SLP 342 (ch 6-11)

source filter theory

vocal fold vibration (source → harmonics) + vocal tract (filter → formants)

formant

vocal tract shape → resonance patterns → produces F1, F2, F3

standing wave

a stable vibration pattern formed when sound waves bounce back and forth inside the vocal tract and reinforce certain frequencies

node

a point in the standing wave where there is no vibration

antinode

a point in the standing wave where there is vibration

oral boundary (lips = open end)

pressure node (low pressure); velocity antinode (high velocity)

glottal boundary (glottis = closed end)

pressure antinode (high pressure); velocity node (low velocity)

f1

tongue and jaw close/openness → vowel height (y-axis)

f2

tongue fronting or backing → vowel advancement (x-axis)

f3

fine articulatory shaping → lip rounding, tongue tip, etc.

extrinsic tongue muscles

movement

intrinsic tongue muscles

shaping

perturbation

small change or disturbance (in a sound or vibration)

jitter

frequency or pitch variation (x-axis)

shimmer

amplitude or loudness variation (y-axis)

wideband spectrogram

(0.005 s) emphasis on timing, formants, voicing; vowels

narrowband spectrogram

(0.025 s) emphasis on frequency and harmonics; consonants

vowel sounds

speech sounds made through an open vocal tract (shaping)

high-front vowels

/i/, /ɪ/ → F1 (low), F2 (high)

mid-front vowels

/e/, /ɛ/ → F1 (mid), F2 (high)

low-front vowels

/æ/ → F1 (high), F2 (high)

central vowels

/ə/, /ʌ/, /ɚ/, /ɝ/ → F1 (mid), F2 (mid)

high-back vowels

/u/, /ʊ/ → F1 (low), F2 (low)

mid-back vowels

/o/ → F1 (mid), F2 (low)

low-back vowels

/a/ → F1 (high), F2 (low)

dipthongs

/aɪ/ → F1: high → low, F2: low → high

/aʊ/ → F1: high → lower, F2: high → low

/oʊ/ → F1: mid → low, F2: low → high

/eɪ/ → F1: mid → low, F2: high → slightly higher

consonants

speech sounds made by one or more areas of constriction in vocal tract

downstream constriction

constriction closer to mouth

upstream constriction

constriction closer to vocal folds or glottal area

egressive sound production

speech produced on the outward flow of air from the lungs

coarticulation

• simultaneously articulating more than one phoneme

• anticipatory (forward) and retentive (backward)

classifying consonants

• place of articulation (where)

• manner of articulation (how)

• voicing (voiced or unvoiced)

voice bar

presence of voicing shown on spectrogram

vocal onset time (VOT)

timing present between release of a stop and onset of voicing

aspiration

breathy sound that may occur between release and voicing (long +VOT)

stops or plosives

• /p/, /b/, /t/, /d/, /k/, /g/

• stop burst, aspiration, VOT visible

fricatives

• /f/, /v/, /θ/, /ð/, /s/, /z/, /ʃ/, /ʒ/, /h/

• frication noise

nasals

• /m/, /n/, /ŋ/

• weak overall energy, nasal murmur

affricates

• /t͡ʃ/ (ch) and /d͡ʒ/ (j)

• stop (silence/burst) + fricative (frication noise) sequence

approximants

• /l/, /ɹ/, /w/, /j/

• vowel-like formants, no noise

prosody

intonation (pitch contour), timing (duration + juncture), loudness (intensity contour)

theory

an explanation of a phenomenon backed by empirical evidence

hypothesis

an educated explanation based on limited evidence; usually serves as the starting point for further investigation

model

simplified processes or systems (conceptual, computational, biological)

degrees of freedom

the issue regarding how the human speech system controls and coordinates the many possible movements involved in producing speech

motor program

pre-structured set of central commands capable of carrying out a movement executive (information processing) + effector (execution) + sensory feedback

output targets

the intended speech goals the CNS aims to produce (ex. acoustic targets, articulatory gestures, aerodynamic pressures)

acoustic targets

desired sound qualities the speech system aims to produce (auditory feedback)

articulatory gestures

coordinated movements of speech articulators used to create speech sounds (internal cognitive map of vocal tract positions)

aerodynamic pressures

airflow and air pressure conditions used to produce and shape speech sounds (respiratory coordination)

sensory feedback

transfer of some output back into input for regulation and error correction

open feedback loop

speech movements are produced by a preplanned motor command and run without using sensory feedback during execution

closed feedback loop

speech movements are continuously adjusted based on sensory feedback

feed-forward models

speech is generated using preplanned motor commands that are executed without needing real-time sensory feedback to guide each movement

mandibular movements

six possible degrees of freedom; four movements used in speech production

mandibular movement doesn't align neatly with phoneme boundaries…

because speech is an overlapping movement and is quite dynamic

directions into velocities of articulators (DIVA)

a computational and neuroanatomical model that explains speech production as the brain sending movement directions or commands that are converted into articulator velocities, while also using sensory feedback to correct errors

dynamical systems

a theory that explains speech as self-organizing movement patterns that emerge from interactions between the brain, body, and environment, where many degrees of freedom are coordinated into a smaller number of functional patterns

spatiotemporal organization

the way speech movements are coordinated in both space and time

spatiotemporal index

a measure of variability in timing and coordination of speech movements

connectionist models

various elements processing information simultaneously (node activation)

speaking task

a specific speech goal or action the speaker is trying to perform

perturbation studies

examining impact of disturbance to a speech production system (bite-block, artificial palate, airway occlusion, etc).

motor learning principles

practice/feedback used to enhance learning + retention of new motor behaviors

limitations of resources

finite resources that must be allocated appropriately

serial processing

each component of the multiple task is processed individually in order

frame-and-content theory

a theory that speech is produced by combining a rhythmic speech frame with speech sound content

slot-and-filler model

a model in which structural positions are filled with speech units such as phonemes or syllables

lack of invariance

no one-to-one relationship between a phoneme and its acoustic features (due to coarticulation and suprasegmental factors)

normalization

process of simplification by filtering out noise

motor equivalence

the same speech sound can be produced using different articulatory strategies

categorical speech perception

hearing speech sounds as distinct categories

quantal speech perception

hearing speech sounds as stable until a threshold is reached

duplex perception

simultaneous perception of non-speech and speech signal

McGurk effect

what you see affects what you hear during speech perception

sinewave speech

composed of sine waves (tracking F1, F2, F3)

gating task

a task where speech is revealed little by little to study how people recognize spoken words (context + prior knowledge)

bottom-up speech perception

understanding speech by starting with the sounds you hear and building up to words and meaning

top-down speech perception

understanding speech by using your knowledge, expectations, and context to help interpret what you hear

motor theory of speech perception

relating sounds you hear to the mouth and vocal tract movements that would produce those sounds

acoustic landmarks

noticeable changes in sound that help listeners identify speech sounds (hearing)

distincitive features

characteristics of speech sounds that help tell phonemes apart

mirror neurons

active when you perform an action and observe an action

indexical properties

characteristics of speech that provides information about the speaker’s identity (age, gender, region, etc.)

native language magnet theory

infants learn the common speech sounds of their native language, and these sounds become perceptual magnets that make similar sounds seem the same