心理语言学 4

Chapter 4 Speech Perception April 6, 2010 Speech perception nThe structure of speech nPerception of isolated speech segments nPerception of continuous speech nPerception of written language Introduction nLanguage comprehension: processes involved in understanding the meaning of written or spoken language. nLanguage comprehension may happen at phonological, lexical, syntactic, and discourse levels respectively. nThis does not necessarily mean that we process language in a strictly serial manner. nRecent models assume that we process information at multiple levels in an interactive way. Introduction nSpeech perception: (speech recognition) the understanding or comprehension of speech in a language. nThe complexity of speech perception Interference of environmental context Variability of the speech signal itself nSpeech recognition is a complex process. Normal environmental conditions are full of distractions that interfere with the fidelity of the speech signal. The variability of the speech signal also influence or distort the acoustic stimulus. nListeners make a series of adjustments in the course of perceptual recognition. The structure of speech nProsodic factors nPhonetics uArticulatory phonetics uAuditory phonetics uAcoustic phonetics Spectrogram Parallel transmission Context- conditioned variation The structure of speech nProsodic factors: factors that may influence the overall meaning of an utterance (stress, intonation and rate) nStress：the emphasis given to syllables in a sentence. (e.g. present, refuse, project, pervert, etc) e.g. He is our undertaker (承办人）. He is our undertaker. （殡仪员） cf. He voted for John. (他替约翰投票表决） He voted for John. (他投票赞成约翰） nIntonation ：the use of pitch to signify different meanings. ( We aim to please. You aim too, please). nRate：the speed at which speech is articulated. (Lets wait and see!) nProsodic factors are sometimes called suprasegmentals. Phonetics nPhonetics: the study of speech sounds Articulatory phonetics: the study of how speech sounds are produced, how to classify and transcribe them Auditory phonetics: the study of the impression a speech sound makes on the hearer. Acoustic phonetics: the study of the physical properties of speech sounds. Articulatory phonetics Articulatory phonetics nDistinction between vowels and consonants: airstream obstruction nConsonants: manner of articulation, place of articulation, voicing (vocal cord vibration) uPlace of articulation: the different vocal organs, or more specifically, different parts of organs that are involved in the production of speech sounds (where air is impeded) uManner of articulation: the degree of obstruction, complete, partial, or a mere narrowing (how the airstream is impeded) Articulatory phonetics nVowels are distinguished from one another chiefly by whether they are produced in the front, center, or back of the mouth, and whether the tongue position is high, middle, or low. nThe cardinal vowel system was invented by Daniel Jones as a means of describing the vowels in any language. nThe utility of distinctive features is that they allow us to describe the relationships that exist among various speech sounds in an economical manner. Daniel Jones nDaniel Jones (18811967), British phonetician . nHe wrote The Pronunciation of English in 1909 and An outline of English Phonetics in 1918. This is considered to be the first comprehensive description of Received Pronunciation. nThe problem of the phonetic description of vowels was a long- standing one. Earlier phoneticians such as Bell and Ellis had suggested a system of reference vowels; and Henry Sweet did much work on the systematic description of vowels. Jones however was the one who is credited with having solved the problem by introducing the concept of cardinal vowels, a system of reference vowels which are taught with much care in the British tradition. nThe International Phonetic Association still uses Joness model. Acoustic phonetics nOne of the most common ways of describing the acoustical energy of speech sounds is called a sound spectrogram. nSpectrogram is a visual representation of a sound, showing its component frequencies. It is measured by a spectrograph, an instrument used in acoustic phonetics. The spectrograph “prints” out a spectrogram on special paper. nA time scale is shown along the horizontal axis and a scale of frequencies along the vertical axis. The greater the intensity (i.e. the louder the sound), the darker the ink. Acoustic phonetics nSpectrograms: description of the acoustical energy of speech sounds nFormant: concentration of acoustical energy (dark bands in spectrograms) nFormant transitions: large rises or drops in formant frequency Acoustic phonetics nEach of the spectrograms contains a series of dark bands, called formants (共振峰), at various frequency levels. nTwo aspects of formants have been found to be important in speech perception. Formant transitions are