Introduction To Discrete Time Signals & System

Download as pptx, pdf, or txt
Download as pptx, pdf, or txt
You are on page 1of 119
At a glance
Powered by AI
Some key takeaways from the document are that discrete time signals can be represented and manipulated in various ways, discrete time systems can be classified into different types like linear/nonlinear and time-invariant/time-variant, and correlation and convolution are important operations used in signal processing.

Some classifications of discrete time systems include linear/nonlinear, time-invariant/time-variant, causal/non-causal, stable/unstable, with examples provided on pages 1-2.

To calculate the correlation rxy(l) of signals x(n) and y(n), you take the convolution of x(n) and the time-reversed y(-n). This is demonstrated on pages 3-4.

hapter 1

troduction to Discrete Time Signals & System:


DiscreteTime Signals representation and Manipulation,
DiscreteTime IIR and FIR Systems, Impulse Response,
(Infinite Impulse Response and Finite Impulse Response)
Transfer Function,
Difference Equation,
Frequency Domain and Time Domain Analysis of IIR filter and FIR filter,
Correlation,
Linear and Circular and Convolution Algorithm,

apter 01 Questions
Dec. 2010
Give any five classifications of Discrete time systems with examples

X (t) = sin(480 t) + 3 sin (720 t) is sampled with Fs = 600 times per sec.

(1) What are the frequencies in radians in the resulting DT signal x (n)?

(2) If x (n) is passed through an ideal interpolator,


what is the reconstructed signal?

June 2011
If x (n) = { 2, -1, 3, 0, 4 } obtain following:
(i) x (-n)
(ii) x (n-1) (iii) x (n+1) (iv) x (-n+2)(v) x (2n)

For a discrete time system whose impulse response h (n) = { 1, -2, 1}

Find the output for input x (n) = { 1, 2, 3, 4 }

Classify following DT System on linearity/ causality and time variance:(i) y (n) = 2x(n) + x(n-1)
(ii) y (n) = x(2n) +2

1. x (n)= [1, 2, 1, 2, 3, 4]

find y (n) = x(1-n) + x(3-n) + 4 x ( 4 n)

2. Explain whether the following signals are power signal or energy signal.
a.

0.5 n u (n)

b.

A cos (n)

3. Determine whether the following signals are periodic or non-periodic.


a.

j( /4) n
x (n) = e

b.

x (n) = cos(50 t) + cos(100 t)

4. Decompose the following signals into even and odd parts


a.
x [n] = [ 1 2 3 4 5 6 7 8 ]
b.

y [n] = [8 4 8 4 8 4 8 4]

Signal
A signal is defined as any physical quantity that varies with:
1. Time,
2. Space or
3. Any other independent variable or variables
- Examples of Signals
- Speech
- Music
- Pictures
- Video
- ECG
- Mathematically, we describe a signal as a
function of one or more independent variables
s1(t ) = 5 t
s2( t) = 20 t2

One independent variable t (time)

f (x, y) = 3x + 2x y + 10 y2
two independent variables x, and y

Signals can be generated by a system:


Speech - Vocal Cords
ECG
- Electrocardiogram: Polarization/ depolarization of ventricles
EEG
- Electroencephalogram

Example of speech signal

- Speech, ECG and EEG signals are function of a single variable:

t, Time
- An image signal is function of two variables:
x, y (Coordinates of an image)
- These signals are generated by some means:
- Speech signal by vocal cord and air flow.
- Image signals by exposing light- sensitive sensors to light
- The signals are generated by a system

Signal processing system


A signal processing system is defined as a device
that perform an operation on the signal
- Amplifier is a system that amplifies a signal
- Filter is a system that suppresses or allows
some frequencies from the signal
- Thus a system is an interconnection of components that
performs an operation on
input signal and produces an output signal.
- The definition of a system can be broadened to include not only
physical devices but also,
software realization of operations

Digital signal processing system


- There are two types of signals:
- Analog signals and
- Digital signals
Analog signals:
- Most of the signals that occur in nature are analog signals.
- The analog signals are a function of a continuous variable
such as time or space,
takes on values in continuous range
- Analog signals can be processed directly, in continuous form
by analog systems (circuits) such as:
Amplifiers, Filters, Frequency multipliers,
- Analog processing:
Direct processing of signals in continuous form (Analog form)

Example of analog processing of signals

Voice amplifier:
- Amplifier is made up of resistors, capacitors, transistors etc.
- Amplifier takes the analog signal (continuous) from microphone,
amplifies it , and produces analog output signal for speaker
Analog processing system:
- Takes analog input, process it in analog form and
produces analog output

Digital Signal Processing:


- Digital signal processing provides an alternative method for
processing the analog signals
- First analog signals are convert to digital signals :
ADC (Analog to Digital converter)
ADC provides the interface between
the analog signal and the digital processor
- Process the digital signals, using digital processor
- Digital processor can be
- Programmable computer,
- Small microcomputer, or
- DSP chip.
- The digital output of the DSP is converted back to analog
( DAC, Digital to Analog converter)

Example of digital processing of signals

- First, input analog signal from microphone is converted to digital signal


by Analog to Digital Converter (ADC)
- Then, the digital signal is processed by Digital Signal Processor (DSP)
- Finally, the digital output produced by DSP, is converted to Analog signal,
by Digital to Analog converter (DAC), for feeding it to speaker

x (t)

Analog
Input signal
t is time

x (n)

Digital
Input signal
n is number
(sample)

y (n)

Digital
Output signal

y (t)

Analog
Output signal

ntages of Digital Signal Processing over Analog Signal processing:

DSP is more flexible: Easily reconfigurable. (Software)


(Reconfiguration of analog system requires redesigning
hardware)
Replication is easy, digital systems can be easily replicated

Accuracy of design in digital systems is high


( Getting high accuracy in analog systems is very difficult
because of tolerances of hardware components)

Storage of digital signals is very easy on magnetic media, memory, CD, DVD
pen drive.

DSP allows implementation of more sophisticated signal processing


algorithms.
(Complex mathematical algorithms can be easily implemented)

The cost of implementation digital systems is low


due to the lower cost of the digital hardware

Applications of Digital Signal Processing:


- Speech processing
- Signal transmission and reception in telephone/ mobile systems
- Image processing
- Seismology ( Study of seismic signals)
- Oil exploration (Analysis of signals traveling through the layers of earth

Limitations of DSP:
1. The use of ADC and DAC may make the processing
comparatively, slower (Analog systems are faster)
2. The power consumption of digital signal processors may be higher
3. Not suitable for signals with huge bandwidth
(ADC, Sampling rate has to be two times the bandwidth)

Nyquist sampling theorem:


If a function x (t) contains no frequencies higher than B hertz,
it is completely determined by giving its ordinates at a series
of points, spaced 1/ (2B) seconds apart.
(Or the signal with the maximum frequency of B Hertz, can be
completely represented by sampling at a frequency of 2 B Hertz)
Sound signals
Audible range: 300 Hertz to 18000 Hertz
For telephone quality sound signals:
Range of frequencies = 4 KHz
thus sampling frequency required is 8 KHz
For CD quality:
Sampling frequency is 44.1 KHz / channel
Range of frequencies = 22 KHz
Audible range of human beings 20 Hz to 20, 000 Hz

Classification of signals
- Signal Processing technique for any signal depends upon the
characteristics of the signal.
Multi-channel and Multidimensional Signals:
MC-Multiple components
of same signal
MD- Multiple inputs
Real valued signal
s1(t) = A sin 3 t
Complex valued signal
s2 (t ) = A e j3 t = A cos3 t + j A sin 3t
Vector representation
The signal generated by:
multiple sources or
multiple sensors
z

can be represented by components of a vector


s1(t)
s3(t ) =
s2(t)

Three component
signal
y

Three components of ground acceleration measured a few kilometers


from the epicenter of an earthquake

0
22

10

12

14

16

18

20

One dimensional/ multidimensional signal


- one-dimensional signal
- If a signal is a function of one independent variable
It is one-dimensional signal
- Otherwise it is multidimensional signal
- If it is dependent upon more than
one independent variables
- An still image can be two dimensional signal
since I (x, y), intensity of light at any location
depends upon x, y coordinates of the point in the picture

- An television picture can be three dimensional

as I (x, y, t), intensity at any x, y coordinates also


depends upon time t (frame time)
An colour TV image can have three components:
Thus a colour TV image is a three channel (red, green and blue),
three dimensional (x, y and t) signal
Ir (x, y, t)
I (x, y. t) = Ig (x, y, t)
Ib (x, y, t)

Classification of signals

1. Continuous time and discrete time signals

2. Continuous valued and discrete valued signals

3. Deterministic and random signals

Continuous-Time and Discrete Time Signals


1a. Continuous-time signals (Analog signal):
- A signal that exists all the time in a given interval
- These signals are defined for every value of time in the interval
- Most of the naturally occurring signals are continuous in nature
- These signal take on values in the continuous interval (a, b)
where a can be - and b can be +
- Mathematically, these signal can be described by functions of
a continuous variable
x1(t) = cos t
x (t)

x2(t) = e - | t |

- < tx<(t)

where
t

1b. Discrete-time signals


The signals are defined only at specific values of time
x (n)

0 1 2 3 4 5 6 7 8 9 10

- The value of signal between 0 and 1, say at 0.5 is not known


- Thus, discrete time signals exists only
at specific values of time
- These time-instants usually are equidistant at equally
spaced intervals.
(but need not be equidistant)
- The discrete time signal is obtained from a continuous time signal
using the sampling at specific values of time

- A discrete time signal is an approximation of the continuous signal


- To improve the accuracy of the approximation
the sampling period (Ts ) is reduced or
the frequency of the sampling Fs is increased

- A typical discrete time signal can be represented by:


x (n) = A cos .n

( = 2 f)

Example of discrete-time signal:


- |tn|

x (tn) = e

, n = 0, 1, 2, . . . [ (x(t0), (x(t1), . . (x(tn), ]

-If index n of discrete-time instants is used as the independent


variable (i.e., a sequence of numbers),
the signal value becomes a function of an integer variable
-Thus a discrete-time signal can be represented mathematically
by a sequence of real or complex numbers

A discrete time signal is represented as x (n)

( instead of x (t) )

- If the time instants tn are equally spaced, tn = nT

- Signal can be x (nT) =

(T = Time period)

[ x(0T), x(1T), . . x(nT) ]

- There are some discrete time signals which are inherently discrete and
do not require the sampling of continuous signal
i.e.
Accumulating a variable over a period of time
(i.e. number of cars/ hour)

Graphical representation of the discrete time signal


x (n)

x (n) = 0.8n
x (n) = 0 for n <0

for n > 0 and


(negative values of n)

2. Continuous-valued and Discrete-Valued signals


- The values of continuous time or discrete time signals
can be:
continuous (taking all possible values) or
discrete
( taking only a finite set of values)
2a. Continuous-Valued signal:
Signal takes on all possible values in
a range ( which can be finite or an infinite)
(y-axis can be divided into infinite number of levels)
( 4.5, 4.52, 4.53, 4.6 . . .)

2b. Discrete-Valued signal:


Signal takes on values only from a finite set of possible values
(y-axis can be divided into finite number of levels)
- Usually, these values are equidistant,
hence can be expressed as an integer multiple of
distance between two successive values.
(2, 3, 4, 5 or 6,

not 2.322, 3.41, 4.35, 6.012)

Digital signal:
A discrete-time signal having a set of discrete values.
For digital processing of a signal:
- It must be discrete in time
- Its values also must be discrete
- Digital signal is obtained by:
First - Obtaining a discrete time signal by
sampling an analog signal at discrete instants in time
Then Quantizing the values of discrete time to
a set of discrete values
Quantization:
It is the process of converting continuous-valued signal into
discrete valued signal by simple rounding / truncation process
or by mapping to a set of finite values

Digital signal with 4 different amplitude values:


{ 1,

(1, 2, 3 and 4)

1, 1, 2, 1, 3, 2, 4, 2, 1}
X(n)

-2 -1 0 1 2 3 4 5 6 7

Discrete time signal


It is defined for every integer value of

n for - < n < +

x (n) is nth sample of the signal


if x (n) is obtained by sampling then x ( n) x (nT)
[ x(0T, x(1T), . . ,x(nT)
where T is sample period
i.e. the time between two successive samples
Alternative representation of DT signals:
to graphical representation
1. Functional representation,
1,
for n = 1, 3
x (n) =
4, for n=2
0,
otherwise
2. Tabular representation:
n ...
-2 -1 0 1 2 3 4 5 6 . . .
--- -------------------------------------------------------x (n) 0 0 0 1 4 1 0 0 0

3. Sequential representation:

An infinite duration signal or


sequential with the time origin (n = 0)
indicated by symbol

x (n) = { . . . 0, 0, 1, 4, 1, 0, 0, . . . }

A sequence which is 0 for n=0


x (n) = { 0, 1, 4, 1, 0, 0, . . . }

The time origin for a sequence which is zero for n < 0


- First leftmost point is considered to be the origin

Finite duration sequence


x (n) = { 3, -1, -2, 5, 0, 4, -1}

(seven point sequence)

A sequence x (n) =0 for n < 0 x (n) = {0,1, 4, 1}

(4-point sequence

Some elementary Discrete-time Signals


1. The unit impulse sequence
1, for
n =0
(n)

Denoted as (n)

0, for n 0

- It is zero everywhere except n=0,


where it has a unit height
- Also referred as a unit impulse
( zero everywhere, except time t = 0)
- It has unit area

Graphical representation
of unit impulse

-2 -1 0 1

2 3 4

2. The unit step signal

Denoted as u (n)
1, for n 0
u (n)
0 , for n <0

Graphical representation
of step signal

-2 -1 0 1

2 3 4

- It is 1 at positive n ( including at n= 0) and


0 at negative n

3. The unit ramp signal

denoted as u r (n ) u (n)
n, for n 0
0 , for n <0

-2 -1 0 1

2 3 4

4. The exponential signal


- a sequence of the form x (n) = a
0<a<1

for all n
a>1

Unit step signal

Unit ramp signal

x (n) = a n

x(n) = a n

x (n) = a n
If the parameter a is real x (n) is a real signal
- if the parameter a is complex valued,

a r e j

where r and are the parameters

n
xx(n)
=
a
(n ) = r n: e j n
= r n ( cos n + j sin n)

Real part

xg (n) = r n cos n

Imaginary part xj (n) = r n sin n


If

r = 0.9

and = /10

xg (n) = r n cos n = (0.9)n cos /10 . N


xi (n) = r n sin n = (0.9)n sin /10 . N
xg (n) and x j (n) are a damped, decaying ( exponential) cosine function
and damped sine function)

Real part xg

Imaginary part xi

x (n ) = r n e j n

can be represented as amplitude and phase function:

Amplitude function |x (n)| = A (n) r

Phase function x (n) = (n ) n ( n/10)

/2

- Phase function is linear with n


However, phase is defined only over the interval:
equivalently, over the interval:
0<2
Thus (n ) is plotted over the finite interval - <
subtract multiplies of 2 , from (n ) before plotting
The graph for phase (n ) is modulo 2

0
2

-<
or 0 < 2

Classification of Discrete-time signals:


1. Even and odd signals (Symmetric and asymmetric signals)
2. Periodic signals and aperiodic signals
3. Energy signal and power signal

1. Even and odd signals (Symmetric and asymmetric signals)


A signal is even / or symmetric if

x (n) = x (-n)

A signal is odd / or asymmetric if

x(-n) = - x(n) or

x (n) = - x (-n)
For an even signal x(1) = x (-1),
For an odd signal x(1) = - x(-1),
- If x (n) is odd then

x (2) = x (-2) . . .
x (2) = - x(-2) . . .

x(0 ) = 0

Examples:
A sine wave is odd,
while a cosine wave is even signal
sin (50) = - sin(-50)
cos (50) = cos ( -50)

x (n)

-4 -3 -2 -1 0 1 2 3 4

even signal

x (n) = x (-n)

x (n)

-4 -3 -2 -1
0 1 2 3 4

odd signal

x (n) = - x (-n)
x(n)

x (0 ) = 0

Any arbitrary signal can be expressed as the sum of the two components:
- One even and the
- Other odd
- The even signal component can be found out by:

xe(n) = [x(n ) + x(-n)]


odd signal component
xo (n) = [ x(n) x(-n)]
Adding both:
x(n) = xe(n) + xo(n)

Find even and odd components of the given discrete signal:


x (n) = {1, 2, 3, 4, 1, 2, 2}

x (-n) = {2, 2, 1, 4, 3

2,

1}

x (n)e = {1.5, 2, 2, 4, 2, 2, 1.5}

x (n)o = {-0.5, 0, 1, 0, -1, 0, 0.5}

mirror image of x (n)


about origin x (0)
= [x (n) + x (-n)] / 2
= [x (n) - x (-n)] / 2

2. Periodic signals and aperiodic signals


- A signal x (n) is periodic with period N (N> 0)
Only if x (n +N ) = x (n)

for all values of n

Smallest value of N for which the above equation hold good


is called the fundamental period
-The signal is non-periodic if there is no value of N
that satisfies the above equation

Check if
x (n)

x (n) = A cos (n + ) is periodic


Here = 2f
= A cos (2f n + )

x (n +N) = A cos (2f (n +N) + )


= A cos (2f n + 2f N + )
For periodicity:
x (n + N) = x (n)
A cos (2f n + 2f N + ) = A cos (2f n + )
For this to be true:
2f N = =0 or = 2k where k is an integer (0, 1, 2, 3 . . )
or f = k / N where both k and N are integers
Thus for periodicity, f = k /N, where both k and N should be integers

1. Check if cos(0.01 n) is periodic?


0.01 = 2f

f = 0.01/ 2

= 1/200 cycles per sample,

f = k/ N ratio of two integers


As k =1 and N is 200, both are integers, so cos (0.01 n) is periodic

2. Check if x (n) = sin 3n


Here
3 = 2f or f = 3/ 2
since f cannot be expressed as fraction of two integers ,
the signal x (n) = sin3n is not periodic

Energy and power signal


Energy of a signal for discrete time signal x (n)

is defined as E |x (n) | 2
n=-
- As the magnitude square is used for x (n),
The definition is applies to real as well as complex-valued signals

- The energy of a signal can be finite or infinite

Energy signal
- If Energy E is finite then x (n ) is called an energy signal
- The finite Energy can be called Ex of signal x (n)

- Many signals possess infinite energy,


but have a finite average power
1
N
Average Power P = lim

N
N+1 n=-N

|x (n)| 2

If the signal is defined over a finite interval - N n N


N
Then
EN
|x (n)|2
n = -N
Then signal energy can be expressed as
N

E Lim

The average power of the signal x (n) is:


1
P lim
---------- EN
N 2N +1
If EN is finite,

then

as N 1/(2N +1) 0
thus Average power P 0
Thus Power for a finite energy signal is zero

EN

Energy and power of a unit step signal u(n)


1 for n = 0
- Step signal u(n) =
0 otherwise

1
0 1 2 3..

E |x (n) | 2 = |u (n) | 2 = 1 + 1 + 1 . . .
=
n=-
0
Since the E is infinite, the unit step signal is not an energy signal

Energy

1
N
The average Power P = lim
-----------
u 2 (n)
N 2N +1
n=0
( summation of u2 (n) is N +1)
1 (N+1)
1 + 1/ N
1
= lim
----------- = lim
----------- = ------N 2N +1
N
2 + 1/N
2
Consequently, the unit step sequence is power signal
and its energy is infinite

Find Energy for the signal x (n) = an u (n)

- As energy of D. T. Signal is:


n = -
=

|an u (n)|2

Since u (n) is 1, for 0 to

EN
for n = 0 to
|an . 1|2

a| < 1

[u (n) step signal]

|x (n)|2

u (n) is a init step


for n = 0 to

=[a2]0 + [a2]1 + [a2]2 + . .


(Geometric series An = 1 + A + A2 + A3 . . . = 1/ 1 A if A < 1)

Thus E = 1/ 1- a2

if |a2| < 1

Explain whether the following signals are power signal or energy signal.
i 0.5 n u (n)
- First calculate energy
E of x (n) is given by:

of the signal

n=
n=
E = | x (n)| 2 =
| 0.5n u (n)| 2
n=-
n=-
-Since this signal is multiplied by unit step: n is from 0 -
( signal is zero for n > 0)
n=
n=
n=
E=

| (1/2) n | 2 =
(1/2) 2 n =
(1/4) n
n=0
n =0
n=0
n=
{Standard Geometric series
A n 1 + A + A2+ A 3 . . . 1/ (1-A) (A = )}
n=0
Thus E = 1 / (1 -1/4 ) = 1/ = 4/ 3 = 1.333
- Thus E is finite
Since Energy is finite it is energy signal

Operations on discrete time signals:


The mathematical transformation from one signal to another
is represented as:
y (n) = T [x (n)]
- Operations involving:
- Independent variable ( time)
- Dependent variable (amplitude)
The basic operations on DT signals are:
1. Time shifting
2. Time reversal
3. Time scaling
4. Scalar multiplication
5. Signal addition and
multiplication

Independent variable, time

Dependent variable,
Amplitude

Operations involving independent variable ( time)


1. Time Shifting
- Shift n by - k in x (n)

Delay
(k is an integer)
y (n) = x (n - k)
for k = 3, y (0) =x (-3)

- if k is negative the shifting results in advancing by |k| units


Future

X (n ) = { -1, 0, 1, 2, 3 , 4, 4, 4, 4, 4}

X(n-3) = { -1, 0, 1, 2, 3 , 4, 4, 4, 4, 4}

shifting delay

k=3

(-3 becomes origin)

X( n +2) { -1, 0, 1, 2, 3 , 4, 4, 4, 4, 4}
advancing
k= -2 (+2 becomes origin)

Note: delay /advance is easy in stored signal


(However, advancing in real time, generated signal is not possible.)

2. Time reversal
Folding/ Reflection about the time origin n = 0:
- Replacing n by n
y (n) = x (-n)
x (n ) =
{2, 2, 2, 0, 1, 2, 3, 4}

x (-n) = {4, 3, 2, 1, 0, 2, 2, 2}

x (-n - 2) = {4, 3, 2, 1, 0, 2, 2, 2}
(Folding and delayed by 2)

(-2 becomes origin)

3. Time scaling:
- Changing in time scale
- Two types down scaling and up scaling
- Down scaling is represented as:
y (n) = T [x (n)]
Example

y (n) = x (2n)

x (n) = { -3, -2, -1, 0, 1, 2, 3, 4, 4, 4, 4, 4, 4, 4}

y (n) is taking every other sample from x (n)


y(0) = x (0) = 4,
y(-1) =x(-2)= 2
y (1) = x(2) = 4
y(-2) =x(-4)= 0
y(2) = x(4) = 4,
Thus y (n) { -2, 0 , 2 ,4 , 4, 4, 4}

x (n)
= { -3, -2, -1, 0, 1, 2, 3, 4, 4, 4, 4, 4, 4, 4}

y (n) = x (2n)
= { -2,
0, 2, 4, 4, 4, 4}

Sampling is reduced to half

If the original analog signal is xa (t) then

x (n) = xa (nT)

The analog signal value at time nT

If y (n ) = x (2n) , then y (n) = xa (2nT) = xa (n 2T)


thus Ts sampling period is increased from T to 2T
sampling frequency Fs is decreased from 1/T to 1/2T
This is down scaling operation

Up scaling
y (n) = x (n / 2 )
x (n) = { 5, 4, 3, 2, 1, 2, 3, 4, 5}

y (0 )= x (0/2) = x(0) = 1
y (1)= x (1/2) = x(0.5) = no sample
y (2 )= x (2/2) = x(1) = 2
y (3 )= x (3/2) = x(1.5) = no sample
y (4 )= x (4/2) = x(2) = 3
y (6 )= x (6/2) = x(3) = 4
y (8 )= x (8/2) = x(4) = 45
Thus y (n) is expanded version of x (n) with y (1), y(3), y(5) . . No sample

-8 -7 -6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6 7 8
The sampling rate is increased from 1/T to 2/T
Upscaling

Addition, multiplication, and scaling of sequences


- Amplitude modification includes addition, multiplication and scaling
of discrete-time signals
4. Amplitude scaling:
- Multiplying by a constant
y (n) = A x (n)
- <n <
5. Signal addition and multiplication
- The sum of two signals
y (n) = x1 (n) + x2 (n)
- <n <
(add corresponding terms)
- The product of two signals
y (n) = x1 (n) . x2 (n) - <n <
(multiply corresponding terms)

At -2, -1 . 0 = 0
-1, 2 . 0 = 0
0, 2 . 1 = 2
1, 1 . 2 = 2
2, 0 . 3 = 0

Discrete time Systems


- DT system is a device/ algorithm that operates on
a discrete-time signal (input/ excitation),
according to some well-defined rule,
to produce another discrete-time signal (output/ response)
- It is a set of operations performed on input signal x (n) to produce
output signal y (n)
x (n) is said to be transformed to y (n)
y (n) T [ x (n) ]

or

x (n )

y (n)

where T is transformation

(transformed to )

Determining the response:


|n|,
-3n 3
x (n) =
0, otherwise

x (n)=

{ 0, 3, 2, 1, 0, 1 ,2, 3, 0}

Y (n) = x (n -1) = { 0, 3, 2, 1, 0, 1 ,2, 3, 0}


past

delayed by 1 y (0) = x ( -1)

Y (n) = x (n+1) = { 0, 3, 2, 1, 0, 1 ,2, 3, 0}


future
Y (n) = 1/3 [x (n+1) + x (n) + x (n-1) ]
mean of the current + old + next values

advanced by 1 y (0) = x (+1)

x (n) taken as reference

= { 0, 1, 5/3, 2, 1, 2/3 , 1 ,2, 5/6, 1, 0, 0}

x(n ) = { 0, 0, 3, 2, 1, 0, 1 , 2, 3, 0, 0, 0}
Y(0) = 1/3[ x (1) +x (0) + x(-1)] = 2/3

y (n) = max ( x (n+1), x (n), x (n-1)


x (n)= { 0, 3, 2, 1, 0, 1 ,2, 3, 0}

=
{ 0, 3, 3, 3, 2, 1, 2, 3, 3, 3, 0 }

n
y (n) =
k=-

x (k) = x (n) +x (n-1) + x ( n-2) + . .

Y ( 0) = x (0) +x (0-1) + x (0-2)


+ x (0-3), +x (0-4) + x (0-5)

x (n)= { 0, 3, 2, 1, 0, 1 ,2, 3, 0}

y (n) = { 0, 3, 5, 6, 6, 7, 9,12,12} ( k = - ) sum of all past values ?

y(0) = x(0) + x(-1) + x(-2) . .


0+3
0 + 3 +2
0 + 3 +2+1
0 + 3 +2+1+ 0 + 1

Classification of Discrete Time Systems


Discrete time system is a device or algorithm that operates on
a discrete time signal
- It is represented as y (n) = T [x (n)]
1. Static and Dynamic systems

(Static - current input)

2. Causal and Anti-causal systems


no future
3. Linear and Non-linear systems

(Causal - Present and past

4. Time variant and Time invariant systems


5. Stable and unstable systems
6. FIR and IIR systems
(Finite Impulse Response/ infinite impulse response)

. Static versus Dynamic systems


Static system:
A system is static if the output of a system depends
only the current input and
not on the past or future input
- Examples of static systems:
y (n) = T [x (n)]
y (n) = 4. x (n),
y (n) = a x (n)
y (n) = log x (n)
y (n) = A cos x (n)
y (n ) = a x (n ) + b x3 (n)
- In each case y (n) requires only present value of input x (n)
- Static systems are also called memory less system
as no memory is required to store previous input values
e. in case of y (n) = 4 . x (n),

y (1) = 4 . x (1),

y (2) = 4 . x (2),

...

Dynamic system
- All the systems that are not static.
- In dynamic systems, the output may depends on the
present as well as
past input signals (Even future signals)
y (n) = x (n) + x (n-2) is a dynamic system
- as for finding y (4) = x (4) + x (2)
past input is required

( current input + past input)

- Thus the dynamic system requires memory to store the past inputs
Dynamic system is also called memory system

Examples:
y (n) = [x (n) + x (n-12)]
for y (15), x (15) and x (3) are required
y (n) = x (n) . x (n-1)
n
y (n) = x (n -k)
Requires Finite memory
k=0
n
y (n) = x (n - k) Requires Infinite memory
k= -

2. Causal and Non-causal system


- A system is causal if the output of the system y (n) at any n
depends on present and past inputs but
not on future inputs
- Example:
y (n ) = x (n)
(present

+ x (n-2)
past

+ x (n-3)
past)

Non-causal system:

- If the output also depends upon the future input


then it is Non causal system
- Example:
y (n) = x (n) + x (n+1 )
(present + future)

Future

y (4) = x (4) + x (5)

- Hence for a non-causal system,


- Future input needs to be predicted to find the present
Other examples:
y (n) = x (2n)
y (n) = x (n) x (n+2)

3. Linear and non- linear


- Linear systems satisfies superposition principle:
- The superposition principle states that the response to
a weighted sum of input signals,
should be equal to the corresponding
weighted sum of the outputs of the system
to individual input signals.
i. e. T[ a1x1(n) + a2x2(n) ] = a1T [ x1(n) ] + [ a2 T [ x2(n) ]
(response to the weighted sum of inputs =
the weighted sum of responses to the individual inputs)

x1 (n)
x2 (n)

a1
a2

y1 (n) = T [a1 x1 (n)

+ a2 x2 (n)]

a1
AND

x1 (n)
x2 (n)

T
T

+
a2

y2 (n) = a1 T [x1 (n)] + a2 T [ x2 (n)]

Check whether the y (n) = n. x (n) is a linear or non linear system?

As y (n) = T [x (n)]
Hence, the system is n . x (n)
For linearity:
Check if T [x1 (n) + x2 (n)] = T [ x1(n) ] + T[x2 (n)] or
y1 (n)
=
y2 (n)
so y1 (n) = T [x1 (n) + x2 (n)]
= n. [x1 (n) + x2 (n)] = n. x1 (n) + n.x2 (n)
y2 (n) = T [ x1(n) ] + T[x2 (n)] = n . x1(n) + n. x2 (n)
Hence y1 (n)

= y2 (n)

The system y (n) = n. x (n) is linear


similarly y (n) = n . x2 (n)

is non-linear can be proved

Time-variant and time- invariant systems


Time-invariant system:
In Time-invariant system
the input /output characteristic does not change with time
- suppose input x (n ) produces y (n) at any stage,
y (n) = T [ x (n)]
- if the input signal is delayed by k units,
the output y (n- k) will be same as y (n)
except that it is delayed by k units
- First delay the input by k samples , and observe the output y (n, k)
- Delay the output by k samples y (n - k)
If y (n- k) = y (n, k) the system is time-invariant

x (n)

y (n)

x (n -k)

y (n -k)

Check whether the following systems are time invariant?


1. y (n) = e

x (n)

- Delay input by k

y (n, k) = e x (n- k)

- Delay the output y (n) by k, y (n - k ) = e

x (n-k)

Since both are equal the system is time-invariant


2. y (n) = n . x (n) is time variant as
1. Delaying input y (n, k) = n. x (n - k)
2. Delaying output y (n - k)
y (n- k) = ( n- k) . x (n- k)
Both are not equal the system is time-variant

Stable and unstable systems :


For a Stable system
a bounded input, produces bounded output
Let

Mx is finite numbers such that Mx < ,

the input is said to be bounded if | x (n )| Mx <


for all value of n
My is a finite number such that My < ,
the output is said to be bounded if |y (n)| My <
for all value of n
- If for a bounded input x (n), there is output
the system is unstable
Check whether the following systems are stable?
1. y (n) = e x (n)
2. y (n) = x (2n)
Both are stable, as for bounded input, there is bounded output

Analysis of Linear Time Invariant systems (LT I ) systems:


- LT I systems are characterized in time domain by
their response to a unit sample sequence
- Any arbitrary signal can be represented as
weighted sum of unit sample sequences
Techniques for analysis of Linear systems:
1. Direct solution of the input-output equation
2. First decomposing the input signals into elementary signals, then
determining the response to each elementary signal and
adding all the responses

Direct solution of the input-output equation


Input output equation has the form
y (n) = F [ y (n-1), y (n-2), . . Y (n-N), x (n), x (n-1), x (n-1), . . . X (n-M)]
F [. . . . ] denotes some function of quantities in brackets
- Specially, for LTI (Linear Time-invariant) systems
- The general form of input-output relationship:
N
M
y (n) = - ak y (n - k) +
k=1
k =0

bk x (n - k)

ak and bk are constant parameters that specify the system


ak and bk are independent of x (n ) and y (n)

and

- The above equation is called a difference equation

- It represents one way to characterize the behavior of a


discrete-time LTI system

Second method of analyzing the behavior of a linear system


- Resolve the input signal x (n) into
weighted sum of elementary signal components [xk (n) ]
- (The elementary signals are selected so that the response
to each component is easily determined)
- Use the linearity property to add the responses to
individual components to obtain the total response (output)
x (n ) = ck xk (n)
k
xk (n) are the

(1)
elementary signal component

ck are the set of amplitudes (weighting coefficients)


in the decomposition of the signal x (n)

If the yk (n) is the response of the system to

the elementary component xk (n)


then yk (n) T [xk (n)]

(2)

Assuming the system response to the ck xk (n)

is ck [ yk (n)]

As a consequence of scaling property of the linear systems

x (n) from (1)

=T [ ck xk (n) ] = ck T [ xk (n)]
k
k
Using additive property of the linear system
y (n) = T [x (n)]

y (n) = ck yk (n)
k

( y k (n) from (2)

(The response to the input x (n ) whose components are ck x (n)


equals to the weighted sum of
the responses to the components)

Resolution of a discrete time signal into impulses


- Let an arbitrary signal be x (n)
{ 1, 2, 3, 1, 2, 1, 3}
-3 -2 -1 0 1 2 3

- Resolving the signal x (n) into a sum of unit sample sequences


let kth component of x (n), xk (n) = (n -k)

[ (n- k)th impulse]

x (n)
where k is the delay of the unit sample sequence

x (3)

- Multiplying x (n) and (n -k)

k =3

-3 -2 -1 0 1 2 3

Since (n - k) is zero everywhere except at n = k (Impulse)


the result of the multiplication will be a sequence:
which is zero at everywhere except at n = k
Therefore x (n). (n- k) = x (k). (n - k)

if we take different values of k we will get:

x (n) =
x (k) (n -k)
k=-
= summation of an infinite number of unit sample sequence , (n -k)
having amplitude values of x (k)

Representation of a signal in terms of


weighted sum of shifted, discrete impulses :
Any arbitrary signal can be represented as
summation of shifted and scaled impulses
x (n) = {0.5, 2, 1, 0.5, 1, 2, 0.5}

- The sample x (0) can be obtained by


multiplying x (0), the magnitude, with a unit function (n)
x (0) ; for n =0
i. e. x (0) . (n) =
0 ; for n 0

= 0.5 . 1 = 0.5

( multiply)

(n)
1
x (0) . (n)
0

0.5

n
-3 -2 -1 0 1 2 3

x (n) = {0.5, 2, 1, 0.5, 1, 2, 0.5}

Similarly other components can be obtained:


:
x (1). (n-1) = 1 . 1
=1
x (2). (n-2) = 2 . 1
=2
x (3). (n-3) = 0.5 . 1 = 0.5
x (-3). (n +3) = 0.5 . 1 = 0.5
x (-2). (n +2) = 2 . 1 = 2
x (-1). (n +1) = 1 . 1 = 1

Thus x (n) = 0.5. (n +3) + 2 . (n +2) + 1. (n +1) +


+ 0.5 . (n) . + 1 . (n -1 ) + 2 . (n -2) + 0.5 . (n -3)
The signal x (n) can be viewed as a summation of scaled and
shifted impulses

in general, x (n) =
x (k). (n -k)
k=-

Let x (n ) = {2, 4, 0, 3}

Resolve into a sum of weighted impulse sequence


since x (n ) is non-zero at the time instant n = -1, 0, 2
we need three impulses at k = -1, 0, 2
x (n) = 2 .(n+1) + 4. (n) + 3 .(n-2)

Impulse response and convolution


y (n) = T [ x (n) ]
x (n)

Output y (n) = transfer function x (n)


y (n)

If input x (n) is a unit impulse (n) then the output of the system is
known as impulse response h (n)
(n)

h (n)

- Therefore impulse response: h (n) = T [ (n)]


(transfer function of unit impulse)
Impulse response completely characterize the system

Convolution sum
Since input signal x (n) can be represented as
weighted sum of discrete impulses

i. e. x (n) =
x (k). (n- k)
(where n is the time index,
k=-
k is a parameter showing
the location of input impulse

Thus output signal y (n) = T [x (n)] = T

x (k). (n -k)
k=-
- If the system is linear:

y (n) = T [x (k). (n- k) ]


k=-

y (n) = x (k). T [ (n- k) ]


(1)
k=-

Let T [ (n -k)]
y (n) =

= h (n, k)

x (k). h (n, k)
k=-

h (n, k) unit pulse response delayed


(From (1)

If the system is time invariant:


h (n, k) = h (n -k)
( Response with
the input delayed =
The response with

the output delayed)


y (n) = x (k) . h (n - k)
k=-
( The output/ response at n is summation of
values of input x at k, multiplied by unit impulse response at n - k,
for all values of k)


y (n) =
k=-

x (k). h (n - k)

The above equation is called convolution sum:

For a Linear Time Invariant (LTI) system


if input sequence x (n) and
the impulse response h (n) is known
The response y (n) can be found out by the convolution sum

y (n) = x (k). h (n - k)
k=-
(Output y at n = sum of values of input x (n) at n = k,
multiplied by unit impulse response at n -k,
for all values of k
The convolution sum is represented as
y (n) = x (n)

* h (n)

Properties of Convolution
1. Commutative Law : x (n) * h (n) = h (n) * x (n)
k
k
y (n) = x (k). h (n - k)
k=-

h ( k). x (n -k).
k=-

2. Associative Law: [ x (n) * h1 (n) ] * h2 (n) = x (n) * [ h2 (n) * h1 (n) ]


3. Distributed Law: x (n) * [ h1 (n) ] + h2 (n)] = x (n) * h1 (n) + x (n) * h2 (n)

Computation of Linear Convolution


1. Graphical Method
2. Tabular method
The linear convolution is given by:

y (n) =

x (k) . h (n - k)
k=-
( Response y (n) at n, is summation of values of x (n) at k,
multiplied by unit impulse response at n - k, for all values of k)
Calculating the value y (n) for time instant n = 0

y(0) =
x (k). h(0 - k) =
x (k). h( - k)
k=-
k=-
( h ( -k) indicates folding )

For n = 1

y(1) =

x (k). h(1 - k)
k=-

=
k=-

x (k). h(- k + 1)

h(- k + 1) indicates shifting of folded signal h


h (-k +1) Indicates h (-k), the folded signal is delayed by 1 sample
For n = 2

y(2) =
x (k). h(1 - k)
= x (k). h(- k + 2)
k=-
k=-
h(- k + 2) indicates shifting of folded signal h
h (-k +2) Indicates h (-k), the folded signal is delayed by 2 sample
And so on

Steps for computing the convolution between x (k) and h (k)


1 Folding: fold h (k )

h ( -k)

2 Shifting: shift h ( -k)

h (-k +1)

3 Multiplication: Multiply

x (k) by h (n k),

find all the product terms for all the values of k


4 Summation: Sum all the product terms to find y (n)
Find y (n) for all values of n, following steps 2 to 4 for each
value of n


Range of n and k , for calculating y (n) =
k=-
Range of n

x (k) . h (n - k)

Lowest value of n in y( n) = lowest value of n in x (n) +


lowest value n in h h (n)

to
Highest value of n in y( n) = highest value of n in x (n) +
highest value n in h (n)

Range of k
The range k will be same as range of n in x (n)

Obtain the impulse response of a system h (n) = { 1, 2, 1, -1}

Input Signal
x (n) = { 1, 2, 3, 1}

Range of n in y (n) = (Lowest value of n in x (n) + lowest value of n in h (n))


to (highest value of n in x (n) + highest value of n in h (n))
n in y (n) = (-1 + 0 = -1)

to (2 + 3 = 5 ) = -1 to + 5

Range of k = range of n in x (n ) = 0 to 3
Determining the response/ output y:
For n = 0: y (0)
h (n) =
{ 1, 2, 1, -1}

y (0) = x (k). h(0- k)


h (-k) = folded h = { -1, 1, 2, 1}
k=0

y (0) = x (k) . h (-k)


k varies from 0 to 3
x (k)
. h (-k )
=
{1, 2, 3, 1} .
{ -1, 1, 2, 1}
= {0, 0, 1, 2, 3, 1} .
{ -1, 1, 2, 1, 0, 0}
aligning origin
= { 0x-1, 0x1, 1x2, 2x1, 3x0, 1x0}
= (0 + 0 + 2 + 2
+0 + 0)

=4

For n = 1: y (1) h (-k) = folded


h(-k) = { -1, 1, 2, 1}
y (1) = x (k) . h (-k +1)
k varies from 0 to 3
x (k)
.
h (-k +1 )
y (1) =
{1, 2, 3, 1} .
{ -1, 1, 2, 1}
= {0, 1, 2, 3, 1} .
{ -1, 1, 2, 1, 0}
= { 0x-1, 1x1, 2x2, 3x1, 1x0,}
=

( 0

+1

+4

+3

For n = 2: y (2)
y (2) = x (k) . h (-k +2)
y (2) =

+0) = 8
k varies from 0 to 3

x (k)
. h (-k +2 )
{1, 2, 3, 1} .
{ -1, 1, 2, 1}

= {1, 2, 3, 1} .

{ -1, 1, 2, 1}

= { 1x-1, 2x1,

3x2, 1x 1}

=(

+6

-1

+2

+1) = 8

For n = 3: y (3)
y (3) = x (k) . h (-k +3)

k varies from 0 to 3

x (k)
. h (-k +3 )
{1, 2, 3, 1} .
{ 0, -1, 1, 2, 1}

y (3) =

= {1, 2, 3, 1, 0} .

{ 0, -1, 1, 2, 1}

= { 1x0, 2x-1, 3x1, 1x2 + 0 x 1}


=

(0

-2

+3

+2 +0 ) = 3

Similarly y (4) = -2
y( 5) = -1
y ( -1) = 1
Entire response of the system { . . . ., 0, 0, 1, 4, 8, 8, 3, -2, -1, 0, 0, . . . }

Obtain the linear convolution of following sequences:


x (n) = { 1, 2, 1, 2}

h (n) = { 1, 1, 1}

Since the sequences are given in terms of n


obtain x (k) and h (k ) by substituting x by k
x (k) = { 1, 2, 1, 2}
h (k) = { 1, 1, 1}

Convolution is given by:

y (n) =
x (k). h (n - k)
k=-
Range of n for y (n) = xL + hL = 0 + 0 = 0 to xh + hh = 2 + 3 = 5

= 0 to 5
Range of k : same as the value of n in x (n) = 0 to 3

Tabulation method of linear convolution


Let x (n) = {x(0), x(1), x(2) }

and h (n) = {h(0), h(1), h(2)}

Step I: Form the matrix as shown below:


x(n)
x(0) x(1)
x(2)
h(0)
h (n) = h(1)
h(2)

Step II Multiply the corresponding elements of x (n) and h (n), and


Step III Separate out elements diagonally as shown below:
x(n)
x(0) x(1)
x(2)
h(0) h(0).x(0) h(0).x(1) h(0).x(2)
h(n) =

h(1) h(1).x(0) h(1).x(1) h(1).x(2)

h(2) h(2).x(0) h(2).x(1) h(2).x(2)

Step IV : Add the elements in each block.


This will give corresponding values of y (n)
Y(0) = h(0).x(0)
Y(1) = h(0).x(1) + h(1).x(0)
Y(2) = h(0).x(2) + h(1).x(1) + h(2).x(0)
Y(3) = h(1).x(2) + h(2).x(1)
Y(4) = h(2).x(2)
Y (n) = {y(1), y(2), y(3), y(4)}
Range of n in y (n) = Lowest value of n in x (n) + lowest value of n in h (n)
to highest value of n in x (n) + highest value of n in h (n)
= (0 + 0 = 0) to (2 + 2 = 4 ) = 0 to + 4
The range of n for y (n) = 0 to 4

Compute convolution y (n) = x (n)* h( n)


x (n) = { 1, 1, 0, 1, 1}

x (n) 1 1 0 1
h (n)
1 1
1 0 1

h (n) = {1. -2, -3, 4}

1
1

.1

-2

-2

-2

-2

-2

. -2

-3

-3

-3

-3

-3

.-3

.4

y (n) = {1, 1-2, 0-2-3, 1+0 -3+4,


= {1, -1,
-5,
2,

Origin
1 -2 +0 +4, -2 -3 +0, -3+4 , 4}
3,
-5,
1,
4 }

x (n) . h.(n) = y (n)


(1+1+0+1+1) . (1 -2 -3 +4) = ( 1 -1 -5 +2+3 -5 +1 +4)
4. 0 = 0

Correlation:
- Correlation is used for the comparison of two signals
- It is measure of degree to which two signals are similar
- Cross correlation
Correlation of two separate signals is known as
cross correlation
- Auto correlation
Correlation of a signal with itself is known as
auto correlation
Cross Correlation:
Correlation between two signals x (n) and y (n)
+
rx y ( l ) = x (n) . y (n - l)
where l = 0, 1, 2, 3 . . .
n=-
Index l is time shift (or lag) parameter and
L
the subscripts x, y on cross correlation sequence rx y ( l )
indicates the sequences being correlated

Auto correlation
The correlation of a signal with itself to determine time delay
between the transmitted signal and the received signal
+

rx x (l) =
n=-

x (n) . x (n - l)

where l = 0, 1, 2, 3 . . .

Compute the cross-correlation between:


x (n) = { 1, 1, 0, 1}

y (n) = { 4, -3, -2, 1}

- Cross correlation of x (n) and y (n) is given by :


+
rx y(l) = x (n) . y (n - l) where l = 0, 1, 2, 3 . . .
n=-
- Range of n:
As y (n) is delayed by l, and x (n) is not changed
so the range of n in the summation is
same as x (n) = -2 to +1
1
So r x y (l) =
x (n) . y (n - l)
n = -2

Range of n in x (n) is -2 to +1
Range of n in y (n) is -2 to +1
Range of l:
since the equation y (n- I) should have maximum value at y (n) = y(1)
nl=1
maximum
Starting value of n in summation is -2
-2 l = 1 for n = -2
I = -1 -2 = -3
or starting value of l = -3

in x (n)

The equation y (n l ) should have minimum value at y n ) = y (-2)


n- I = -2
but the summation should stop at n =1
in x (n)
1- I = -2
or I = 3
So the range of I = -3 to + 3

The range of n = -2 to +1
The range for l is -3 to 3
1
rx y(l) =
n = -2

x (n) . y (n - l)

where l = -3 to + 3

Now x (n) = { 1, 1, 0, 1} and y (n) = { 4, -3, -2, 1}

Rx y (-3) for I =-3, n = -2 to +1


n=
-2
-1
0
1
x (n) . y (n - l)
r xy(-3) = x(-2) .y(-2 +3 ) + x(-1). y (-1 +3) + x(0) . y( 0 +3) + x(1). Y(1+3)
x (-2). Y(1)
+ x(-1) y (+2)
+ x(0)y(+3)
+ x(1). Y(4)
1 . 1
+ 1 .0
+0 . 0
+ 1 .0
= 1
+
0
+ 0
+0
=1

rxy (-2) for I =-2, n = -2 to +1


n=
-2
-1
0
1
x (n) . y (n - l)
r xy(-2) = x(-2) y(-2 +2 ) + x(-1) y (-1 +2) + x(0)y( 0 +2) + x(1). Y(1+2)
x(-2) y (0 )
+ x(-1) y (+1)
+ x(0)y(+2) + x(1). Y(3)
1 . -2
+ 1 .1
+0 . 0
+ 1 .0
-2
+1
+0
+0 =1
rxy (-1) for I =-1, n = -2 to +1
r xy(-1) = x(-2) y(-2 +1 ) + x(-1) y (-1 +1) + x(0)y( 0 +1) + x(1). Y(1+1)
x(-2) y(-1)
+ x(-1) y 0)
+ x(0)y(+1) + x(1). Y(2)
1 . -3
+ 1 . -2
+0 . 1
+ 1 .0
-3
+ -2
= -5
rxy (0) for I = 0, n = -2 to +1
rxy(0) = x(-2) y(-2 +0 ) + x(-1) y (-1 +0)
= x(-2) y(-2)
+ x(-1) y (-1)
1 . 4
+ 1 . -3
4
+ -3

+ x(0)y( 0 +0) + x(1). Y(1+0)


+ x(0)y( 0)
+ x(1). Y(1)
+ 0 . -2
+ 1 .1
+0
+1
=2

rxy ( 1) for I =1, n = -2 to +1


r x y(1) = x(-2) y(-2 -1 ) + x(-1) y (-1 -1) + x(0)y( 0 -1) + x(1). Y(1-1)
= x(-2) y(-3 )
+ x(-1) y (-2) + x(0)y(-1) + x(1). Y(0)
1 . 0
+ 1 .4
+ 0 . -3
+ 1 .-2
0
+4
+0
-2
Similarly rxy(2) = -3 and rxy(3) = 4
Rxy(l) = { 1, -1, -5, 2, 2, -3, 4}

=2

Simple method to calculate correlation


Solution using property of correlation
rx y (l ) = x (n) * y (-n)

( = convolution of x (n) and y (-n))

- So take x (n) as it is:

x (n) = { 1, 1, 0, 1}

- Fold y (n):
y (n) = { 4, -3, -2, 1}

y (-n) = { 1, -2, -3, 4 )

Obtain the convolution of x (n) and y (-n) by matrix method


y(-n) x(n)
1
-2
-3
4
rx y(l) =
rxy(l) =

1 1
1 1
-2 -2
-3 -3
4 4

0 1
0 1
0 -2
0 -3
0 4

(1, 1 -2, 0-2-3, 1+0-3+4, -2+0+4, -3+0,


(1,

-1,

-5,

2,

2,

-3,

4}
4}

Determine the autocorrelation of the following signal


x (n) = (1, 2, 1, 1}

Let x1 (n) be {1, 2, 1, 1}

x2 (n) be {1, 2, 1, 1}
Folding x2 (n)

x2(-n)= { 1, 1, 2, 1}

So rxx (l) = x (n) * x (-n)

x1(n)
1, 2, 1, 1
x2(-n)
1 1 2 1 1
1 1 2 1 1
2 2, 4, 2, 2

1 1, 2, 1, 1
{ 1,
+ 3,1+2+2,
+5, 1+1+4+1,
+ 7,
5, ,+2+1,
3, +11}
}
x x(l)
rx xR(I)
= ={ 1,
2 +1,
1++2+2,

Properties of Correlation
1. The result of autocorrelation is maximum
when the signal matches with itself and there is no phase shifting.
2. Auto-correlation is an even function

rx x (l) = rx x (- l)
3. The cross-correlation is not commutative.
That means rx y(l) ry x (l)

FIR and IIR


From the convolution sum:

y (n) =
x (k). h (n - k)
k=-
h (k) is the impulse response of the system
FIR
If h (k) is of finite duration, the system is called
Finite Impulse Response (FIR) system
IIR
If h (k) is of infinite duration, the system is called
Infinite Impulse Response (IIR) system

END

2. x (t) = sin(480 t) + 3 sin (720 t) is sampled with Fs = 600 times per sec.
(1) What are the frequencies in radians in the resulting DT signal x (n)?
(2) If x (n) is passed through an ideal interpolator,
what is the reconstructed signal?
To find DT signal frequencies sample the CT signal
Put

t = n Ts = n / Fs = n/600
Ts sampling time period
Fs sampling frequency = 1/ Ts
x ( t) = sin ( 480 t ) + 3 sin (720 t)
x [ n]
= sin ( 480 n / 600) + 3 sin (720 n/ 600)
= sin ( 0.8 n ) + 3 sin ( 1.2 n )
1.2 n = (2 0.8) n = -0.8 n
= sin (0.8 n ) + 3 sin (-0.8 n) = - 2 sin (0.8 n)
= - 2 sin (w n)
thus w = 0.8 radians
To find reconstructed signal put n = t . Fs = 600 t
= - 2 sin ( 0.8 . 600 t) = - 2 sin ( 480 t)

You might also like