Abstract :
In a typical digital filter or digital signal processing class, students are asked to design a digital filter of a specified type and fixed order. The design may or may not be satisfactory for any application but almost always in industry, the best design is required that would satisfy certain desired specifications which are usually predefined on the basis of system considerations. Unfortunately, more often than not the topic of designing the best filter for the application is not treated in the classroom. In this tutorial, design methodologies will be described that would yield FIR as well as IIR filters that would satisfy prescribed specifications. Two types of designs will be explored. Closed-form methods based on some classical techniques and iterative methods based on optimization. The tutorial will draw heavily from the proposer´s past teaching experience and research results such as the design of elliptic IIR digital filters, optimization-based techniques such as the design of equalized IIR filters using quasi-Newton methods, and the use of enhanced Remez methods for the design of FIR filters that would satisfy precise predefined specifications. Participants will receive a free license of Dr. Antoniou´s DSP software D-Filter (see http://www.d-filter.ece.uvic.ca for details.)
