Advanced Digital Signal Processing Techniques

Course: Advanced Digital Signal Processing Techniques

Code: 3ФЕИТ05011

ECTS points: 6 ECTS

Number of classes per week: 3+0+0+3

Lecturer: Prof. Dr. Dimitar Taskovski

Course Goals (acquired competencies): Successful completion of the course will provide students with: familiarization with all relevant discrete transformations and their features that enable processing in the domain of transformation; ability to recognize situations in which a particular transformation shows better performance than another; introduction to fast algorithms for the implementation of discrete transformations. Also, with a successful completion of the course, the student will demonstrate knowledge and understanding of: the basics of multi-rate processing, the basic idea of ​​the filter banks, the design of filter banks, and the link between the filter banks and the wavelet transformation.

Course Syllabus: A brief review of signals. Definition and basic properties of discrete transformations. Orthogonal discrete transformations: DFT, Hartley’s (DHT), Karhunen-Loeve (KLT), Cosine (DCT), Lapped (LOT), Wavelet (DWT), Walsh-Hadamard (WHT). Fast Algorithms: Concept and selected examples. Some typical DSP applications in the transform domain: filtering, spectrum estimation, coding, adaptive filtering, multi-rate processing. Introduction to the basic theory of multi-rate processing: decimation, interpolation and sample-rate conversion. Two-channel filter banks: QMF filter bank, filter bank with perfect reconstruction, paranitary, biorthogonal, and linear phase filter banks. Lattice structure of filter banks with perfect reconstruction. Link of filter banks with a wavelet transform. Lifting implementation of the wavelet transform. Application of wavelet transform for analysis and signal compression.

Literature: