Digital Filter Design (DFD)
Dr. Anurag Singh (anurag@iiitnr.edu.in)
Syllabus
Discrete signal analysis: Review of discrete time signals, systems , Sampling and transforms: DTFT, Laplace and Z-
transform.
8
Discrete Fourier Transform (DFT): Need of DFT and its evolution from DTFT, DFT properties, fast Fourier transform
(FFT) algorithms, Circular convolution, linear convolution using DFT.
LTI systems and their frequency domain analysis: LTI systems , LTI systems as Frequency selective filters,
difference between digital and analog filters , Ideal filter characteristics, lowpass, highpass, bandpass and bandstop filters, 6
Paley-Wiener criterion, digital resonators, notch filters, comb filters, all-pass filters, inverse systems, minimum phase,
maximum phase and mixed phase systems.
Design of digital filters:
Design of FIR filters- Design of FIR filters using windows, frequency sampling, Remez algorithm and least 5
mean square error methods;
Design of IIR filters- Design of IIR filters using impulse invariance, bilinear transformation and frequency
transformations. 5
Filter Applications 4
Laboratory Work:
Generation and representation of different types of signals with multiple frequencies, Frequency domain analysis of signals through
DTFT and DFT, Convolution process (Linear and circular), Implementation of the different types of digital IIR and FIR Filters, Analyze
the effects of filters with varying parameters, Problems based on design and implementation of filters for practical world problems.
Books
Text Books:
1. J.G. Proakis, D.G. Manolakis and D. Sharma, Digital Signal Processing, Pearson, 3 rd edition,
(2013).
2. S. Salivahnan, “ Digital signal Processing”, Tata Mcgraw Hill
3. A.V. Oppenheim, and R.W. Schafer, Discrete-Time Signal Processing, Pearson, (2002).
Reference Books:
1. S. K. Mitra, Digital Signal Processing: A computer based approach, Tata McGraw Hill, 2nd
edition.
2. B. P. Lathi,"Signal Processing and Linear Systems", Oxford University Press, 1998
3. M. Li Tan, Digital Signal Processing: Fundamentals and Applications, Elsevier, (2008).
4. Tamal Bose, Digital Signal and Image Processing, Wielly, (2004).
Evaluation Scheme
End Term: 50
Quiz: 15
Course Project (DSP Applications): 25
Continuous Lab/class assessment: 10
Digital Filters
What are filters:
Anything that has ability to separate two or more things based on their physical
characteristics.
Examples ?
Filters in context of signal processing: called frequency selective filters
Why Frequency Selective?
What are Digital Filters?
Filters that handle digital data for any signal processing task, e.g. noise canceller in your
headphone
Digital Filters in Digital Signal Processing
(DSP)
Digital filters are the integral part of every DSP operations, be it
voice communication over phone,
beautifying your pictures for social media,
playing music or recording your voice.
In every DSP operation, digital filters are used for
Removing noise
Recovering your desired signal
Discarding unwanted signal components/frequencies, etc.
So, for studying digital filters, understanding the related DSP is a must
Definition of Digital Signal Processing
Processing of the signals through digital means, either in software or hardware
Evolution of DSP
DSP evolved with the faster development of digital computers and the associated digital hardware
Advantages of DSP/Digital domain processing
DSP gives faster, reliable and more economical way of doing many signals processing tasks
More flexibility: DSP hardware allows programmable operations, hence any task can be performed though change
in the software
High precision: More degree of precision compared to analog processing
DSP allows more portability of data, low cost storage, high reproducibility, etc.
Disadvantages
Introduces distortion: Due to digital data formation through Sampling/Quantization
Slower process: Due to Sampling, its slow for high bandwidth signals
Applications-DSP/Digital Filters are Everywhere
Sound applications
Compression, enhancement, special effects, synthesis, recognition, echo cancellation,…
Cell Phones, MP3 Players, Movies, Dictation, Text-to-speech,…
Communication
Modulation, coding, detection, equalization, echo cancellation,…
Cell Phones, Satellite Receiver,…
Automotive
GPS, Active Noise Cancellation,…
Medical
Magnetic Resonance, Tomography, Electrocardiogram,…
Military
Radar, Sonar, Space photographs, remote sensing,…
Image and Video Applications
DVD, JPEG, Movie special effects, video conferencing,…
Mechanical
Motor control, process control,…
Standard DSP Steps
DSP Example- Voice recording/playing
DSP-Example voice amplification
Review of Signals and Systems
Types/Classification of signals- unit impulse/delta signal, Causal/noncausal,
Type of Systems- Linear/nonlinear, Time invariant, causal/non-causal, etc..
LTI systems and their response analysis (in time-domain, z-domain, Laplace
domain)
Sampling theory
Fourier analysis- FS, FT, DTFS, DTFT DFT
Sampling
Sampling Rate
fs= 1/Ts
1 2 2
Ts (t ) t n T T
n
s
Ts
cos( s t ) cos(2 s t ) . . .
Ts
s
1 n
Nyquist Rate G ( )
Ts
G(w n )
n
s
fs > 2B
fs = 2B
g (t ) g (t ) Ts (t )
Aliasing Examples
15
Aliasing
The black is the original 16
signal
The red dots are the samples
Obviously, the red is a poor
representation of the signal
Sampling frequency of a signal play
an important role in deciding various
filter design parameters
How to avoid aliasing ?
17
Eliminate high frequencies before sampling
Apply a low-pass filter
A low-pass filter allows low frequencies through, but attenuates (or reduces) high frequencies
Sample at higher rate
Transforms: Laplace and Z-transform
Limitation of Fourier analysis: It can't be used for frequency domain analysis of every type
of signals and systems.
Examples?
Laplace and Z-transforms are the frequency domain analysis tools for unstable signals and
systems- An extension of Fourier analysis
Laplace is for analog/continuous signals & systems-> An extension of Continuous Fourier
Transform (FT)
Z-transform for digital signals & systems-> An extension of discrete time FT (DTFT)
Since filers are digital devices, z-transform is used for the frequency domain analysis of
filters
Concept of frequency in analog and digital domain
Two types of frequency
1) Analog- cyclic (Cycles/sec) and radian/angular frequency (Radian/sec)
2) Discrete/digital frequency (Cycles or radian/sample)
• Discrete/digital frequency, = radian/sample= T
=or
[]
In terms of cyclic frequency= cycles/sample
= or [0,1] also called Normalized Frequency
Frequency relations
