Print › Acoustics

Acoustic Landmarks and Distinctive Features: Words are represented in memory as a sequence of segments, each of which consists of a bundle of binary (on or off) distinctive features. Perception is based on landmark detection, using points of minimal and maximal change.
Adaptive Response: The ways in which the individual compensates for a perterbation- online compensation
Artificial Palate: An appliance, inserted into the mouth and attached to the upper teeth that changes the contour of the palate. (Provides information about how the articulation of an individual phoneme may be disrupted, as well as information about how a speaker might adapt to or learn to compensate for the impediment, and what, if any, carryover might occur once the impediment is removed.)
Bite-Block: Typically a semi-rigid material made from a dental molding impression placed in a frontolateral position between the upper and lower central incisor teeth. It is kept in place during utterance production. It has been shown that normal speech is not adversely affected by fixation of the mandible with a bite-block.
Categorical Perception: Stimuli that vary along some continuously changing dimension are perceived as belonging to two or more discrete categories. (People will "chunk" stimuli into categories)
Clear Speech: Subjects are instructed to speak as clearly or precisely as possible. Characterized by decreased speech rate.
Closed Feedback Loop: The controlling action is dependent upon the output.
Coarticulation: Adjustment of articulator movements to target more than one speech sound simultaneously. Influence of one sound on another.
Cohort Theory: The beginning of a word is perceived and then all words that have that beginning-the cohort-are reviewed, from which the correct word is selected. Cohort is formed based on similar acoustic features.
Connectionist Model: A group of models of speech motor control that are based upon nonlinear and nonhierarchical sets of components
Data Driven: The data obtained from the acoustic signal drives, or directs, the listener's perception of speech.
Degrees of Freedom: Ways in which the tongue muscle can move
Direct-Realist Theory: Perception consists of a single step from the acoustic signal to the percept.
Duplex Perception: The simultaneous perception of nonspeech and speech stimuli extracted from statements of the speech signal.
Dynamic Systems: The study of forces that result in the movement of objects. To take infinite degrees of freedom and reduce them to an effective number of degrees of freedom
Fairbanks (1956): Closed-loop model, auditory, tactile & kinesthetic feedback, discrepancies are detected by a feedback loop, correction detection in comparator.
Feed-forward models: Propose that adjustments in speech production are made at the periphery of the system, that is, to the actual articulatory gestures. Errors can be corrected more quickly.
Frame Theory: Based on the concept that the infant's babbling consists of a cyclic alternating opening and closing of the mouth arising from the movement of the mandible (the "frame") superimposed upon vowel and consonant production. With maturation, the cyclic opening and closing evolves into the syllable structure (the frame) and the segmental articulation of phonemes (the content). The syllable is rooted in a rhythmic stress pattern which originates in an area of the brain different from where phonemic content originates.
Gating: t-tr-tre-tres-tresp-trespa-trespas-trespass. Listener tries to identify the word after the presentation of each successive fragment. Takes 1/3 of a second to identify word
Hybrid Theory: Combination of Look Ahead and Frame Theories-Two phases of articulatory gesture. Low velocity (anticipate/important) and high velocity.
Jakobson, Fant, and Halle (1963) and Chomsky and Halle (1968): Phonemes are viewed as bundles of features: voicing, noise, burst. The features are called distinctive. Each speech sound has a physiological bundle of distinctive features, and an acoustical bundle of features. Does not seem to be a universal set of distinctive features.
Kozhevnikov and Christovich (1968): Open loop, motor commands specify complete syllable, used duration and reaction time data.
Ladefoged (1972): Spatial-acoustic target: spatial target, access to auditory representation, brain uses rules to relate these.
Lashley (1957): Several interacting systems: speakers intention, store of images and lexicon, motor organization, temporal ordering. Syntax provides temporal control:orders words, orders motor activity.
Lexical Information: Selecting the correct phoneme based upon which word was most likely to have been spoken, given the content of the utterance
Liberman, Cooper, Shankweiler, and Studdert-Kennedy (1987): Speech sound is encoded in the acoustic signal. Starts with phonemes, ends with speech signal. Phonemes are processed in parallel producing coarticulation. No single direct conversion from phoneme to sound. Multiple conversations are needed.
MacNeiliage (1970): Motor equivalence, gamma loop control, spatial targets, motor commands for speech are not invariant, open loop.
Markov Chains: Probabilistic math function chain
McGurk Effect: A film was created in which a speaker produced the syllable /ga/ but the syllable /ba/ was dubbed into the sound track. The viewers heard /da/ instead of /ga/ or /ba/. Articulatory information from both the auditory and visual perceptions was combined to yield a different phoneme.
Models: Means of simplifying complex systems or processes (approximate real world systems)
Motor Equivalence: A single articulatory gesture can sometimes be achieved with different patterns of muscle activation. (different behavioral responses to the same stimulus)
Motor Equivalence 2: Individuals demonstrate different motor behaviors in response to the same stimulus.
Motor Planning Ahead: Efficiency is achieved when one anticipates. Scan ahead mechanism = fast.
Motor Programs: Preconstructed set of central commands capable of carrying out a movement
Motor Theory of Speech Perception: Speech perception is based on invariant articulatory gestures. That is, speech perception uses the motor commands of the speech production system as the units of perception. The listener assesses her own knowledge of how sounds are produced, and then uses that reference to process the perception of sounds produced by another individual.
Native Language Magnet Theory: Holds that the phonetic categories of one's native language are organized as prototypes, which begin in infancy by attracting or assimilating nearby members of the same phonetic category. Such prototypes are called "perceptual magnets".
Normalization: refers to the process of simplification by smoothing out variability or "noise" and the unnecessary or superfluous variability to capture better the essence of a signal
Ohman (1966): Mathematical model of articulation. 50 radial planes in the VT. Model includes :state properties and dynamic rules.
Open Feedback Loop: The mechanism that controls the action is not a central part of the activity itself, but rather an outside agent. Control is independent of the action.
Output Target: Acoustic signal
Path: Sequence of positions in space occupied by an object or articulator (spatio-temporal)
Percept: That which the listener hears-the "object"- as contrasted with the actual acoustic event.
Perception: a subjective occurrence that rarely, if ever, matches exactly the actual physical event being perceived.
Perterbations: Unanticipated, small disturbances
Peterson and Shoup (1966): Describe sounds (perceptual) in terms of physiological data, acoustical data, and transformations from physiological to acoustical data.
Quantal Change: The abrupt transition from one state to another due to a continuously changing variable
Quantal Perception: The phenomenon of discontinuous categorical perception
Sapapple: Short utterances like "Buy Bobby a poppy"
Sensory feedback: a transfer of a portion of the systems output back to the input for regulation and error correction. Changes motor movement.
Serial Ordering: Expect things to come in certain order
Sinewave Speech: Composed of sine waves that track the center frequencies of F1, F2, and F3 of a naturally produced sentence. In other words, a "real" sentence, as spoken by a human being is recorded. (Zoo/Lunchtime class example)
Slots: Specific positions in the syllables to which the content can be placed
Spatio-Temporal Organization: The movements of the articulators relative to some frame of reference
Speaking Task: Vary one feature systematically while holding the others constant. Experimental control-one variable at a time.
Speech Perception Theories: (Majority) Focus on identification of phonemes within the speech signal.
Suprasegmental: The level of meaning of an utterance carried by features that are superimposed upon the segmental level, including intonation, stress, and duration. Also called prosody.
Target Models: Speech is considered the acoustical result of attaining phoneme targets
TRACE Model: Speech units are arranged into three interactive levels: features, phonemes, and words. Each level is comprised of nodes, or processing units. For example, at the feature level, one node would detect voicing, another would detect lip rounding. Activation occurs form the bottom upward.
Trajectory: References the timing of the sequence of positions (spatio-temporal)
Unit of Analysis: Could be phonemes (sounds), syllables, words, or articulatory gestures(probably greater than the phoneme -phrases- because of coarticulation)
Voice Onset Time (VOT): Duration between the release of the closed portion of the plosive and the onset of the voicing
Word-Recognition Theories: Focus on the process by which a sequence of phonemes is transferred to lexical representations held in the memory of the listener.