1/60
Looks like no tags are added yet.
What is Hockett’s Easter Egg Analogy?
We have distinct speech sounds (phonemes) in our brains, yet when spoken they blend together, but the listeners brain is able to decode the mess into distinct words
What is coarticulation? Give an example
Coarticulation is when production of a speech sound is influenced by nearby sounds
ex. /di/ vs /du/ the lips are already rounding and tongue is getting in position for the upcoming vowel
What are the two types of coarticulation?
Anticipatory (regressive) coarticulation & Preservative (progressive, carryover) coarticulation
Anticipatory (regressive) coarticulation
A sound is influenced by a later sound — you anticipate the upcoming movements
ex. the lip rounding on /k/ in “cool”
Preservative (carryover, progressive) coarticulation
A sound is influenced by an earlier sound (lingering effect)
ex. devoicing of /z/ in “it’s”
Why does coarticulation create both segmentation and invariance problems?
For segmentation, you can’t find isolatable phonemes
you can’t figure out when one phoneme ends and when the next one starts
Lack of invariance
the same phonemes are acoustically different in different contexts
How does the /d/ sound differ in “dee” and “do” and why?
F2 drops in “do” and rises in “dee”
tongue is high in front /I/ and lips are spread for dee
tongue is back and lips are rounded in “do”
How can a following nasal sound affect the vowel before it?
the velum gets lower early in anticipation of the nasal so the vowel becomes partially nasalized
/i/ in “bean”
What is the difference between anticipatory vs preservative coarticulation?
Anticipatory: a later sound influences an earlier one
Preservative: an earlier sound influences a later one
Why can we perceive speech so quickly even though sounds overlap?
“speech code”
encoding of phonemes by speaker
decoding of phonemes by listener
How does lip rounding affect /s/ in “suit” compared with “seat”?
/s/ is a lower frequency in suit due to lip rounding
How are pitch, loudness, and voice quality different?
Pitch = perceived frequency — how high or low a voice sounds
Loudness = perceived intensity —- how loud or soft a voice sounds
Voice quality = the perceptual characteristics that distinguish one voice from another
Name two factors that can influence voice quality
Vocal fold vibration — degree of vocal fold closure, glottal airflow and air leakage
Vocal tract characteristics — shape, length, tongue and jaw position
How can vocal fold closure and airflow affect the sound of a voice?
If vocal folds aren’t completely closed, this can cause air leakage which leads to breathy voice
Very tightly closed vocal folds stopping one vocal fold from vibrating leads to creaky voiced (vocal fry)
Very strongly closed vocal folds leads to harsh/strained voice
Whisper is when the VF are apart and don’t vibrate
Why can a voice have more than one voice-quality characteristic?
This is because the different voice quality characteristics overlap
Why can’t a listener determine the exact physiological cause from voice quality alone?
auditory descriptions alone can’t establish a physiological cause because different physiological conditions can produce similar-sounding voices and one cause can produce different qualities
What are suprasegmentals?
Properties of speech that are not confined to single segments
involve manipulation of
f0 (pitch)
Intensity (loudness)
duration (length)
Prosody
can add linguistic/communicative structure — can affect the message itself
Paralinguistics
convey information about the speaker and/or nonlinguistic context ex. voice quality, speaking style
Prosody features
intonation
rhythm
tempo & timing
loudness
paralinguistic features
voice quality
whispered
breathy
creaky
harsh
speaking style
What is intonation and how can it distinguish a statement from a yes/no question?
Intonation is the manipulation of F0 (pitch) across an utterance
Statements typically end with a falling pitch whereas a yes/no question ends with a rising pitch
What is a continuation rise?
A rise in pitch at the end of an utterance to signal that more is to come
Pitch declination (downdrift)
lowering of pitch and loudness as the utterance unfolds
length effect
longer utterances starting with a higher pitch
What acoustic features can make a syllable sound stressed?
increased duration (length)
higher or more prominent pitch (f0)
greater intensity (louder)
fuller vowel quality
How can contrastive stress change what a speaker emphasizes?
Placing stress on certain words signals contrasts
ex. stressing “I” in “I didn’t take the book”
or “take” in “I didn’t take the book”
What is the difference between sentence stress & lexical stress?
Sentence stress is to highlight important information whereas lexical stress is distinguish lexical meanings or grammatical structures
word-class stress
to emphasize information content
given new stress
to emphasize new information
contrastive stress
to emphasize contrasting information
What is the difference between a trochee and iamb? Give on example of each.
Trochee is stressed (Strong) followed by unstressed (weak) Sw
ex. tiger, apple, water, menu
Iamb is (wS) unstressed followed by stressed
ex. giraffe, hotel, balloon, today, deny
What information about speech can we see from a talker’s face?
Lip movement:
opening & closing: timing cues
lip rounding & spreading: vowel and consonant cues
Jaw and facial movements:
articulatory information
How can matching & mismatching visual speech cues affect perception?
Matching visual cues improves accuracy and speed whereas mismatching cues can distort or change what is heard
Why might seeing a talker’s face help in background noise?
the brain automatically combines visual cues—such as lip movements, teeth, tongue positions, and facial expressions—with auditory signals to decode speech
Why does looking at the mouth not necessarily mean better comprehension?
it doesn't show how much information is actually used or understood
Name two reasons listeners may differ in their use of visual speech
hearing status (ex. hearing loss)
age
attention
experience
How does clear speech differ from conversational speech?
clear speech is produced is when the talker is trying to be understood
longer movements
distinct articulation
prosodic changes
Conversational speech contains more reduction and articulatory overlap
clear speech is simply not…
talking louder
Why is clear speech not simply talking more slowly?
involves many changes
timing: longer vowels/consonats and changes in pauses
spectral detail: vowels may become more distinct
voice: f0 and intensity may change
What is the Lombard effect? Name two common changes.
Speakers change their vocal output when exposed to background noise.
common changes:
increased vocal intensity & f0
duration and spectral characteristics
Why doesn’t an acoustic change always improve communication?
increased intensity doesn’t equal improved phonetic distinctiveness
slower speech does not necessarily maximize intelligibility
a listener can look at the mouth without effectively using visual cues
How can listeners adapt when speech becomes difficult to hear?
combining auditory input with visual speech and linguistic context
pay more attention
Why is it important to measure intelligibility or comprehension when studying adaptation?
acoustic differences alone do not tell us whether communication has improved
intelligibility and comprehension measures show whether an adaptation actually helps the listener
What is the difference between aphasia and dysarthria?
Aphasia is a language disorder that affects language processing whereas dysarthria is a motor speech disorder that makes it difficult to produce speech — poor control of speech muscles
Name three ways clinicians can use acoustic analysis
characterize a speech or voice difficulty
track changes over time
evaluate treatment outcomes
What does automatic speech recognition do?
Converts a spoken audio signal into text
Why can ASR make mistakes with accents, noise, or coarticulation?
speech varies across speakers
coarticulation makes sound boundaries hard to locate
noise and context change what is recognized
What does text-to-speech do and how does it differ from ASR?
TTS converts text to speech whereas ASR converts speech to text
Why do pronunciation, stress, and intonation matter in synthesized speech?
they affect intelligibility and naturalness
wrong pronunciation or stress can change word identity or meaning and make speech sound robotic
How does infant-directed speech typically differ from adult directed speech?
IDS has higher pitch, wider pitch range, exaggerated intonation contours, slower rate, shorter utterances, more repetition
What is vowel hyper articulation and what might it help listeners do?
vowel hyper articulation is the stretching of the vowel space so vowels like /I/, /a/, /u/ are produced more extremely and sit farther apart
it may make vowel categories more distinct and help infants learn and distinguish the vowel categories of their language
What did the IDS and Foreigner-directed speech comparison predict about pitch, affect, and vowels?
Both IDS and FDS would be adapted for listeners with limited language knowledge, but they would differ by function
what did the IDS and FDS comparison suggest about the teaching function of vowel hyperarticulation?
The vowel hyper articulation found in both IDS and FDS (even without infant-specific pitch and affect) suggest that it is tied to didactic language-teaching purpose rather than only to emotional or attention getting goals.
How do attentional, emotional, and didactic functions differ across speech styles?
Attentional: IDS & PDS
getting and holding the listener’s attention (high pitch, exaggerated intonation)
Emotional: IDS & PDS
Expressing affect and bonding (positive, warm tone)
Didactic function: IDS & FDS
In the DIVA model, how do feedforward and feedback control differ?
feedforward: PLAN
learned motor plan —> Produce sound
fast and predictive
works without checking each sound first
Feedback: CHECK & CORRECT
produce —> compare with target
detect auditory/sensory mismatch
correct the movement
What is learned during babbling compared with imitation learning?
Babbling: Establish systemic relationships
exploring articulator space
learning to predict sensory consequences of movement
Imitiation Learning: Establish phoneme-specific relationships
determining the movements to reach auditory and somatosensory target regions that define specific sounds
learning to predict necessary movements to achieve particular sensory outcome
Name two methods researchers use to study infant speech perception
High amplitude sucking paradigm
Monitor heart rate
ERPs: MMN
Head-Turn Paradigm
Preferential looking paradigm
How does habituation and dishabituation help test whether an infant hears a difference?
Habitaution- infants are repeatedly presented with one sound until their response declines
Dishabituation- if the response rebounds the infant noticed the change
Why does no rebound in an infants response not prove the infant cannot hear the difference?
Infants discriminate VOT differences that cross the adult phoneme boundary but do not discriminate equal sized differences within a category
why does infant speech discrimination not require the ability to babble?
Perception develops before production and infants can discriminate sounds based on hearing alone
ability to hear differences does not depend on being able to produce them and young infants already show discrimination before babbling begins