lec 7 Flashcards
what is the simple vowel target model? (2)
- theory that each vowel has a canonical form which is invariant across contexts
- vowels are typically characterized by their first 2-3 formants
what is F1 and F2 for the vowels /a/ and /i/?
- /a/ F1 = 800, F2 = 1200
- /i/ F1 = 200-300, F2 = 2300
is the simple vowel target model accurate? why or why not?
- no!
- because vowel form varies from context to context
if everybody’s F1 and F2 are slightly different for /a/ and /i/, how does the brain process this?
- theory: has to do with the ratio between the formants
- that ratio = “acoustic fingerprint”
explain how double helmholtz applies to the F1 and F2 of /a/
- comparing to F1 and F2 of schwa (500, 1500).
- F1 for /a/: tongue is low = increased A1, increasing F1.
- F2 for /a/: tongue being back = increased V2, decreasing F2.
explain how double helmholtz applies to the F1 and F2 of /i/
- comparing to F1 and F2 of schwa (500, 1500).
- F1 for /i/: tongue goes front = increased V1, decreasing F1.
- F2 for /i/: tongue goes front = reduced V2, increasing F2.
what is the cardinal vowel system?
- organization of vowels in reference to most extreme high/front and low/back tongue positions
T or F: lax vowels are longer and tense vowels are shorter.
false – lax vowels are shorter (and are followed by consonants) and tense vowels are longer.
T or F: lax vowels do NOT occur in the nucleus of open syllables.
true
when working with clients, which should you aim to elicit first: vowels or consonants?
vowels because they provide structure.
what is perceptual magnet?
attracting neighbouring elements to sounds because you don’t have an equivalent to the actual sound in your perception
what difference do you notice between american vs UK english in terms of vowel plots?
- american english vowel plot is more spread out (less overlap in tongue position)
if a parent asked you to show proof that your therapy is effective, would you show them a vowel plot of their child’s progress or IPA transcription? why?
- vowel plot because this is ACOUSTICS (empirical)
- IPA is perception-based
what are the problems with simple vowel target model? (3)
- speaker normalization: biological differences = varying F1, F2, F3 values.
- target undershoot: running speech = articulatory targets not being met (yet listeners understand).
- vowel duration: temporal and dynamic variations.
what is non-linear scaling?
in relation to speaker normalization, non-linear scaling = brain using stats to neutralize differences