Advanced Digital Signal Processing
Selected topics in digital signal processing such as digital speech processing, multidimensional digital signal processing, spectrum estimation, space-filling curves, and error analysis.This course will examine a number of advanced topics and applications in one-dimensional digital signal processing, with emphasis on optimal signal processing techniques. Topics will include modern spectral estimation, linear prediction, short-time Fourier analysis, adaptive filtering, plus selected topics in array processing and homomorphic signal processing, with applications in speech and music processing.
At the end of this course, students will be able to:
1. Analyze multirate DSP systems.
2. Determine coefficients for perfect reproduction filter banks and wavelets.
3. Choose parameters to take a wavelet transform, and interpret and process the result.
On completing this course, students are expected to:
Analyse and evaluate the properties of LTI systems in terms of z-transforms.
Understand the sampling theorem and perform sampling on continuous-time signals by applying advanced knowledge of sampling theory (i.e. aliasing, quantisation errors, pre-filtering).
Apply the concepts of all-pass and minimum-phase systems to analyse LTI systems and address complex design problems.
Evaluate design problems related to frequency selective processing and design FIR/IIR filters.
Construct systems for spectral estimation of real signals by applying advanced knowledge of Fourier techniques.
Judge implementation aspects of modern DSP algorithms.
Apply the relevant theoretical knowledge to design and analyse a practical discrete-time signal system, such as a radar, image, speech, audio, bio-medical or wireless system.