You are on page 1of 6

1/16/2020

Signals

DIGITAL SIGNAL PROCESSING


EEE F434 Lecture 1-3
Sarang Dhongdi

Signals Signal types


 Function of independent variable such as time, distance,  CT signal vs DT signals
position, pressure etc.
 Natural vs synthetic (man-made) signals
 Speech – Air pressure as function of time at a point in space
– 1D
 Real valued vs complex valued
 B&W image – Light intensity as function of two spatial
coordinates -2D  Even or Odd signals

 B&W video – Sequence of images – w.r.t. time -3D  Energy or power signals

 Colour video – multi-channel (3 channel), three 3D signals  Deterministic vs random

Digital Signal Processing Purpose ?

The purpose of such processing may be to estimate


It is concerned with the representation of signals by characteristic parameters of a signal or to
sequences of numbers or symbols and the transform a signal into a form which is more
processing of these sequences. desirable in some sense.

1
1/16/2020

Scope of DSP – No visible sign of saturation Cell Phone


Field Applications
SPACE -Space photograph enhancement
-Data compression
-Intelligent sensory analysis by remote space probes
MEDICAL -Diagnostic imaging (CT, MRI, ultrasound, and others)
-Electrocardiogram analysis
-Medical image storage/retrieval
TELECOMM -Voice and data compression
-Echo reduction
-Signal multiplexing
-Filtering
MILITARY -Radar
-Sonar
-Secure communication
INDUSTRIAL -Oil and mineral prospecting
-Process monitoring & control
-Nondestructive testing

What does this course offer? What does this course offer?
- In teaching – - In lab
- Review of “Signals and Systems” - Tools to Explore
- Analog filter design - MATLAB (Z transform, FFT functions etc.)
- Digital filters design - Simulink (Filter Design toolbox)
- Types of filters (LP, HP, BP, BS)
- Realization of digital filters
- Finite word length effects
- Multirate DSP

Why Digital? Advantages


 Signals and coefficients (of digital signal  Guaranteed Accuracy
operations) are specified as binary words.  Perfect Reproducibility
 No drift due to temp and age
 Easy storage and retrieval
 Greater reliability, smaller size, lower cost, low
power consumption, higher speed
 Multiplexing functionality / time-sharing
 Greater flexibility
 Ease of operation at lower frequencies (compared  Superior performance
to analog which requires large size components)

2
1/16/2020

Disadvantages Historical perspective


 For Large Bandwidth applications  Discrete time signal processing has a rich history.
 ADC/DAC
 Until the early 1950, the signal processing was typically
done using analog systems using electronic circuits.
 Finite word length effect
 Digital computers were expensive and had limited
capabilities.

 Signal processing could not generally be done in real


time.

Historical perspective Historical perspective


 Rather digital computers were used to simulate a  When signals were being processed, certain algorithms
signal processing system before implementing it on were developed out of flexibility of digital computer,
and had no apparent practical implementation on
analog hardware – flexible experimental analog system –Ex. Cepstrum analysis.
environment.
 It prompted the notion of “all-digital” implementation
 Essentially it was to approximate or simulate an of signal processing systems.
analog signal processing system.
 Algorithm of Cooley and Tukey– FFT – this was
inherently discrete time concept.
 Speed, size and cost were main limitations.

Historical perspective Historical perspective


 FFT algorithm reduced the computation time of FT  Rather than mere approximation of analog systems,
by orders of magnitude, permitting implementation DTSP became an important field of investigation in
of signal processing algo with lesser processing time it’s own right.
– interactive experimentation was possible.
 DTSP is a dynamic and rapidly growing field, it’s
 Considering that FFT algo can be implemented on fundamentals are well formulated.
special-purpose digital hardware, many signal
processing algorithms appeared to have practical
implementation.

3
1/16/2020

Digital signal processing – Digital Filters Block Diagram


 Need of Digital filter?
 Adaptive filters – periodic changes in coefficients
 Frequency selective filters with adjustable cut-off
frequencies
 Linear phase filter
 Multirate signal processing
 Many specific algo developed only in/for digital domain
 Cepstrum analysis
 Homomorphic filtering

Spectrum view Nyquist sampling rate

where
 fs = sampling frequency

 BW = bandwidth of input signal

Aliasing distortion undersampling

4
1/16/2020

Discrete time impulse function Discrete time unit step function

kronecker delta u[n]


δ[n]
1 1

n n

Discrete time exponentials


Exponentially varying DT sinusoids

(0.9)ncos (nπ/6)
(0.8)n 1 (-0.8)n
1

n n

(1.2)n
1

n
(1.1)ncos (nπ/6)

5
1/16/2020

Discrete time sinusoids Non-uniqueness


cos (nπ/12 + π/4) cos (3πn/7)

cos (11πn/7)

Fundamental range of frequencies


Example
cos (2nπ)
x1 (t )  c o s (6  t )
x 2 ( t )  c o s (1 4  t )
cos (nπ/8) or cos (15nπ/8)
x3 (t )  c o s(2 6 t )
Sampling rate is 10 Hz for all the signals cos (nπ/2) or cos (3nπ/2)

x1 ( n )  c o s ( 0 .6  n )  c o s ( 0 .6  n )
cos (nπ)
x 2 ( n )  c o s (1 . 4  n )  c o s ( 0 .6  n )
x 3 ( n )  c o s ( 2 .6  n )  c o s ( 0 .6  n )

You might also like