Quiz 4 Flashcards
Vowel Analysis 2 | Consonants 1
What is vowel normalization?
Using methods/techniques to account for variability in speakers
Normalization aims to make vowel comparisons ________ and _________
consistent; reliable
What measure do we use to make vowel normalization?
Z-scores
Name 4 goals of vowel normalizations
- Take out physiological variation caused by MOUTH size.
- Preserve sociolinguistic diffs in vowel QUALITY
- Preserve PHON. distinctions among vowels
- Model COGnitive processes in vowel normalization
Name the 4 vowel normalization techniques?
- Vowel-intrinsic method
- Vowel-extrinsic method
- Speaker intrinsic method
- Speaker extrinsic method
Vowel norm. - vowel intrinsic method (e.g.)
- using info. from a SINGLE vowel token - 1 formant in 1 vowel
- F frequency adjustments : adjusting F values based on average values for similar speakers
Vowel norm. - vowel extrinsic method (e.g.)
- Comparing F values across diff. vowels by the SAME individual - 1+ formants in 1+ vowels
- Grand mean normalization : normalizing vowels using a GRAND mean calculated from all vowels spoken by the person
Vowel norm. - speaker intrinsic method (e.g.)
normalize based on data from a single speaker’s vowels
- Individual speaker baseline : making a baseline for formant freqs. for a SINGLE speaker
Vowel norm. - speaker extrinsic method (e.g.)
take data from a MULTIPLE speaker’s vowels
- Mean adjustment method : normalizing using mean values derived from a group of speakers
Hawaiian is known for what kind of stops?
Voiceless
The acoustic cues of stops
- Closure duration
- Prevoicing and aspiration
- VOT
- Formant transition
Pressure becomes _______ in stops at what point?
bigger; Between a complete-closure of the airflow (velar port is closed) and when pressure grows and RELEASES the STOP
________ vibrations, and ______ open vocal _____ are additional gestures that helps to produce different TYPES of stops
vocal fold; wide; folds
What kinds of stop exist
voiced and voiceless
Air pressure = _____
amplitude
True or F : it IS useful to use absolute amplitude
F - relative amplitude is better to use
VOT is longer in both __________ and ____________ stops
aspirated; voiceless
T or F : VOT is NEGATIVE in voiced stops like [b, g]
T
What is and what does a voiced bar do?
A voiced bar is at the BOTTOM of the spectrogram, representing pre-voicing, before the vowel, and is used to distinguish contrasts
Vocal pulses in spectrograms =
full cycles
Which set contains sounds that are all likely to have a negative VOT?
A. [b, d, g]
B. [p, b, m]
C. [p, t, k]
A. [b, d, g]
What can formant transitions show us?
Give cues of PLACE of articulation
Give examples of formant transitions
- LOWERING of ALL Fs by bilabial stops
- Dentals RAISE F2
- Velars bring F2 and F3 nearer to eachother
What is the ‘velar’ pinch
Shown in the spectrogram, velar sounds will ‘pinch’ F2 and F3 together (e.g. GAG)
Which of the following scenarios requires vowel normalization? (Two correct answers.)
A. A study comparing vowel formants from 10 native speakers of Nepali of whom 5 are old and 5 and young.
B. Analyzing two speech samples from a 40-year old Japanese politician who spoke at two events 6 months apart.
C. Analyzing the spectral properties of a UBC student speaker’s recorded speech in a quiet laboratory setting.
D. A study that wants to compare formant frequencies of vowels produced by Korean, Hausa, and Zulu languages.
A. A study comparing vowel formants from 10 native speakers of Nepali of whom 5 are old and 5 and young.
D. A study that wants to compare formant frequencies of vowels produced by Korean, Hausa, and Zulu languages.
Which of the following are NOT true about vowel normalization methods? (Two correct answers.)
A. Vowel-extrinsic methods use information from different vowels produced by the same individual.
B. Vowel-intrinsic methods use information from multiple vowel tokens produced by the same individual.
C. Speaker-intrinsic methods use information from a single speaker’s vowels.
D. Speaker-extrinsic methods use information from multiple speakers.
B. Vowel-intrinsic methods use information from multiple vowel tokens produced by the same individual.
Which of the following categories of sounds will likely have positive VOT? (Two correct answers.)
A. Voiceless aspirated stop
B. Voiced unaspirated stop
C. Voiced aspirated stop
D. Voiceless unaspirated stop
A. Voiceless aspirated stop
D. Voiceless unaspirated stop