Algorithms of Signal Processing

The course is a substitute for:

Algorithms of Signal Processing (31ASI)

Lecturer:

Tutor:

Supervisor:

Department of Circuit Theory

Synopsis:

The aim is to give the principles of algorithms used in modern digital systems. They are studied: filtration in time and frequency domain, structures of systems, parametric methods (AR, MA, ARMA), linear prediction (LPC), LPC characteristics, DFT and LPC based spectral analysis. The exercises are focused on solution of project in a chosen topic - usually algorithms implementation and simulation in MATLAB. Required knowledge of digital signals and systems theory.

Requirements:

The elaboration and representation of one selected problem is the condition for the credit.

Syllabus of lectures:

1. Short-term and long-term signal characteristics in time and frequency domain

2. Overlap-add (OLA) method of signal synthesis

3. Filtration in time and frequency domain, lattice structure

4. Parametric methods (AR, MA, ARMA), linear prediction (LPC)

5. LPC characteristics, DFT and LPC based spectral analysis

6. Application of spectral analysis for speech and industrial processing

7. Cepstral analysis, echo detection and suppression

8. Application of cepstral analysis for signal detection in noise

9. Robust parameterization of signals in speech processing and prediction

10. Correlation analysis, delay compensation, noise cancellation

11. Coherence analysis, signal detection and separation

12. Application of adaptive filtering in spectral analysis and noise cancellation

13. Time-frequency analysis, nonlinear frequency transformations

14. Orthogonal transformations, data compression

Syllabus of tutorials:

1. Estimation of short-time and long-time signal characteristics

2. Implementation of method OLA (OverLap-and-Add)

3. Filtration in the frequency domain, spectral subtraction

4. DFT filter banks with linear and non-linear frequency axis

5. LPC - normal equations, autocorrelation method

6. Covariance method of normal equation solution

7. Fast algorithms for AR parameters computation, lattice structures in AR modelling

8. Real and complex cepstrum, aplication for smoothed spectral estimation

9. Implementation of cepstral analysis, signal reconstruction

10. Spectral and cepstral distances, signal detection in noise

11. Cepstral liftering and applications

12. Coherence analysis and evaluation of coherence function in MATLAB

13. Appliaction of discrete cosine transform

14. Semester project presentation, credit

Study Objective:

Study materials:

1. Sovka, P., Pollák, P.: Selected methods of digital signal processing (in Czech language). Ediční středisko ČVUT Praha,2003

2. Uhlíř, J., Sovka, P.: Digital signal processing (in Czech language). Ediční středisko ČVUT Praha,1995. Monografie ČVUT FEL

3. Davídek, V., Sovka, P.: Digital signal processing and implementation (in Czech language). Ediční středisko ČVUT Praha,2001

4. Oppenheim, A.V., Shafer, R.W.: Discrete-Time Signal Processing. Prentice-Hall, Inc., New Jersey, 1998

Note:

Further information:

No time-table has been prepared for this course

The course is a part of the following study plans:

• Electronics - Electronic Systems- structured studies (compulsory course)