nasals approximants rhotics and laterals
play

Nasals, approximants, rhotics, and laterals C s m v hc v ng m hc - PowerPoint PPT Presentation

Nasals, approximants, rhotics, and laterals C s m v hc v ng m hc Formant ranges, F1 and F2: F1: 300 1000 Hz Lecture 15 F2: 800 2500 Hz Recap Recap Last time we saw that stops consist of both a closure


  1. Nasals, approximants, rhotics, and laterals Cơ sở âm vị học và ngữ âm học Formant ranges, F1 and F2: F1: 300 ∼ 1000 Hz Lecture 15 F2: 800 ∼ 2500 Hz Recap Recap • Last time we saw that stops consist of both a closure • Voiceless stops can be identified by concentrations of ( cấu âm tắc ) and a burst ( tiếng nổ ). energy ( khu vực năng lượng ) in the release burst. 5000 k p t Frequency 0 0 0.738 From Kent & Read (1992), The Acoustic Analysis of Speech Time (s)

  2. Recap Nasals • Like stops, nasals involve a vocal tract constriction ( cấu âm thắt trong đường dẫn âm )... • Voiced stops, on the other hand, were better identified using formant transitions ( chuyển tiếp formant ). simmer http://www.cns.nyu.edu/ ∼ david/courses/perception/lecturenotes/speech/speech.html From Ladefoged (2003), Phonetic Data Analysis Nasals Nasals • ...but like vowels, nasals have a periodic glottal source • Energy (amplitude) is lower than for vowels – in part ( nguồn thanh hầu có chu kỳ ). because nasal membranes absorb the sound (màng nhầy hút âm thanh). simmer From Ladefoged (2003), Phonetic Data Analysis

  3. Nasals Nasals • Like stops, the crucial information is contained in the • Since the oral tract is completely closed, this effect is formant transitions. enhanced, giving rise to antiformants ( formant yếu ). • In practice, the velar nasal is usually the clearest. dimmer dinner dinger Nasals Formant transitions preceding nasals • Like stops, the crucial information is contained in the formant transitions. − [m] : falling F2 ← • In practice, the velar nasal is usually the clearest. − [n] : level F2 ← − [N] : rising F2 ← dimmer dinner dinger (reverse for transitions following nasals) From Ladefoged (2003), Phonetic Data Analysis

  4. Nasals Approximants • F1 tends to be low (250-300Hz) and F2 around 2500Hz. In between (where vowels have F2) there is little energy. • Vowels are periodic – complex repeating waves ama ana a N a generated by regular vibrations of the vocal folds. • Consonants may have an aperiodic component, but can still be voiced or voiceless • There are also (largely periodic) segments that lie somewhere in between: approximants. Approximants Approximants • Vowels are periodic – complex repeating waves • Vowels are periodic – complex repeating waves generated by regular vibrations of the vocal folds. generated by regular vibrations of the vocal folds. • Consonants may have an aperiodic component, • Consonants may have an aperiodic component, but can still be voiced or voiceless but can still be voiced or voiceless • There are also (largely periodic) segments that lie • There are also (largely periodic) segments that lie somewhere in between: approximants. somewhere in between: approximants.

  5. Approximants Approximants • Acoustically, [j] [w] are very similar to vowels, with • Acoustically, [j] [w] are very similar to vowels, with striations and formants, but often with less energy striations and formants, but often with less energy • For this reason they are sometimes called semivowels . • For this reason they are sometimes called semivowels . • The constriction for [j] [w] is narrower than for [i] [u], but • The constriction for [j] [w] is narrower than for [i] [u], but formant values are similar formant values are similar Approximants Approximants • Acoustically, [j] [w] are very similar to vowels, with striations and formants, but often with less energy • For this reason they are sometimes called semivowels . • The constriction for [j] [w] is narrower than for [i] [u], but formant values are similar From Borden, Harris, & Raphael (1994)

  6. Approximants Approximants • [w] and [j] tend to differ from [u] and [i] in that they tend • What constitutes a diphthong vs. a semivowel + a vowel not to have a steady-state portion ( một phần không đổi ). can be controversial... From Kent & Read (1992), The Acoustic Analysis of Speech From Kent & Read (1992), The Acoustic Analysis of Speech Rhotics (‘r’-tính) Rhotics (‘r’-tính) • The defining acoustic signature of [ô] is the third formant • The defining acoustic signature of [ô] is the third formant (F3) making a steep fall to below 2kHz. (F3) making a steep fall to below 2kHz. • This effect is referred to as ‘r-colouring’ ( r-hoá ). • This effect is referred to as ‘r-colouring’ ( r-hoá ). • [ô] – and other rhotics ( [r], [R], [ö] ) – are the only • [ô] – and other rhotics ( [r], [R], [ö] ) – are the only approximants for which F3 is (perceptually) important approximants for which F3 is (perceptually) important

  7. Tongue configuration types of Rhotics (‘r’-tính) American English /r/ ([ ô ]) • The defining acoustic signature of [ô] is the third formant (F3) making a steep fall to below 2kHz. • This effect is referred to as ‘r-colouring’ ( r-hoá ). • [ô] – and other rhotics ( [r], [R], [ö] ) – are the only approximants for which F3 is (perceptually) important Examples of corresponding American English tongue configuration types for sustained /r/ as identified from MRI by Tiede et al. (2004). Adapted from Tiede et al. (2007), ‘Variability of North American English /r/ Production in Response to Palatal Perturbation’, Haskins Internal Workshop on Speech Production and Motor Control. Rhoticization Taps ( âm vỗ ) and trills ( âm rung ) 3: Ç: From Ladefoged (2003), Phonetic Data Analysis

  8. Where F3 is useful Lateral approximants • [l]: faint formants at a) low frequency and b) about • [ô] : characterised by low F2 and especially low F3. 1000-1500 and 2500-3000 Hz, followed by abrupt change in amplitude when the tongue tip breaks away. Frequency (Hz) Time (ms) From Ladefoged (2006), A Course In Phonetics From Ladefoged (2006), A Course In Phonetics Clear vs. dark [l] Clear vs. dark [l] • Catalonian (clear, top) vs. Mallorcan (dark, bottom) /l/s 5000 [ l ] • Allophones in many F2 English dialects: dark /l/ 5000 [ l ] in coda (‘feel’), clear /l/ F2 Frequency (Hz) in onset (‘leaf’) Frequency (Hz) 0 • In Newcastle English, 5000 initial /l/ is light; in [ ɫ ] 0 Recasens & Espinosa (2005), ‘Articulatory, Leeds, initial /l/ is dark F2 positional and coarticulatory characteristics for 5000 clear /l/ and dark /l/: evidence from two Catalan (Carter & Local, 2007) [ ɫ ] dialects’, JIPA 35(1), 1-25. F2 • In Scottish English, /l/ can be dark in both 0 Time positions. 0 Time

  9. Clear vs. dark [l] Clear vs. dark [l] 5000 5000 [ l ] [ l ] • Allophones in many • Allophones in many F2 F2 English dialects: dark /l/ English dialects: dark /l/ in coda (‘feel’), clear /l/ in coda (‘feel’), clear /l/ Frequency (Hz) Frequency (Hz) in onset (‘leaf’) in onset (‘leaf’) 0 0 • In Newcastle English, • In Newcastle English, 5000 5000 initial /l/ is light; in initial /l/ is light; in [ ɫ ] [ ɫ ] Leeds, initial /l/ is dark Leeds, initial /l/ is dark F2 F2 (Carter & Local, 2007) (Carter & Local, 2007) • In Scottish English, /l/ • In Scottish English, /l/ can be dark in both can be dark in both 0 0 Time positions. Time positions. Spectrogram practice • What distinguishes these approximants?

Recommend


More recommend