DIGITAL SIGNAL PROCESSING

Dr Rashmi Patil
Program:B. Tech. (E&TC) Semester: B.TECH Semester-VI

Course:Digital Signal Processing Code: BET6419

Teaching Scheme Evaluation Scheme

FA SA Total
Lectur
Tutorial Credit Hours FA2
e 1
3 -- 3 3 20 20 60 100

2
� Prior knowledge of:
� Analog ,Digital signals, Signal Transforms ,Mathematics Is
essential

� Objectives:

� 1. To provide background and fundamentals for the analysis and

processing of digital signals.
� 2. To make students familiar with DFT and the computations of
FFT algorithms.
� 3. To describe the design procedure and types of realization of
digital filters to the students.
3
� 4. To explain the necessity of Digital Signal Processors, their
architectures and applications
�
� Outcomes:
After learning the course, the students will be able to

� 1. Explicate various stages of Digital Signal Processing

� 2. Select proper tools to evaluate system response using frequency
transformation techniques like DFT, FFT
� 3. Design the FIR digital filters for given specifications
� 4. Realize the IIR digital filters using different structures
� 5. Explain applications of digital signal processing algorithms to
various areas like medical, speech, image
� 6. Illustrate various architectures and features of digital signal
processors
4
�
� Syllabus:
1. DSP Introduction
2. Transforms
3. FIR Filter Design
4. IIR Filter Design
5. DSP Applications
6. Digital Signal Processor Architecture,GPU architecture
� Self Study: Sampling,Nyquist criteria ,Basics of Z
transform,Microprocessor

5
� Text Books:
� John G.Proakis, D.G.Manolakis, "Digital Signal Processing:Principles,Algorithms
and Applications", Pearson Education, 4th edition
� A.Nagoor Kani,”Digital Signal Processing”.McGrawHill 2nd edition
� Venkataramani.B, Bhaskar.M, “Digital Signal Processors, Architecture,
Programming and Application”,Tata McGraw Hill, New Delhi, 2003.
� Reference Books:
� S.K. Mitra, "Digital Signal Processing Computer Based Approach", Tata MacGraw
Hill , 3rd edition
� Avtar Singh, S.Srinivasan DSP Implementation using DSP microprocessor with
Examples from TMS32C54XX -Thamson 2004.
� Salivahanam, A Vallavaraj, C. Guanapriya , "Digital Signal Processing", Tata
MacGraw Hill , 1st edition
� Rulph Chassaing, Donald Relay ”Digital Signal Processing and Applications with
TMS3206713 and TMS320C6416DSK”, 2nd Edition, Wiley 2014
6
� P. Ramesh Babu, "Digital Signal Processing" Scitech publication, 4th edition
NPTEL COURSES
� https://nptel.ac.in/courses/117/102/117102060/
� https://www.classcentral.com/course/youtube-electrica
l-digital-signal-processing-47650/classroom
� .https://ocw.mit.edu/courses/res-6-008-digital-signal-pr
ocessing-spring-2011/

7
UNIT 1
� DSP Introduction: Basic Elements of DSP and its
requirements, Sampling and Reconstruction,
Advantages & Disadvantages, Mapping between
analog frequencies to Digital frequencies, System
Stability and Correlation

8
Definition

� Anything that carries information can be called as signal.

� It can also be defined as a physical quantity that varies
with time.
� The process of operation in which the characteristics of a
signal Amplitude, Shape, Phase,Frequency,etc.
undergoes a change is known as signal processing.

9
SIGNAL PROCESSING
� Humans are the most advanced signal processors – speech and pattern
recognition, speech synthesis,…
� • We encounter many types of signals in various applications – Electrical
signals: voltage, current, magnetic and electric fields,… – Mechanical
signals: velocity, force, displacement,… – Acoustic signals: sound,
vibration,… – Other signals: pressure, temperature,…
� • Most real-world signals are analog – They are continuous in time and
amplitude – Convert to voltage or currents using sensors and transducers
� • Analog circuits process these signals using – Resistors, Capacitors,
Inductors, Amplifiers,…
� • Analog signal processing examples – Audio processing in FM radios –
Video processing in traditional TV sets

10
ANALOG SIGNAL PROCESSING
� Most of the natural signals are analog
⚫ Analog signal processing
� Differentiation, Integration, Filtering, Amplification
� Implemented via passive or active electronic circuitry.

11
LIMITATIONS OF ANALOG SIGNAL PROCESSING
� • Inflexibility to changes
� Accuracy limitations due to – suffer from Component tolerances –
Undesired nonlinearities
� • Limited repeatability due to – Tolerances – Changes in environmental
conditions
� Difficulty of storing information
� Difficulty of implementing certain operations – Nonlinear operations –
Time-varying operations
� • Temperature Variation
� • Sensitivity to electrical noise
� • Limited dynamic range for voltage and currents

12
DIGITAL SIGNAL
� Signal: a function of independent variables (time,
distance, position….)
� Independent variable-Continuous/discrete
� Discrete time signal
⚫ A sequence of numbers
� Analog signal
⚫ continuous time signal with continuous amplitude
� Digital Signal
⚫ A discrete time signal with discrete valued amplitudes

13
14
WITHOUT DSP???????
� We would not have
⚫ Digital/internet audio or video
⚫ Digital recording, CD, DVD, MP3 players
⚫ Digital cameras
⚫ Digital and cellular telephones
⚫ Digital satellite and TV
⚫ powerful tools to analyze, visualize data
⚫ Image and video editing systems
� Medical instruments would be less efficient or unable to
provide useful information for precise diagnosis (digital
ECG analyzer, digital X-ray, medical imaging
systems……..) 15
WHY PROCESSING
� To obtain better quality signal
⚫ E.g. noise filtering

Signal
Processing

Analog Digital Mixed

16
DIGITAL SIGNAL PROCESSING
� Digital signal

Multiplication

Addition Delay

DSP
17
WHY DSP
� Cost effective
� Less sensitive to component tolerances and
environmental changes
� Amenable(accept all the changes easily) to full
integration
⚫ Analog circuits: Inductors and transformers
� Accuracy can be increased by increasing the word length
(cost increases)
� Dynamic range can be increased by floating point
operations
� Time sharing of a processor is possible (time 18
multiplexing)
PROS AND CONS OF DIGITAL SIGNAL PROCESSING
� • Pros
� – Accuracy can be controlled by choosing word length
� – 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
19
� – Quantization and round-off errors
� --Expensive for small applications
APPLICATION OF DSP
1. Telecommunication: Echo cancellation in telephone networks,
Telephone dialling application, Modems, Line repeaters, Channel
multiplexing, Data encryption, Video conferencing, Cellular
phone, FAX.
2. Consumer electronics: Digital Audio,/TV, electronic music
synthersizer, educational toys, FM stereo applications, Sound
recording applications.
3. Instrumentation and control: Spectrum analysis, Digital filter,
PLL, function generator, Servo control, Robot control, process
control.
4. Image processing: Image compression, Image enhancement, Image
analysis and recognition.
5. Medicine: Medical diagnostic instrumentation such as
20
Computerized Tomography (CT), X-ray scanning,
Magnetic resonance imaging, Spectrum analysis of ECG and EEG
signals to detect the various disorders in heart and brain, Patient
monitoring.
6.Speech processing: Speech analysis methods are used in automatic
speech recognition, speaker verification and speaker
identification. Speech synthesis techniques includes conversion of
written text into speech.

7.Seismology: DSP techniques are employed in the geophysical

exploration for oil and gas, detection of underground nuclear
explosion and earthquake monitoring.
8.Military: Radar signal processing, Sonar signal processing,
Navigation, secure communications.
21
BASIC ELEMENTS OF DSP

22
 ANALOG-TO-DIGITAL AND DIGITAL-TO-ANALOG
CONVERSION(X)
Many signals of practical interest, such as speech, biological
signals, seismic signals, sonar signals, and various communications
signals such as audio and video signals are analog. To process
analog signals by digital means, it is first necessary to convert them
into digital form. This procedure is called analog-to-digital (A/D)
conversion, and the corresponding devices are called A/D
converters (ADCs).
A/D conversion is a three-step process. Sampling, Quantization
and coding
 SAMPLING
� This is the conversion of a continuous-time signal into a discrete-time signal
obtained by taking "samples" of the continuous-time signal at discrete-time
instants.
� For periodic or uniform sampling, if xa(t) is the analog input to the sampler,
the discret time output is
xa(nT) =x(n),
 where T is called the sampling interval.and its reciprocal 1/T = Fs, is
called the sampling rate (samples per second) or the sampling
frequency (hertz).
For Periodic sampling
•Relationship between the time variables t of continuous-time and n
of discrete-time signals is

•Relationship between the frequency variable F (or Ω) for analog

signals and the frequency variable f (or ω) for discrete-time signals.
To establish this relationship, consider an analog sinusoidal signal of
the form

which, when sampled periodically at a rate Fs = 1/T samples per

second, yields
 While a discrete time sinusoidal signal is expressed as

By compare equations frequency variables F and f are linearly related

Or
The frequency variable f is called as relative or normalized frequency
•The range of the frequency variable F or Ω for continuous-time sinusoi
are

While the range for discrete-time sinusoids

So the frequency of the continuous-time sinusoid when sampled at a

rate Fs = 1/T must fall in the range
 So max.value of F and Ω are

The highest frequency in a continuous time signal that can be

uniquely distinguished is Fmax.For the frequencies above
this,sampling introduces ambiguity called aliasing.
To understand aliasing
Let the sampling of a continuous-time sinusoidal signal
xa (t) = A cos (2π Fot + θ) …………(1)
with a sampling rate Fs=1/T results in a discrete-time signal
x (n) = A cos (2 π fon + θ ) …………(2)
where fo = FO/FS, is the relative frequency of the sinusoid.
If we assume that -Fs/2 ≤ Fo ≤ Fs/2, the frequency fo of x(n) is in
the range -1/2≤ fo ≤ 1/2, which is the frequency range for
discrete-time signals.
 � In this case, the relationship between Fo and fo is
one-to-one, and hence it is possible to identify (or
reconstruct) the analog signal xa(t) from the samples x(n).
� On the other hand, if the sinusoids
xa (t) = A cos (2π Fkt + θ)
Where Fk = Fo + k Fs, k = ±1, ±2....
are sampled at a rate Fs, then the frequency Fk is outside
the fundamental frequency range -Fs/2≤ F ≤ Fs/2.
And the sampled signal is
�

which is identical to the discrete-time signal in (2) obtained by

sampling (1). Thus an infinite number of continuous-time sinusoids
is represented by the same discrete-time signal.
 So if we are given the sequence x(n), an ambiguity exists as to which
continuous-time signal xa(t) these values represent. As a result
These fequencies Fk = FO + kFs for -∞ < k < ∞ (k integer)(Dec07,6)
are indistinguishable from frequency Fo after sampling and hence
they are aliases of Fo.
The relationship between the frequency variables of the
continuous-time and discrete-time signals is shown in fig.
SAMPLING THEOREM
Theorem-If the highest frequency component in a signal is
fmax,then the signal should be sampled at the rate of
atleast 2 fmax for the samples to describe the signal
completely.

Fs ≥2 fmax

Sampling at less than the rate specified by the sampling

theorem leads to a folding over or 'aliasing' of 'image'
frequencies into the desired frequency band so that the
original signal cannot be recovered if we were to convert
the sampled data back to analog.
� Figure shows the sampling process, which can be regarded as the
multiplication of the analog signal x(t) by a sampling function,
p(t).

� p(t) consists of pulses of unit amplitudes, width dt and period T.

The spectra of x(t), p(t), and their product are shown in Figure

� X'(f) is the convolution of X(f) and P(f) because multiplication in

the time domain is equivalent to convolution in the frequency
domain.
• In sampled waveform the spectrum is the same as the original
analog spectrum, but repeats at multiples of the sampling
frequency, Fs. The higher order components which are centered on
the multiples of Fs , are referred to as image frequencies.
• If the sampling frequency, Fs is not sufficiently high the image
frequencies centered on Fs will fold over or alias into the base
band frequencies . In this case, the information of the desired
signal is indistinguishable from its image in the foldover region.

The overlap or aliasing occurs about the point FN, that is half the
sampling frequency point. This frequency point is variously called
the folding frequency, Nyquist frequency, and so on.
QUANTIZATION
� In quantization and encoding, each analog sample is
assigned to one of 2B values where B is the number of ADC
bits. Quantization introduces an error which cannot be
removed. The level of the error is the function of the number
of bits of the ADC, being approximately equal to one-half an
LSB .
� For example, a 12-bit ADC with an input voltage range of ±10
V will have an LSB of 20/212 V, that is 4.9 mV, and a
quantization error 2.45 mV.
� For an ADC with B binary digits ,the number of quantization
levels is 2B and interval between the levels, that is the
quantization step size, q is given by

where Vfs, is the full-scale range of the ADC with bipolar signal
inputs. For example a sine wave input of amplitude A ,the
quantization step size becomes q = 2A/2B
 The maximum quantization error, for the case where the values
are rounded up or down, is ±q/2.
The quantization error for each sample, e, is normally assumed
to be random and uniformly distributed in the interval ±q/2
with zero mean. In this case, the quantization noise power, or
variance, is given by

For the sine wave input, the average signal power is A2/2. The
signal-to-quantization noise power ratio (SQNR), in decibels, is
This is a theoretical maximumvalue.In practice,achievable SQNR is
less than this value. However, the SQNR increases with the
number of bits, B .
The digital samples, x(n), which in many cases are in a binary
form, are next encoded into a form suitable for further
manipulation. Encoding means assigning discrete codes to the
quantized samples.
NUMERICALS

37
NUMERICALS

38
NUMERICALS

39
NUMERICALS

40
41
42
43
44
45
46
47
SIGNAL REPRESENTATION
Notations:
⚫ Analog signal:
⚫ Digital Signal:
Signal representation

Units

48
RELATIONS
� What is the relation between F and f

� What is the relation between and

� What is the relation between t and n

49
FREQUENCY RANGES
� What is the frequency range of and

� What is the frequency range of and

50
FREQUENCY MAPPING

51
RELATIONS AMONG FREQUENCY VARIABLES

52