Digital Processing of Speech Signals

The course is a substitute for:

Digital Processing of Speech Signals (X31DZR)

Lecturer:

Tutor:

Supervisor:

Department of Circuit Theory

Synopsis:

Articulation and perception of speech. Discretization and quantization, data compression, quality measures. Statistical parameters of speech signal. Linear prediction, cepstral parameters and their application. Text to speech synthesis. Fundamentals of automatic speech recognition (ASR). ASR based on Hidden Markov chains. Application of Artificial Neural Nets in ASR. Dialogue systems, large vocabulary systems an continuous speech recognition. Voice controlled information systems.

Requirements:

Syllabus of lectures:

1. Articulation model, speech signal perception

2. Speech signal and its transmission

3. Speech signal compression algorithms

4. Bit rate in transmission systems - quality measures

5. Statistics of digital signal of speech

6. Parametric methods of speech signal coding

7. Algorithms of low bit rate coding

8. Enhancement of noisy speech signal

9. Phonetic description of speech fundamentals, text to speech synthesis (TTS)

10. Automatic speech recognition systems (ASR)

11. Hidden Markov models (HMMs) in ASR

12. Artificial neural nets (ANNs) in speech recognition

13. Continuous speech recognition - ideas and algorithms

14. Examples of TTS and ASR systems commercially offered

Syllabus of tutorials:

1. PC lab - MATLAB - quantization and sampling of speech signal, quantization noise

2. PC lab - MATLAB - sampled speech signal statistics

3. PC lab - MATLAB - speech signal compression in the time domain (DPCM, ADPCM)

4. PC lab - MATLAB - spectral analysis of segmented speech signal, phoneme characteristics

5. PC lab - MATLAB - speech signal parametrization (LPC, cepstrum)

6. PC lab - MATLAB - speech signal parametrization (LPC, cepstrum)

7. PC lab - MATLAB - basic methods of speech recognition, spectral distance, (DTW)

8. Class - principles of Hidden Markov Models and application, Hidden Markov Toolkit

9. PC lab - HTK experiments, design of HMM isolated word recogniser

10. PC lab - individual project (IP), implementation of speech processing algorithm

11. PC lab - IP realisation

12. PC lab - IP realisation

13. PC lab - IP presentation, student group discussion

14. Credit - concluding discussion and notes

Study Objective:

Study materials:

1. Deller, J. R., Proakis, J. G., Hansen, J.H.L.: Discrete Time Processing of Speech Signals. New York, Macmillan, 1993

Note:

Further information:

No time-table has been prepared for this course

The course is a part of the following study plans:

• Computer Science and Engineering (elective course)