Study Guide 2015-2016

SGN-23006 Advanced Filter Design, 5 cr

Additional information

Suitable for postgraduate studies

Person responsible

Tapio Saramäki

-->

Requirements

Final examination and 2 assignments out of 3.

Learning Outcomes

After finalizing the course with a good grade, the student will be really aware of, among others, on the following aspects: - What is a digital filter and how to analyze its performance using its transfer function and frequency response as well as its magnitude, phase, group delay, and phase delay responses? - Advantages and drawbacks when comparing infinite-impulse response (IIR) and finite-impulse response (FIR) filters with each other. – Four types of linear-phase FIR filters and their use in practice. - Design and implementation of FIR digital filters using both traditional approaches and more sophisticated approaches leading to efficient implementations - Design and implementation of IIR digital filters using both traditional approaches and more sophisticated approaches leading to efficient implementations - Various finite word-length effects in the implementation digital filters in both theory and practice: (a) output noise due to multiplication round-off errors; (b) filter scaling; (c) finite word length effects on the variations in the filter coefficients as well as their impact on various kinds of oscillations in IIR filters. - When and why is it beneficial to utilize multirate filtering when generating DSP algorithms? - Design and implementation of various kinds of decimators or interpolators for various practical applications. - Basic characteristics of various kinds of Nth-band FIR and IIR decimators and interpolators as well as their optimization and use in practical applications - How to use the modified Farrow structure for interpolation by an arbitrary factor and the transposed modified Farrow structure for decimation by an arbitrary factor. - What are multirate filter banks, transmultiplexers, and discrete-time wavelet filter banks and how to synthesize them for practical applications?

Content

Content Core content Complementary knowledge Specialist knowledge
1. What is a digital filter and how to analyse its performance with the aid of various responses such as its transfer function and frequency response as well as its magnitude, phase, group delay, and phase delay responses? Various structures to implement the very same transfer function.  Introductory filtering examples, the roles of the poles and zeros in providing their contributions to various responses of linear-phase FIR and IIR filters, and the significant differences between IIR and linear-phase FIR filters for shaping the passband response of the filter, namely, the poles of IIR filters accomplish efficiently this shaping, whereas FIR filters have only zeros to perform this duty.  During the lectures, some extra information not included in the lecture notes is given. 
2. Filter synthesis procedure in nutshell, including typical criteria for filter responses along with examples, and illustrative descriptions on the use of minimax, least-squared, and maximally-flat approximation criteria.  The lecture notes provide a review on those various structures, which are commonly used for implementing digital filters, to make students aware of the terminologies, which are used about these structures.   
3. Filter synthesis procedure in nutshell, including typical criteria for filter responses along with examples, and illustrative descriptions on the use of minimax, least-squared, and maximally-flat approximation criteria.  The characteristics of the four types of linear-phase FIR filters and their use in practical application.   
4. Design and implementation of FIR digital filters using both traditional approaches and more sophiscated approaches leading to efficient implementations.  The characteristics of the four types of linear-phase FIR filters and their use in practical application. How to design minimum-phase FIR filters?  The lecture notes review various alternatives of synthesizing computationally-efficient linear-phase FIR filters. 
5. Design and implementation of IIR digital filters using both traditional approaches and more sophiscated approaches leading to efficient implementations.  An efficient Remez-type algorithm developed by the lecturer for designing classical IIR filters and their generalizations is described.   
6. Finite word-length effects in digital filters when using the fixed-point two's complement arithmetic; Attractive properties of the two's complement arithmetic; The commonly used model for estimating the output noise due to the multiplication roundoff errors; Scaling of the cascaded-form IIR filters by using three commonly used scaling norms and the resulting trade-offs between probabilities of the overflows and resulting output noises; Finite word length effects on the variations in the filter coefficients as well as their impact on various kinds of oscillations in IIR filters.  Finite word-length effects in practice: (a) How to easily quantize the coefficient values of direct-form linear-phase FIR filters?; (b) How to easily quantize the coefficient values of IIR filters, which are implemented as a cascade of second- and first-order blocks?; (c) How to easily quantize the coefficient values of IIR filters, which are implemented as a parallel connection of two allpass filters?; (d) The validity of the noise model, which is commonly used to estimate the output noise due to the multiplication roundoff errors.    
7. Design and implementation of efficient decimators and interpolators - A comprehensive review     
8. Polynomial-based interpolation for signal processing and communications applications     
9. Design and implementation of multirate filter banks including conventional frequency-selective banks and discrete-time wavelet banks - A comprehensive review  More information on multirate filter banks can be found in T. Saramäki and Robert Bregovic', Multirate Systems and Filter Banks," Chapter II in Multirate Systems: Design & Applications, edited by Gordana Jovanovic-Dolocek, Idea Group Publishing, 2002, pp. 27-85.   

Instructions for students on how to achieve the learning outcomes

Course is graded on the basis of answers to exam questions. Very good grade is obtained when exam questions are correctly answered and 2 of 3 homeworks are accepted. Course acceptance threshold is approx. half of the maximum exam points. The third homework is a volunteer work and is prized with increasing the exam result by one grade provided that the threshold is passed.

Assessment scale:

Numerical evaluation scale (1-5) will be used on the course

Study material

Type Name Author ISBN URL Additional information Examination material
Other literature   Design of computationally efficient FIR fillters using periodic subfilters as building blocks in The Circuits and Filters Handbook, Second Edition, edited by W.-K. Chen, CRC Press, Inc., 2002, pp. 2654-2677.   Tapio Saramäki       Only some parts are included in the study material.   Yes   
Other literature   Finite impulse response filter design, Chapter 4 in Handbook for Digital Signal Processing, edited by S. K. Mitra and J. F. Kaiser, John Wiley and Sons, New York, 1993, pp. 155-277.   Tapio Saramäki       Only some parts are included in the study material.   Yes   
Other literature   Multirate Systems and Filter Banks, Chapter II in Multirate Systems: Design & Applications, edited by Gordana Jovanovic-Dolocek, Idea Group Publishing, 2002, pp. 27-85.   Tapio Saramäki and Robert Bregovic’       Only some parts are included in the study material.   Yes   
Summary of lectures   Advanced Filter Design   Tapio Saramäki         Yes   

Prerequisites

Course Mandatory/Advisable Description
SGN-11000 Signaalinkäsittelyn perusteet Mandatory    
SGN-11006 Basic Course in Signal Processing Mandatory    
SGN-21006 Advanced Signal Processing Mandatory    



Correspondence of content

Course Corresponds course  Description 
SGN-23006 Advanced Filter Design, 5 cr SGN-2016 Digital Linear Filtering I, 5 cr +
SGN-2056 Digital Linear Filtering II, 4 cr +
SGN-2106 Multirate Signal Processing, 6 cr +
SGN-2156 System Level DSP Algorithms, 4 cr
 

Last modified 19.01.2015