Description: Credit 4U/5G. Computer-aided design of digital filters; quantization and round-off effects; FFT algorithms; number-theoretic algorithms; multirate signal processing; DSP architectures and programming.

Prerequisite: ECE 317

Topics:

1. Computer-Aided Design of Digital Filters
   IIR Filter Design
   Minimum Mean-Square Error
   Deczky's Method
   Group Delay Equalization
   FIR Filter Design
   Frequency Sampling
   Equiripple Filters
   Parks-McClellan Algorithm
   Linear-Phase FIR Filters
   FIR Differentiators
2. Quantization and Round-Off Effects
   Quantization of Analog Sampling
   Quantization of IIR Filters
   Quantization of FIR Filters
   Round-Off Effects of IIR Filters
   Round-Off Effects of FIR Filters
   Limit Cycle Oscillations
   Overflow Scaling
3. Fast Fourier Transform (FFT) Algorithms
   Radix-2 FFT Algorithm
   Goertzel Algorithm
   Chirp-z Transform Algorithm
   Quantization and Round-Off Effects of DFT
   Quantization and Round-Off Effects of FFT
4. Number-Theoretic Algorithms
   Number Theory
   Lagranage Interpolation
   Chinese Remainder Theorem
   Fast Convolution Algorithms
   Cook-Toom Algorithm
   Winograd Algorith
5. Hilbert Transform
   Sufficiency Theorems: Casual Sequences
   Sufficiency Theorems: Finite-Length Sequences
   Magnitude and Phase
   Fourier Transform
   Complex Sequences
6. Multirate Signal Procesing
   Decimation
   Interpolation
   Sampling Rate Conversion
   Phase Shifters
   System Interfacing
   Narrowband Lowpass Filters
   Filter Banks
   Subband Coding
   Quadrature Mirror Filters
   Transmultimplexers
   Oversampling Converson
7. DSP Microprocessors
   Architectures
   Instruction Sets
   Programming Applications

:Lab Topics:

1. Architecture and instruction set of a DSP microprocessor
2. Assembly language programming of a DSP, software
   environment
3. Four experiments in assembly language programming a DSP
   for digital signal processing
4. Implementation of equiripple FIR filters
5. Round-off and quantization effects
6. Limit cycles
7. Implementating a Hilbert transform