FDM vs TDM
Frequency division multiplexing(Analog Exchange)
•
•
•
•
•
•

1 chl
300 Hz to 3400 Hz
3chls : Basic Group(BG): 12 – 24 KHz
12chls: Sub Group : 4BG: 60 -108KHz
60chls:SuperGroup:5SubGp: 312 – 552 KHz
900chls:Master Group:16 super Gp 312 – 4028 KHz
2700chls:Super Master Group: (3 MG):312 -12336
KHz
1
Time division multiplexing
•
•
•
•
•

Voice frequency 4KHZ
Sampling rate 8kilo samples per second
Each sample requires 8bits
1 voice frequency requires 64kbps
32 voice frequencies (PCM) requires
2Mbps
2
System Capacity
2Mb

PCM

8Mb

Ist order

34Mb
2nd order
140Mb/STM1
3rd order

Higher order

No. of voice channels
30

120

2Gb/STM16

480

1920

30000

3
Chapter 2
Discrete-Time signals and systems
•
•
•
•
•

Representation of discrete-time signals
(a)Functional
(b)Tabular
(c)Sequence
Examples of singularity functions
Impulse, Step and Ramp functions and shifted
Impulse, Step and Ramp functions
4
Energy and power signals
• If E=infinity and P= finite then the signal is
power signal
• If E=finite and P= zero then the signal is
Energy signal
• If E=infinite and P= infinite then the signal
is neither energy nor power signal

5
Periodic and non-periodic signals
• X(n+N) = x(n) for all n, and N is the
fundamental time period.

6
Block diagram Representation of
Discrete-Time systems
•
•
•
•
•
•
•

Adder
A constant multiplier
A signal multiplier
A unit delay element
A unit advance element
Folding
Modulator
7
Classcification of Discrete-time
systems
•
•
•
•
•
•

Static vs dynamic systems
Time invariant vs time-variant systems
Linear vs nonlinear systems systems
Causal vs non causal systems
Stable vs unstable systems
Recursive vs non recursive systems
8
Static vs dynamic systems
• A discrete time system is called static or
memoryless 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. In any other case, the system is
said to be dynamic or to have memory.

9
Time invariant vs time-variant
systems
• A system is called time-invariant if its
impute-output characteristics do not change
with time.
• X(n-k)
y(n-k)

10
Linear vs nonlinear systems systems
• Linear system obeys the principle of
superposition
• T[a x1(n) + b x2(n)] = a T[x1(n)] + b T[x2(n)]

11
LTI system
• The system which obeys both Linear
property and Time invariant property.

12
Causal vs non causal systems
• A system is said to be causal if the output of
the system at any time depends only on
present and past inputs, but not depend on
future inputs. If a system doesnot satisfy
this definition , it is called noncausal
system.
• All static systems are causal systems.
13
Stable vs unstable systems
• An arbitrary relaxed system is said to be
bounded input – bounded output (BIBO)
stable if an only if every bounded input
produces a bounded output.
• y(n) = y2(n) + x(n) where x(n)=δ(n)

14
Recursive vs non recursive systems
• Recursive system uses feed back
• The output depends on past values of
output.

15
Interconnection of discrete time
systems
•
•
•
•

Addition/subtraction
Multiplication
Convolution
Correlation (a) Auto (b) Cross correlation

16
Convolution
• Used for filtering of the signals in time
domain. Used for LTI systems only
• 1.comutative law
• 2.associative law
• 3.d17istributive law

17
18
Correlation
• Auto correlation improves the signal to
noise ratio
• Cross correlation attenuates the unwanted
signal

19
Digital signal Distortion

Distortion less transmission y(n) = α.x(n-l)
20
Random signal vs Deterministic
signal
• Deterministic signal: can be represented in
mathematical form since the present, past
and future values can be predicted based on
the equation
• Random signal: can not be put in
mathematical form. Only the average, rms,
peak value and bandwidth can be estimated
with in a given period of time.
21
Advantages of digital over analog
signals
•
•
•
•
•
•

Faithful reproduction by reshaping
Information is only 2 bits (binary)
Processed in microprocessor
Signals can be compressed
Error detection and correction available
Signals can be encrypted and decrypted
22
Disadvantages
• Occupies more bandwidth
• Difficult to process the digital signals in
microwave frequency range since the speed
of the microprocessor is limited (3 GHz)
• Digital signals have to be converted into
analog signals for radio communication.

23
Advantages of Digital signal processing
over analog signal processing
•
•
•
•
•
•
•
•

Flexible in system reconfiguration
Accuracy , precise and better tolerances
Storage
Low cost
Miniaturization
Single micro processor
Software operated (programmed)
Artificial intelligence
24
Multi channel and Multidimensional
Signals
• Multichannel:
• S(t) = [ s1(t), s2(t), s3(t)…]
• Multi Dimensional: A value of a signal is a
function of M independent variables.
• S(x,y,t) = [ s1(x,y,t), s2(x,y,t), s3(x,y,t)…]

25
Analog-discrete-digital-PCM
•
•
•
•
•
•

Analog frequency
Demo of Time and frequency signals
Digital frequency
Periodic and aperiodic signals
Sampling theorem
A to D / D to A conversion
26
Variable sampling rate

27
x(t)

|X(f)|

t

0

-f

(a)
x

δ

∞

(t) =

n = -∞

δ (t-nT

s)

...
s

-2T

s

0

2T

s

4T

m

∞

1
T s

δ (f-nf

n = -∞

s

...
t

-2f

s

s

-f

0

s

fs

2f

f
s

(d)
δ

(t)

|X

s

(f)|

...
-4T

s

-2T

s

0

2T

s

4T

t
s

(e)

Ch2. Formatting

)

...

(c)
x s (t) = x(t) x

f

f

(b)

X δ (f) =

...
-4T

0

m

...
-2f

s

-f

s

-f

m

0
(f)

[Fig 2.6] Sa mp ling the orem u sing th e freq u en cy
co nvo lution pro pe rty o f the Fou rier tra nsform

fm

fs

2f

f
s

28

[6]
Sampling
• The operation of A/D is to generate sequence by taking
values of a signal at specified instants of time
• Consider a system involving A/D as shown below:

x(t )

x (t)

Analog-todigital
converter

x(nT )
x[n]
xh(t)

0

T

2T 3T 4T 5T

Digital-toanalog
converter

xh (t )

x(t) is the input signal and xh(t) is the
sampled and reconstructed signal

t

T = sample period

29
Impulse Sampling
What is Impulse Sampling?
– Suppose a continuous–time signal is given by x(t),
-∞ < t < +∞
– Choose a sampling interval T and read off the value of
x(t) at times nT, n = -∞,…,-1,0,1,2,…,∞
– The values x(nT) are the sampled version of x(t)

30
Impulse Sampling
• The sampling operation can be represented in a block
diagram as below:
∞

δ T (t ) = ∑ δ (t − nT )
n = −∞

x(t )

xS (t )

• This is done by multiplying the signal x(t) by a train of
impulse function δT(t)
• The sampled signal here is represented by xS(t) and the
sampling period is T

31
Impulse Sampling
• From the block diagram, define a mathematical
representation of the sampled signal using a train of δfunction

xs ( t ) =

∞

∑ x( nT )δ ( t − nT )

n = −∞

∞

= x( t ) ∑ δ ( t − nT )
n = −∞

= x ( t )δ T ( t )
continuous –time signal
function

Train of periodic impulse 32
function
Impulse Sampling
• We therefore have

• And

1 ∞ jnω 0t
δT (t) = ∑ e
T n = −∞

It turns out that the
problem is much
easier to understand
in the frequency
domain. Hence, we
determine the
Fourier Transform
of xs(t).

1 ∞ jnω 0t
xs ( t ) = x( t ) ∑ e
T n = −∞
1 ∞
= ∑ x ( t ) e jnω 0t
T n = −∞
33
Impulse Sampling
• Looking at each term of the summation, we have the
frequency shift theorem:
CTFT

x( t ) e jnω 0t ↔ X ( ω − nω 0 )

• Hence the Fourier transform of the sum is:

1 ∞
X s ( ω ) = ∑ X ( ω − nω 0 )
T n = −∞
34
Impulse Sampling
i.e. The Fourier transform of the sampled signal is simply
the Fourier transform of the continuous signal repeated at
the multiple of the sampling frequency and scaled by 1/T.
X(ω)

1

ω

0
Xs(ω)

1/T

-2ω0

-ω0

0

ω0

ω

2ω0 35
Nyquist Sampling Theorem
• The Nyquist sampling theorem can be stated as:

If a signal x(t) has a maximum frequency content (or
bandwidth) ω max, then it is possible to reconstruct x(t)
perfectly from its sampled version xs(t) provided the
sampling frequency is at least 2 × ω max

36
Nyquist Sampling Theorem
• The minimum sampling frequency of 2 × ωmax is known as
the Nyquist frequency, ωNyq
• The repetition of X(ω) in the sampled spectrum are known
as aliasing. Aliasing will occur any time the sample rate is
not at least twice as fast as any of the frequencies in the
signal being sampled.
• When a signal is sampled at a rate less than ωNyq the
distortion due to the overlapping spectra is called aliasing
distortion

37
Multichannel Tx-Rx
• Station A

cable

Station B

38
39
Design of discrete time systems

40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
PERIODIC AND APERIODIC SIGNALS OF ANALOG AND
DISCRETE

56
57
58
Computing the Z-transform: an
example = α nu[n]
x[n]
• Example 1: Consider the time function

X(z) =

∞

∑

x[n]z −n =

n=−∞

=

1
1− αz

−1 =

∞

∑

n=0

α n z −n =

∞

∑ (αz −1 )n

n=0

z
z −α
59
Another example …

x[n] = −α nu[−n − 1]

X(z) =

∞

∑

−1

x[n]z −n =

n=−∞

∑

−α n z −n =

n=−∞

•

∑

n=−∞

(αz −1 )n =

∑ (αz −1 )n

n=−∞

l = −n;n = −∞ ⇒ l = ∞;n = −1 ⇒ l = 1
−1

−1

∞

∞

l=1

l=0

∑ −(zα −1 )l = 1 − ∑ (zα −1 )l = 1 −

1

1
=
1 − zα −1 1 − αz −1
60
61
62
63
64
65
66
67
68
69
70
71
Frequency domain of a rectangular pulse

Fourier coefficients of the rectangular pulse train with time
period Tp is fixed and the pulse width tow varies.

72
73
74
75
SAMPLING IN TIME DOMAIN

76
77
78
79
80
81
82
83
84
85
86

Tdm fdm

  • 1.
    FDM vs TDM Frequencydivision multiplexing(Analog Exchange) • • • • • • 1 chl 300 Hz to 3400 Hz 3chls : Basic Group(BG): 12 – 24 KHz 12chls: Sub Group : 4BG: 60 -108KHz 60chls:SuperGroup:5SubGp: 312 – 552 KHz 900chls:Master Group:16 super Gp 312 – 4028 KHz 2700chls:Super Master Group: (3 MG):312 -12336 KHz 1
  • 2.
    Time division multiplexing • • • • • Voicefrequency 4KHZ Sampling rate 8kilo samples per second Each sample requires 8bits 1 voice frequency requires 64kbps 32 voice frequencies (PCM) requires 2Mbps 2
  • 3.
    System Capacity 2Mb PCM 8Mb Ist order 34Mb 2ndorder 140Mb/STM1 3rd order Higher order No. of voice channels 30 120 2Gb/STM16 480 1920 30000 3
  • 4.
    Chapter 2 Discrete-Time signalsand systems • • • • • Representation of discrete-time signals (a)Functional (b)Tabular (c)Sequence Examples of singularity functions Impulse, Step and Ramp functions and shifted Impulse, Step and Ramp functions 4
  • 5.
    Energy and powersignals • If E=infinity and P= finite then the signal is power signal • If E=finite and P= zero then the signal is Energy signal • If E=infinite and P= infinite then the signal is neither energy nor power signal 5
  • 6.
    Periodic and non-periodicsignals • X(n+N) = x(n) for all n, and N is the fundamental time period. 6
  • 7.
    Block diagram Representationof Discrete-Time systems • • • • • • • Adder A constant multiplier A signal multiplier A unit delay element A unit advance element Folding Modulator 7
  • 8.
    Classcification of Discrete-time systems • • • • • • Staticvs dynamic systems Time invariant vs time-variant systems Linear vs nonlinear systems systems Causal vs non causal systems Stable vs unstable systems Recursive vs non recursive systems 8
  • 9.
    Static vs dynamicsystems • A discrete time system is called static or memoryless 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. In any other case, the system is said to be dynamic or to have memory. 9
  • 10.
    Time invariant vstime-variant systems • A system is called time-invariant if its impute-output characteristics do not change with time. • X(n-k) y(n-k) 10
  • 11.
    Linear vs nonlinearsystems systems • Linear system obeys the principle of superposition • T[a x1(n) + b x2(n)] = a T[x1(n)] + b T[x2(n)] 11
  • 12.
    LTI system • Thesystem which obeys both Linear property and Time invariant property. 12
  • 13.
    Causal vs noncausal systems • A system is said to be causal if the output of the system at any time depends only on present and past inputs, but not depend on future inputs. If a system doesnot satisfy this definition , it is called noncausal system. • All static systems are causal systems. 13
  • 14.
    Stable vs unstablesystems • An arbitrary relaxed system is said to be bounded input – bounded output (BIBO) stable if an only if every bounded input produces a bounded output. • y(n) = y2(n) + x(n) where x(n)=δ(n) 14
  • 15.
    Recursive vs nonrecursive systems • Recursive system uses feed back • The output depends on past values of output. 15
  • 16.
    Interconnection of discretetime systems • • • • Addition/subtraction Multiplication Convolution Correlation (a) Auto (b) Cross correlation 16
  • 17.
    Convolution • Used forfiltering of the signals in time domain. Used for LTI systems only • 1.comutative law • 2.associative law • 3.d17istributive law 17
  • 18.
  • 19.
    Correlation • Auto correlationimproves the signal to noise ratio • Cross correlation attenuates the unwanted signal 19
  • 20.
    Digital signal Distortion Distortionless transmission y(n) = α.x(n-l) 20
  • 21.
    Random signal vsDeterministic signal • Deterministic signal: can be represented in mathematical form since the present, past and future values can be predicted based on the equation • Random signal: can not be put in mathematical form. Only the average, rms, peak value and bandwidth can be estimated with in a given period of time. 21
  • 22.
    Advantages of digitalover analog signals • • • • • • Faithful reproduction by reshaping Information is only 2 bits (binary) Processed in microprocessor Signals can be compressed Error detection and correction available Signals can be encrypted and decrypted 22
  • 23.
    Disadvantages • Occupies morebandwidth • Difficult to process the digital signals in microwave frequency range since the speed of the microprocessor is limited (3 GHz) • Digital signals have to be converted into analog signals for radio communication. 23
  • 24.
    Advantages of Digitalsignal processing over analog signal processing • • • • • • • • Flexible in system reconfiguration Accuracy , precise and better tolerances Storage Low cost Miniaturization Single micro processor Software operated (programmed) Artificial intelligence 24
  • 25.
    Multi channel andMultidimensional Signals • Multichannel: • S(t) = [ s1(t), s2(t), s3(t)…] • Multi Dimensional: A value of a signal is a function of M independent variables. • S(x,y,t) = [ s1(x,y,t), s2(x,y,t), s3(x,y,t)…] 25
  • 26.
    Analog-discrete-digital-PCM • • • • • • Analog frequency Demo ofTime and frequency signals Digital frequency Periodic and aperiodic signals Sampling theorem A to D / D to A conversion 26
  • 27.
  • 28.
    x(t) |X(f)| t 0 -f (a) x δ ∞ (t) = n =-∞ δ (t-nT s) ... s -2T s 0 2T s 4T m ∞ 1 T s δ (f-nf n = -∞ s ... t -2f s s -f 0 s fs 2f f s (d) δ (t) |X s (f)| ... -4T s -2T s 0 2T s 4T t s (e) Ch2. Formatting ) ... (c) x s (t) = x(t) x f f (b) X δ (f) = ... -4T 0 m ... -2f s -f s -f m 0 (f) [Fig 2.6] Sa mp ling the orem u sing th e freq u en cy co nvo lution pro pe rty o f the Fou rier tra nsform fm fs 2f f s 28 [6]
  • 29.
    Sampling • The operationof A/D is to generate sequence by taking values of a signal at specified instants of time • Consider a system involving A/D as shown below: x(t ) x (t) Analog-todigital converter x(nT ) x[n] xh(t) 0 T 2T 3T 4T 5T Digital-toanalog converter xh (t ) x(t) is the input signal and xh(t) is the sampled and reconstructed signal t T = sample period 29
  • 30.
    Impulse Sampling What isImpulse Sampling? – Suppose a continuous–time signal is given by x(t), -∞ < t < +∞ – Choose a sampling interval T and read off the value of x(t) at times nT, n = -∞,…,-1,0,1,2,…,∞ – The values x(nT) are the sampled version of x(t) 30
  • 31.
    Impulse Sampling • Thesampling operation can be represented in a block diagram as below: ∞ δ T (t ) = ∑ δ (t − nT ) n = −∞ x(t ) xS (t ) • This is done by multiplying the signal x(t) by a train of impulse function δT(t) • The sampled signal here is represented by xS(t) and the sampling period is T 31
  • 32.
    Impulse Sampling • Fromthe block diagram, define a mathematical representation of the sampled signal using a train of δfunction xs ( t ) = ∞ ∑ x( nT )δ ( t − nT ) n = −∞ ∞ = x( t ) ∑ δ ( t − nT ) n = −∞ = x ( t )δ T ( t ) continuous –time signal function Train of periodic impulse 32 function
  • 33.
    Impulse Sampling • Wetherefore have • And 1 ∞ jnω 0t δT (t) = ∑ e T n = −∞ It turns out that the problem is much easier to understand in the frequency domain. Hence, we determine the Fourier Transform of xs(t). 1 ∞ jnω 0t xs ( t ) = x( t ) ∑ e T n = −∞ 1 ∞ = ∑ x ( t ) e jnω 0t T n = −∞ 33
  • 34.
    Impulse Sampling • Lookingat each term of the summation, we have the frequency shift theorem: CTFT x( t ) e jnω 0t ↔ X ( ω − nω 0 ) • Hence the Fourier transform of the sum is: 1 ∞ X s ( ω ) = ∑ X ( ω − nω 0 ) T n = −∞ 34
  • 35.
    Impulse Sampling i.e. TheFourier transform of the sampled signal is simply the Fourier transform of the continuous signal repeated at the multiple of the sampling frequency and scaled by 1/T. X(ω) 1 ω 0 Xs(ω) 1/T -2ω0 -ω0 0 ω0 ω 2ω0 35
  • 36.
    Nyquist Sampling Theorem •The Nyquist sampling theorem can be stated as: If a signal x(t) has a maximum frequency content (or bandwidth) ω max, then it is possible to reconstruct x(t) perfectly from its sampled version xs(t) provided the sampling frequency is at least 2 × ω max 36
  • 37.
    Nyquist Sampling Theorem •The minimum sampling frequency of 2 × ωmax is known as the Nyquist frequency, ωNyq • The repetition of X(ω) in the sampled spectrum are known as aliasing. Aliasing will occur any time the sample rate is not at least twice as fast as any of the frequencies in the signal being sampled. • When a signal is sampled at a rate less than ωNyq the distortion due to the overlapping spectra is called aliasing distortion 37
  • 38.
    Multichannel Tx-Rx • StationA cable Station B 38
  • 39.
  • 40.
    Design of discretetime systems 40
  • 41.
  • 42.
  • 43.
  • 44.
  • 45.
  • 46.
  • 47.
  • 48.
  • 49.
  • 50.
  • 51.
  • 52.
  • 53.
  • 54.
  • 55.
  • 56.
    PERIODIC AND APERIODICSIGNALS OF ANALOG AND DISCRETE 56
  • 57.
  • 58.
  • 59.
    Computing the Z-transform:an example = α nu[n] x[n] • Example 1: Consider the time function X(z) = ∞ ∑ x[n]z −n = n=−∞ = 1 1− αz −1 = ∞ ∑ n=0 α n z −n = ∞ ∑ (αz −1 )n n=0 z z −α 59
  • 60.
    Another example … x[n]= −α nu[−n − 1] X(z) = ∞ ∑ −1 x[n]z −n = n=−∞ ∑ −α n z −n = n=−∞ • ∑ n=−∞ (αz −1 )n = ∑ (αz −1 )n n=−∞ l = −n;n = −∞ ⇒ l = ∞;n = −1 ⇒ l = 1 −1 −1 ∞ ∞ l=1 l=0 ∑ −(zα −1 )l = 1 − ∑ (zα −1 )l = 1 − 1 1 = 1 − zα −1 1 − αz −1 60
  • 61.
  • 62.
  • 63.
  • 64.
  • 65.
  • 66.
  • 67.
  • 68.
  • 69.
  • 70.
  • 71.
  • 72.
    Frequency domain ofa rectangular pulse Fourier coefficients of the rectangular pulse train with time period Tp is fixed and the pulse width tow varies. 72
  • 73.
  • 74.
  • 75.
  • 76.
  • 77.
  • 78.
  • 79.
  • 80.
  • 81.
  • 82.
  • 83.
  • 84.
  • 85.
  • 86.