Department of Electronics and Communication
Engineering
Digital Signal Processing
& Applications
Subject Code - ECN-312
Course Introduction
1.
Course Content
Review of discrete-time signal and system analysis; Advantages and typical applications
of DSP.
2. Sampling and Quantization: Sampling and discrete-time processing of continuous time
signals, Sampling of low-pass and band-pass signals; Uniform and non-uniform
quantization, Lloyd-Max algorithm, Log-companding, A-law, µ-law; Adaptive
quantization and prediction
3. Orthogonal transforms: Properties and applications of DFT, implementing linear time
invariant systems using DFT, circular convolution, linear convolution using DFT; Fast
Fourier Transform, FFT algorithms: Decimation in time, decimation in frequency;
Goertzel algorithm; Application of transform in speech, audio, image and video coding,
Karhunen-Loeve Transform, DCT, JPEG and MPEG coding standards.
4. Digital Filter design techniques: IIR and FIR filters, filter design specifications; Design of
digital IIR filters: Impulse invariant, and bilinear transformation techniques for
Butterworth and Chebyshev filters; Design of FIR filters: Windowing, frequency sampling
filter design, optimum approximations of FIR filters
5. Multi-rate Signal Processing: Fundamentals of multirate systems, Decimation and
interpolation, application of Multirate DSP in sampling rate conversion; Filter banks;
Polyphase structures; Quadrature-mirror filter bank; Wavelet transform and its relation to
multi-rate filter banks; applications to speech and audio coding.
6. Basic concept of Adaptive Digital Signal Processing: Adaptive Wiener filter and LMS
algorithm; Applications of adaptive filtering to echo cancellation and equalization
Marks Distribution
• Assignments: 15 %.
• Mid-term Exam: 35%
• End Term Exam: 50%
S. No. Name of Books / Authors/ Publishers Year
1. Proakis, J.G, “Digital Signal Processing: Principle, Algorithms and 2007
Applications”, 4th Ed., Pearson.
2. Oppenheim, A.V. and Schafer, R.W, “Discrete-Time Signal Processing”, 3rd 2009
Ed., Pearson.
3. Porat, B. “A Course in Digital Signal Processing” , 1st Ed., John Wiley & Sons. 1996
4. Mitra, S.K., “Digital Signal Processing A Computer-Based Approach”, 4th Ed., 2010
McGraw-Hill.
5. Weltch , T.B., Wright, C.H.G. and Morrow, G.M., “Real-Time Digital Signal 2012
Processing from MATLAB to C with the TMS320C6x DSPs.”, 2nd Ed., CRC
Press.
Signal, system and Signal Processing
Signal: any quantity that varies with respect to
time, space or any other variable or variables.
Examples: 2
,
S1 (t) =t S2 (x, y) =x + y 2
Speech signal: N
S =å Ai (t)sin(2p Fi (t)t + q i (t))
i=1
Function
Natural of timeElectrocardiogram (ECG): Heart
signals:
activity.
Electroencephalogram (EEG): Brain Activity.
Signal, system and Signal Processing
System: Device that performs an operation on a
signal.
Example: Filter to remove noise and
interference from a system is a system.
“ When signal is passed through a system, signal
is processed”
Digital As Compared to Analog processing
Advantages:
• Re-Configurability
• Accuracy
• Portability
• In some cases Cost-effective.
Limitations:
Speed of A/D and D/A converters
Continuous time vs. discrete time signals
Continuous Signals:
Defined for any continuous time interval [a, b] where
(a, b) (, )
Discrete time Signals:
Defined only at discrete time intervals.
Theoretically, Intervals might be equidistant or not!
Practically, intervals are predefined by sampling
frequency.
Sampling: Continuous time to
discrete time conversion
If sampling frequency is 1 KHz, what would be
sampling time between the two samples?
One Millisecond!
Continuous valued Vs. Discrete
Valued signals
Continuous Valued Signal:
*Can take any value finite or infinite.
Discrete Valued Signal:
*Can take only few selected values.
Process of continuous to digital
value conversion is known as : Quantization!!
Digital signals have discrete time and
discrete values!!
Continuous time Sinusoidal Signal
Radians/second Phase in radians
xa (t + T ) =xa
1
T=
F
Theoretically, T
can go to infinity
for F=0.
Continuous time complex exponential
Signal
j 2 p Ft+q
xa (t) =Ae
By Euler’s identity: e ±jf =cos(f ) + j sin(f)
Theoretically frequency is
positive quantity,
however negative
frequency is introduced
for mathematical
convenience!
xa (t) =A cos(W t + f )
=A(e j (W t+f ) + e- j (W t+f ) )
Discrete time Sinusoidal Signal
No. of samples
Discrete time Sinusoidal Signal
Proof:
Discrete time Sinusoidal Signal
Discrete-time sinusoids whose angular frequencies are separated by an
integer multiple of 2π are identical.
Proof:
1
Any sinusoid resulting from or f is identical to
2
the sinusoid resulting from .
We call sinusoid with an alias of sinusoid with .
We regard all sinusoid within the range or
1 1 unique.
f
2 2
Discrete time Sinusoidal Signal
Above is case for cosine signal, what would be the
case for sinusoidal signal?
What will happen if varies from to 2 .
Sampling of Analog Signals
Discrete Signal Analog Signal
Sampling
Frequency
Sampling of Analog Signals
Relation between analog frequency or (F) and
ω or (f ).
Sampling of Analog Signals
Inserting new values of ω or (f):
Ranges:
Analog frequency or (F)
Highest frequency in
discrete domain is
Discrete frequency ω or (f ).
Example
Consider two waveforms!
Fs=40 Hz, write down discrete time signal
expression!
Example
(a) What is minimum sampling frequency required for x2.
(b) .
(c) .
4
cos( n)
3
Example
(a)What is Nyquist rate for following signal:
Answer:
If sampled at Nyquist rate, write down expression for discrete signal
(b)For
Answer:
(c) If
Answer:
Sampling rate should be greater than Nyquist
Criterion!
Practice Problems from Proakis
1.1, 1.2,
1.3, 1.9,
1.4, 1.10
1.5, 1.11,
1.6, 1.12
1.7, 1.15
1.8,
