Version Okt 2007 Speech Science W Barry Institut
Version Okt. 2007 Speech Science W. Barry Institut für Phonetik, Universität des Saarlandes IPUS
Term Programme 1 Wk 1 : Wk 2 : Wk 3 : Wk 4 : Wk 5 : Wk 6 : Wk 7 : • What is Speech Science - scientific goals? Übung 1 • Capturing and representing Speech Übung 2 • Speech production I – Breathing Übung 3 • Speech production II – Speech organs and articulation; Übung 4 • Speech production III – Speech motor control 1; Übung 5 • Speech production IV – Speech motor control 2; Übung 6 • Acoustic structure of speech I – Source-Filter Übung 7
Term Programme 2 Wk 8 : Wk 9 : Wk 10: Wk 11: Wk 12: Wk 13: • Acoustic structure of speech II – Vowels Übung 8 • Acoustic structure of speech III – Consonants Übung 9 • Variable acoustics – constant perception Übung 10 • What do we perceive? Sounds/syllables/words? Übung 11 • What happens in fluent speech? – Articulation and acoustics; Übung 12 • What do we produce when we speak? Sounds/syllables/words? ; (distribute trial exam) Übung 13: Discussion of trial exam
Term Programme 3 Wk 14: • Discussion of practice exam Wk 15: • Final exam Coursebooks: R. D. Kent (1997). The Speech Sciences. San Diego/London: Singular Publishing Group, Inc. G. J. Borden, K. S. Harris & L. J. Raphael (1994). Speech Science Primer. Physiology, Acoustics and Perception of Speech. (3 rd edition). Baltimore: Williams & Wilkins. B. Pompino-Marschall (2003). Einführung in die Phonetik. (2 nd edition) Berlin/New York: Walter de Gruyter H. Reetz (1999). Artikulatorische und akustische Phonetik. Wissenschaftlicher Verlag Trier (WVT).
What is Speech Science? Topics • Scientific goals • Areas of Speech Science --------- • "Homework": a) Kent, Chap. 1, pp. 1 -20 (K) b) Borden, Harris & Raphael, Chap. 2, pp. 14 -23 (BHR) Deutsch: c) Pompino-Marschall, Teil I, S. 1 -10; Teil II, S. 13 -16 (P-M)
Scientific goals • Science: Seeks to explain the “Hows? ” of the world • What are the “Hows? ” of speech? • For any question “How? ” there has to be a “What? ” • So … what is speech? The systematic movement of our articulators? The sound patterns we perceive? The sound waves that travel from speaker to hearer? Anything else?
What is speech? • All of the things mentioned ……. but more! • We really need a “communication framework” for speech to actually take place • The articulatory movements The acoustic pressure patterns The sound patterns that we perceive HAVE TO BE RELATED TO A MESSAGE • ……… so what do we mean by “message”?
The “message” of speech? • …. . is the information the speaker conveys to the listener Which is? • Mostly, the semantic content of the utterance (though this can be metaphoric or otherwise indirect) • Also, the speaker’s attitude to the content …. . and to the person(s) being talked to • But speech also reflects things about the speaker (indexical information: sex, age, health, mood ……) • So the relationship between message and sound patterns is much more than how /i a u/ and /p t k/ are produced.
Speech as a meta-message? • Our definition of “message” can be problematical for speech research: • Speech research is often associated with unnatural utterances … (something our definition rejects as “genuine speech”? ) Words set in carrier sentences Word lists containing target sounds Repeated phrases……. etc. However…. . • Within the framework of the research question they DO convey information to the listener (= to the investigator) …. As a meta-message = a message about the linguistic form So, within a particular theoretical question it can be valid material. • But such speech can never be considered “natural speech”
So, do we now know WHAT speech is? • A message encoded in language and expressed phonetically (i. e. through the medium of sound) …. . • …. . which means that a speaker is active, an acoustic signal is produced…. . • …. . and it is assumed that the speaker is talking to someone. I. e. , a hearer receives and processes the signal (decodes the message) This “speech chain” defines the broad area within which Speech Science asks its questions.
Speech Chain neural processes neuromuscul. process Hearing Articulation Stimulus neural transform processes Acoustics Measurement Speech signal CNS Observation domains system Speech signal time as a function of Observation domains within the speech chain (adapted from P-M. p. 14)
Speech Chain II From Denes & Pinson, 1993, p. 5
Observation domains I • Neural processes (production and perception) – increasingly of interest (psychology & medicine) EEG (Electroencephalography) and MEG (Magnetoencephalography) MRI (Magnetic Resonance Imaging) • Global information about neural activity: EEG and MEG provide high temporal resolution of activity in small areas of the brain. MRI shows activity in all areas of the brain but with low temporal resolution. • Nothing is clear from one observation (noisy signals). Patterns emerge from the average of many repetitions.
Observation domains II • Neuro-muscular processes (production) EMG (Electromyography)
Observation domains III • Articulation (Production) – Traditional observation domain in Phonetics. Nasal cavity Hard palate Soft palate Pharynx Tongue Epiglottis Jaw Larynx Trachea
Observation domains IV • Acoustic Signal – Increasingly central to speech research from 1945 onwards. • It has a central position between speaker and hearer; it is the product of articulation and the input to the perception process. • Different representations give us information about the distribution of acoustic energy in time or in frequency. b) a)
Observation domains V • Auditory system (the physiological and neurological foundation of perception) Outer ear Inner ear Middle ear Semi-circular canals balance nerve auditory nerve Stapes Incus Cochlear Ear canal Scala vestibuli Scala tympani Eardrum & Malleus Endolymph Oval window duct Round window The peripheral auditory system (ear)
Observation domains VI • Stimulus transformation (psycho-acoustics, psycho-phonetics, speech perception) • What are the acoustic properties that keep vowels and consonants, monophthongs and diphthongs, stops and fricatives, voiced and voiceless consonants apart? • The “same sound” is acoustically different when it is produced by different speakers – very different if the speakers are men vs. women vs. children. Why do they sound the same? • The “same sound” is acoustically different when it is produced in different contexts.
Summing up…… • Different observation domains give us different perspectives on what people do to communicate. • Examining the phonetic events in relation to the message gives us: a) a lot of knowledge about the structure of speech b) insight into how the properties of speech affect the message • This serves many purpose: -The formulation of theories of speech (production and perception) -The development of pronunciation-teaching methods and therapies for the speech impaired -Applications in speech and language technology
- Slides: 19