FI:P132 Acoustic Processing of Speech - Course Information
P132 Current Methods for Acoustic Processing of Speech
Faculty of InformaticsAutumn 2000
- Extent and Intensity
- 15/0. 2 credit(s) (plus extra credits for completion). Recommended Type of Completion: zk (examination). Other types of completion: k (colloquium).
- Teacher(s)
- Hynek Heřmanský (lecturer), prof. PhDr. Karel Pala, CSc. (deputy)
- Guaranteed by
- prof. PhDr. Karel Pala, CSc.
Department of Machine Learning and Data Processing – Faculty of Informatics - Course Enrolment Limitations
- The course is also offered to the students of the fields other than those the course is directly associated with.
- fields of study / plans the course is directly associated with
- Informatics (programme FI, B-IN)
- Informatics (programme FI, M-IN)
- Upper Secondary School Teacher Training in Informatics (programme FI, M-IN)
- Upper Secondary School Teacher Training in Informatics (programme FI, M-SS)
- Information Technology (programme FI, B-IN)
- Syllabus
- Speech signal carries information from many sources and not all information sources are relevant for the given task. The goal of acoustic analysis of speech signal is to extract the relevant information while supressing the rest. Therefore, the acoustic analysis module is the front-end for systems for speech and speaker recognition as well as it is an integral part of speech coding systems. The course first discusses purpose of the acoustic analysis, historical techniques of the analysis, and conventional digital processing techniqes such as the short-term Fourier analysis, the cepstral analysis, and the linear predictive analysis. However, the emphasis of the course is on currently evolving techniques which employ knowledge of \ properties of human auditory and speech perception. Reasons for such auditory-like approaches are discussed and some of relevant properties of human auditory system reviewed on the level necessary for understanding their engineering emulation in auditory-like analysis of speech. The main body of the course describes the underlying philosophy and implementations of techniques such as Perceptual Linear Prediction (PLP), RelAtive SpectTrAl (RASTA) processing, data-guided discriminant approaches to spectral and temporal processing of time-frequency plane, multi-band speech processing, TempoRAl Pattern (TRAP) analysis, and discriminant nonlinear mapping approaches to acoustic speech analysis. It should give a student a good working knowledge of the state-of-the-art in this important area of speech processing and also some intuitions into general principles of processing \ of other information bearing sensory signals. This intensive short course consists of 15 hours of lectures and10 hours of discussions. A typical day would consist of three-hour lecture in the morning and two-hour afternoon discussions related to the morning material. The course is concluded by an oral exam.
- Language of instruction
- English
- Further comments (probably available only in Czech)
- The course is taught: in blocks.
- Enrolment Statistics (Autumn 2000, recent)
- Permalink: https://is.muni.cz/course/fi/autumn2000/P132