CS-590.74

Introduction to Speech Science and Technology

Postgraduate

Great advances in machine learning have led to the employment of “data-hungry” methods with the goal of improving human-machine communication. Databases should be treated not only quantitatively but also qualitatively so as to contribute to the development of modern AI techniques. This course focuses on the qualitative characteristics of speech and of speech databases. The aim is to acquaint Computer Science students with Voice and Speech in a broader sense and induct them into Speech Science.  
Speech Science is the experimental study of speech communication, examining speech production and perception as well as speech signal analysis and processing. It originates from Phonetics, the branch of Linguistics that studies speech sounds, but utilizes empirical methods and techniques adopted from other sciences, such as physics, physiology, psychology, in order to define the physical and physiological dimensions of speech sounds as well as their perceptual characteristics. This additional information on speech sounds can be used in Speech Technology, e.g., synthetic speech, speech recognition, text-to-speech applications, etc.
            The course is an introduction to the basics of acoustic and auditory phonetics and to the study of speech production and perception with experimental techniques, such as recording, spectrographic analysis, electroglottography, stroboscopy, etc. We will address issues relating to articulation, hearing, production and perception of speech, as well as recording, analysis and annotation of speech data.
            The course is offered to Computer Science students as well as to students of Linguistics/Philology, Medicine, Physics, Education and other departments, wishing to acqaint themselves with the mechanism of human speech and hearing. Interdisciplinary cooperation among students from different departments and educational backgrounds is particularly encouraged.
 
 
Goals

1. To acquire knowledge about the structure and function of the human speech and hearing mechanism
2. To distinguish sounds based on voice, manner and place of articulation
3. To understand and appreciate the connection between speech production and perception
4. To learn good practice in making digital speech recordings for research purposes
5. To acquire skills in the spectrographic analysis of speech
6. To appreciate the difficulty but also usefulness of speech segmentation
7. To acquire skills in manual and automatic speech annotation on various linguistic levels
8. To acquire knowledge about the prosodic and paralinguistic features of speech
9. To get acquainted with various experimental techniques of recording voice and speech data
10. To appreciate the significance and value of integrating Speech Science knowledge in Speech Technology

4

CS-100, CS-112, CS-215 or with instructors’ permission

https://elearn.uoc.gr/