CSE 304-Digital Signal Processing

Dr. Amir A. Khan

Office : A-218, SEECS
9085-2162; amir.ali@seecs.edu.pk

Lecture-2

CSE-304 Digital Signal Processing

Dr. Amir / SEECS-NUST

Signal Processing
Input
signal

Signal
Processing
System

Output
signal

continuous-time signal
signal

discrete-time signal
signal

continuous-time signal
analog
analog output
system
system

discrete-time signal

discretediscretetime
time output
system
system

digital signal

digital signal
signal

Lecture-2

Function : extract (output) desired

information (e.g. filtering,
parameter estimation)

digital
digital output
system
system

CSE-304 Digital Signal Processing

Dr. Amir / SEECS-NUST

Signal Processing
Input
signal

analog
signal

Signal
Processing
System

A/D

digital
signal

Output
signal

DSP

Function : extract (output) desired

information (e.g. filtering,
parameter estimation)
digital
signal

D/A

analog
signal

Analog input analog output

Digital recording of music
Analog input digital output
Touch tone phone dialing
Digital input analog output
Text to speech
Digital input digital output
Compression of a file on computer
Lecture-2

CSE-304 Digital Signal Processing

Dr. Amir / SEECS-NUST

Limitations of Analog Signal Processing

Accuracy limitations due to

Component tolerances
Undesired nonlinearities
Limited repeatability due to
Tolerances
Changes in environmental conditions
Temperature
Vibration
Sensitivity to electrical noise
Inflexibility to changes
Difficulty of implementing certain operations
Nonlinear operations
Time-varying operations
Difficulty of storing information

Lecture-2

CSE-304 Digital Signal Processing

Dr. Amir / SEECS-NUST

Pros and Cons of Digital Signal

Processing

Pros
Accuracy can be controlled
Repeatable
Sensitivity to electrical noise is minimal
Dynamic range can be controlled using floating point numbers
Flexibility can be achieved with software implementations
Non-linear and time-varying operations are easier to implement
Digital storage is cheap
Digital information can be encrypted for security
Price/performance and reduced time-to-market
Cons
Sampling causes loss of information
A/D and D/A requires mixed-signal hardware
Limited speed of processors
Quantization and round-off errors
Lecture-2

CSE-304 Digital Signal Processing

Dr. Amir / SEECS-NUST

Discrete Time Signals

Obtained from periodic sampling of analog signals x[n] xa [nT ] (T=sampling interval)

x[n] is nth-sample of sequence x[n] where n extends from inf to +inf

Unit Impulse Signal

Unit Step Signal

Courtesy : Book Oppenheim

Lecture-2

CSE-304 Digital Signal Processing

Dr. Amir / SEECS-NUST

Discrete Time Signals

Sampling/Sifting Property of Unit Impulse Signal

x[n] [n n0 ] x[n0 ] [n n0 ]

h[k ] [n k ] h[n]

Impulse Response

Difference between continuous and discrete time complex

exponentials/sinusoidal sequences

x[n] Ae j (0 2 ) n Ae j0 n e j 2n Ae j0 n

x[n] A cos (0 2r )n A cos0 n

Complex exponentials sequences/sinusoids with frequencies
0+2r are indistinguishable from each other
In discrete complex exponentials/sinusoids, we need consider frequencies
only over a range of 2 (- to or 0 to 2)
Lecture-2

CSE-304 Digital Signal Processing

Dr. Amir / SEECS-NUST

Discrete Time Systems

x[n]

y[n]=T(x[n
])
Impulse Response ?

Moving Average
k M
1
y ( n)
x(n k )

M 1 M 2 1 k M
1

Courtesy : Book Oppenheim

Lecture-2

MM
11
hh((nn))
((nnkk))

M
M1 M
M2 11kkMM1
1

11

hh((nn)) M
M11M
M2211
00

M
M11nnM
M22
otherwise
otherwise

Very efficient in smoothing

out the signal :
averaging = low pass filtering
CSE-304 Digital Signal Processing

Dr. Amir / SEECS-NUST

Properties of Systems
1. Memory / Memoryless Systems
2. Causal / Non-causal Systems
3. Time Invariant /Variant System
4. Linear / Non-linear Systems

9
Lecture-2

CSE-304 Digital Signal Processing

Dr. Amir / SEECS-NUST

Basic System Properties

Memory / Memory-less Systems
System is Memory-less if output for each value of independent
variable at given time is dependent only on the input at that same
time
2
2

y[n] (2 x[n] x [n])

Memory-less Systems

Memory Systems
n

x[k ]

y[n] x[n]

y[n]

y (t ) x(t )

y[n 1] x[n]

v(t ) i (t ) R

Accumulator

n 1

x[k ] x[n]

Average
Lecture-2

1
y[n] ( x[n 1] x[n] x[n 1])
3

CSE-304 Digital Signal Processing

Dr. Amir / SEECS-NUST

Basic System Properties

Invertible / Non-invertible Systems
System is Invertible if distinct input results in distinct outputs.

y[n] 2 x[n]
1
w[n] x[n] y[n]
2
x[n]

y[n]

w[n]=
x[n]

Is Accumulator an invertible System ?

Courtesy : Book Oppenheim

Lecture-2

CSE-304 Digital Signal Processing

11
Dr. Amir / SEECS-NUST

Basic System Properties

Linear / Non-linear Systems (Superposition)

x1 (t ) x 2 (t ) y1 (t ) y 2 (t )

Additivity
Homogenity

ax(t ) ay (t )
Linearity condition

ax1 (t ) bx 2 (t ) ay1 (t ) by 2 (t )
Examples

y (t ) x 2 (t ), Non Linear

Lecture-2

CSE-304 Digital Signal Processing

Dr. Amir / SEECS-NUST

Basic System Properties

Time (Shift)-Invariant Systems

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

x(n)
x(n-k)

TT[[]]

y(n)

x(n)

y(n-1)

x(n-1)

y(n-2)

x(n-2)

Lecture-2

y(n-k)

CSE-304 Digital Signal Processing

Dr. Amir / SEECS-NUST

Linear Time (Shift) Invariant Systems

Superposition and Time Invariance allow the complete
characterization of a system to any input knowing only the
impulse response

x(n)

xx((nn)) xx((kk))((nnkk))

k k

y ( n)

TT[[]]

x(k )T [(n k )] x(k )h(n k )

Impulse Input

Lecture-2

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

yy((nn))TT xx((kk))((nnkk))
kk

CSE-304 Digital Signal Processing

Impulse response
with some shift

Dr. Amir / SEECS-NUST

LTI-system CharacterizationConvolution

x(n)

Lecture-2

h(n)
h(n)

CSE-304 Digital Signal Processing

x(n)*h(n)

Dr. Amir / SEECS-NUST

Convolution-Demonstration

Syste
m

Syste
m

Courtesy : Dr. Murtaza

Lecture-2

CSE-304 Digital Signal Processing

Dr. Amir / SEECS-NUST

Convolution-Demonstration

Syste
m

y[n]=x[0]h[n-0]+x[3]h[n-3]+x[6]h[n-6]

y[n]

x[k ]h[n k ]

k
Syste
m

Lecture-2

CSE-304 Digital Signal Processing

Dr. Amir / SEECS-NUST

LTI system Properties-Convolution

x(n)

hh11(n)
(n)

hh22(n)
(n)

y(n)

x(n)

hh22(n)
(n)

hh11(n)
(n)

y(n)

x(n)

hh11(n)*h
(n)*h22(n)
(n)

y(n)

These
These systems
systems are
are identical.
identical.
Lecture-2

CSE-304 Digital Signal Processing

Dr. Amir / SEECS-NUST

LTI system Properties-Convolution

hh11(n)
(n)
+

x(n)

y(n)

hh22(n)
(n)

x(n)

hh11(n)+h
(n)+h22(n)
(n)

y(n)

These
These two
two systems
systems are
are identical.
identical.
Lecture-2

CSE-304 Digital Signal Processing

Dr. Amir / SEECS-NUST

LTI System Properties-Causality

Causal systems --- output for y(n0) depends only on x(n) with n n0

y[n] 2 x[n]
y[n] x[n] x[n 1]
A causal system whose impulse response h(n) satisfies

hh((nn)) 00 for
for nn 00

Are all real-world systems causal?

Not-exactly
Are all real-time systems (working on single sample acquisition principle)
causal?
Yes indeed
20
Lecture-2

CSE-304 Digital Signal Processing

Dr. Amir / SEECS-NUST

LTI System Properties-Stability

Stability of LTI Systems
Bounded Input produces a Bounded Output (BIBO)
Bounded Input

x[n] B

With x[n] as input, output would be

y[n]

For stability we check the output magnitude

x[k ]h[n k ]

h[k ]x[n k ]

y[n]
y[n]

h[k ]

x[n k ]

Since |x[n]|<B |x[n-k]|<B, with a maximum value of B

y[n] B

h[k ]

Lecture-2

CSE-304 Digital Signal Processing

Dr. Amir / SEECS-NUST

LTI System Properties-Stability

y[n] B

h[k ]

Output would thus be bounded if

h[k ]

For an LTI system to be stable, the condition is thus that its impulse
response is absolutely summable
Think about Finite Impulse Response (FIR) and Infinite Impulse
response (IIR) systems from stability perspective
FIR systems are always stable
Lecture-2

CSE-304 Digital Signal Processing

IIR systems can be unstable

Dr. Amir / SEECS-NUST