227-0677-00L  Speech Processing I

SemesterAutumn Semester 2014
LecturersB. Pfister
Periodicityyearly recurring course
Language of instructionGerman
Comment"Speech Processing I" takes place for the last time in fall 2014.

AbstractFundamentals of speech signal processing and introduction to text-to-speech synthesis and speech recognition.
ObjectiveKnowledge of the basics in speech processing. Acquisition of practical experience in this field. Comprehension of the fundamental problems of text-to-speech synthesis and speech recognition and selected solutions.
ContentAnalysis, representation and properties of speech signals: Time and frequency domain representations, quasi-stationarity, formants, pitch, short-time analysis, spectrum, autocorrelation, linear prediction, homomorphic analysis.
Fundamental problems of speech synthesis: Relations between text and speech; methods of speech production; prosody control.
Fundamental problems of speech recognition: Variability of speech signals, speech features for speech recognition, pattern matching (distance measures, dynamic programming), and introduction to statistical speech recognition with hidden Markov models.
Lecture notesThe following textbook will be used: "Sprachverarbeitung - Grundlagen und Methoden der Sprachsynthese und Spracherkennung", B. Pfister und T. Kaufmann, Springer Verlag, ISBN: 978-3-540-75909-6
Prerequisites / NoticePrerequisites:
Knowledge in digital signal processing and digital filters is helpful.