) 0, h(3) 1, h(4) 0, h(5) 1, h(6) 0, .....


h(1) 0.707, h(2) 0, h(3) 1, h(4) 0, h(5) 1, h(6) 0,

h(n) 0, for n 0, so the system is non causal.

For stability,

h(n) ;

n
So the system is unstable.
2

sin

iii) h(n) sin n (n)

h(n) 0, for n 0, so the system is causal.


For stability,

h(n) 0 0 .... 1 0 0 ... 1 So the system is stable.

2n
iv) h(n) e u(n 1)

e2 n

h( n)
0

for n 1
for n 1

h(n) 0, for n 0, so the system is causal.


For stability,

h(n) ; e

2n

So the system is unstable.

n 1

Sampling of analog signals:


The periodic sampling is described by the relation x(n) = xa(nT) ,Where
x(n) is the discrete time signal obtained by taking samples of the analog signal
xa(t) every T seconds. The time interval T between successive samples is
called the sampling period and

1
T

= Fs is called the sampling rate.

Fig. Periodic sampling of an analog signal.


Consider an analog sinusoidal signal of the form,
x a ( t ) A cos(2 Ft )

Which when sampled periodically at a rate Fs T samples/sec, results in


2 Fn
x a ( nT ) A cos(2 FnT ) A cos

Fs

or 2 f 2 FT T

Fs

Sampling process: It is the first process performed in A/D conversion.


Shanons sampling theorem:
Shannons sampling theorem states that, a signal can be exactly reconstructed from
its samples if the sampling frequency is greater than twice the highest frequency of
the signal.
Consider an arbitrary signal m(t) is a real valued band limited signal satisfying the
condition M(w)=0 for m Suppose the signal m(t) is sampled instantaneously
at a uniform rate once every Ts seconds. As a result an infinite sequence of most
spaced Ts seconds apart denoted by ms (t ) is obtained.
The Fourier transform of the impulse train is obtained as follows:

T (t )

(t nTs)

The complex Fourier series representation of ,

T (t )

ce
j

nt

where s

2
Ts

T /2

1
1
where cn
T (t )e js nt dt

Ts T / 2
Ts

(t )

1 jn t
s

e
n T
s

FT [T (t )]

FT e jns t

Ts n
1

j n t jt
j ( n )t dt
s
s
e
dt e
2 ( ns )
e

( n s )
FT [T (t )]
Ts n
2

Thus the Fourier transform of a unit impulse train is also a similar impulse train.
Now the multiplier or sampled output is,

ms (t ) m(t )T (t ) m(t )

(t nTs ) m( nTs ) (t nTs )


n
n

The Fourier transform of the multiplier output is

FT [ ms (t )] M s ( )
M s ( )

1
2

[ M ( ) * s

( n s )]
n

M ( n s )
n

Ts
1

M s ( ) represents a continuous spectrum which is periodic with a period equal to

1
. M s ( ) will repeat periodically without overlap as long as f s 2 fm and
Ts
m(t) can be recovered by low pass filtering the sampled signal M s (t ) . The
minimum sampling rate f s 2 fm samples per second is called the Nyquist rate.
On the other hand if m(t) is sampled at a rate less than the Nyquist rate that is

f s 2 fm the shifted components of M s ( ) overlaps . Because of this overlaps it


is no longer possible to recover m(t) from M s (t ) by low pass filtering . Since the
spectral components in these regions of overlaps add and the signal is distorted .This
distortion that occurs when a signal is sampled too slowly is called aliasing.
Data reconstruction process:
The spectrum of the sampled signal is,

M s ( )

1
TS

M ( n s )
n

When n=0, M ( ) Ts M s ( )

Which is band limited (ie) M ( ) 0 for | | m


Where Ts

ms (t )

(sampling rate)

m( nTs ) (t nTs )
n

m( nTs ) sampled signal


Taking Fourier transform on both sides of equation (2),

M s ( )

jnTs

m nTs e
n

Substituting (3) in (1),

M ( )

j nTs
m nTs e
n

For | | m , equation (4) shows that the Fourier transform M ( ) of the band
limited signal m(t) is uniformly determined by its sample values m( nTs ) .
Taking inverse Fourier transform of equation (4),

1 m
j (t nTs )
jt
m( nTs )e
d
d
M ( ) e

2
2 m m n
m j (t nT )

1
s d
m( nTs ) e
m (t )
n

m
2m

m (t )

e j (t nTs ) m

m ( nTs )
m (t )
j (t nTs )
2m n
m
1

e jm (t nTs ) e jm (t nTs )

m(t ) m( nTs )
2 jm (t nTs )
n

m (t )

Sin m (t nTs )
m( nTs )
m (t nTs )
n

m(t )

Sin (2 f mt n )

m(nTs )
m(nTs ) Sinc 2 f m t n
(2 fmt n)
n
n

(5)

Equation (5) provides an interpolation formula for reconstructing the original signal
m(t) from the sequence of sample values m(nTs ) with the sinc function

Sinc 2 f m t

playing the role of an interpolation function. Each sample is multiplied by a delayed


version of the interpolation function and all the resulting waveforms are added to
obtain m(t).
Equation (5) also represents the response of an ideal lowpass filter of bandwidth fm
and zero transmission delay, which is produced by an input signal consisting of the
sequence of samples m(nTs ) by passing it through an ideal low pass filter of
bandwidth fm.

ALIASING:
When the sampling rate exceeds the Nyquist rate 2fm all replicas of M(w) involved
in the reconstruction of Ms(w) move farther apart, and there is no problem in
recovering the original signal m(t) from its sampled version ms(t).
When f s 2 fm , in the construction of the spectrum of sampled signals Ms(w), the
frequency shifted replicas of the original spectrum M(w) overlap. In this case the
high frequencies in M(w) get reflected into the low frequencies in Ms(w).
If the sampling rate f s 2 fm the original signal m(t) cannot be recovered exactly
from its sampled version ms(t) and information is there by lost in the sampling
process. The distortion that occurs when a signal is sampled too slowly is called
aliasing.
Another factor that contributes to aliasing is the fact that a signal cannot be finite in
both time and frequency. Therefore this violates the strict band limited requirement
of the sampling theorem. So whenever a time limited signal is sampled there will be
some aliasing produced by the sampling process. Accordingly in order to combat the
effects of aliasing in practice, two corrective measures are used.
i)prior to sampling, a low pass pre aliasing filter is used to attenuate through high
frequency components of the signal, that lie outside the band of interest.
ii) The filtered signal is sampled at a rate slightly higher than the Nyquist rate.
Analog to digital conversion is done in three steps. This process is illustrated in Fig.

(i)

Sampling :This is the conversion of a continuous time signal into a discrete


time signal obtained by talking samples of the continuous time signal at
discrete time instants . Thus, if xa(t) is the input to the samples, the output
Xa(nt) = x(n), where T is called the sampling internal .

(ii)

Quantization : This is the conversion of a discrete time continuous valued


signal into a discrete time discrete valued (digital signal)signal. The value of
each signal sample is represented by a finite set of possible values. The

difference between the unquantized sample x(n)


xq(n)is called the quantization error.
(iii)

and the quantized output

Coding :In the coding process, each discrete value xq(n) is represented by a b
bit binary sequence.

Quantization of continues amplitude signals:


The process of converting a discrete time continuous amplitude signal into a digital
signal by expressing each sample value as a finite number of digits is called
quantization. The error introduced in representing the continuous valued signal by
a finite set of discrete value levels is called quantization error.
Let xq(n) denote the sequence of quantized samples at the output of the quantizer.
Hence xq(n) = Q[x(n)]. Then the quantization error is a sequence eq(n) defined as
the difference between the quantized value and the actual sample value.
Thus eq(n) = xq(n)-x(n).
Quantization is done two ways.
(i)

Truncation (excess digits are discarded)

(ii)

Rounding (rounded off to next integer level). The distance between two
successive quantization levels is called the quantization step size (or)
resolution. The quantization error eq(n) in rounding is limited to the range of

to

. In other words, the instantaneous quantization error cannot exceed

half of the quantization step. If x min & x max represent the minimum and
maximum values of x(n) & the number of quantization levels, then

x max x min
L1

The quantization error decreases and the accuracy quantizer increases as the
number of quantization level increases.

The quantization error eq(t) = xa(t) xq(t)


Let denotes the time that xa(t) stays within the quantization level. The
mean-square error power Pq is,
1
Pq
2

e q (t ) dt e2q (t ) dt
2

eq ( t )

Pq

t
2

2
2
t
dt

2
12
0
1

If the quantizer has b bits of accuracy and the quantizer covers the entire
range 2A, the quantization step is

2A
4 A2
A2
. Hence, Pq 22 b 12 22 b 3
2b

The average power of the signal xa(t) is Px

A2
2

. The signal to quantization

noise ratio (SQNR) is,

Px
A2 / 2
SQNR
2 2b
3 / 2 22 b 3 / 2 L
Pq
A / 2 3

SQNR ( dB ) 10log 1.5 22 b 1.76 6b

This implies that, the SQNR increases approximately 6dB for every bit added
to the word length.
Coding of quantized samples:
The coding process in an ADC assigns a unique binary number to each
quantization level. If there are L levels, we need at least L different binary
numbers. With a word length of b bits, the number of bits required in the
coder is the smallest integer greater than or equal to log 2L

Fig. sampling and quantization of sinusoidal signal

A Digital communication link carries binary coded words representing samples of


an input signal. The link is operated at 10,000 bits/s and each input sample is
quantized into different voltage levels.
i) what is the sampling frequency and the folding frequency?
ii) what is the nyquist rate for the signal x(t)?
iii) what are the frequencies in the resulting discrete time signal x(n)?
iv) what is the resolution ?

You might also like