Speech recognition Flashcards
Acoustic signal
Speech sounds or patterns of pressure changes
Articulators
Structures involed in speech production, including lips, teeth, tongue, jaw, soft palate
Sound spectrogram
Graph depicting intensity of speech sound frequencies
Y-axis - frequency
X-axis - sound produced
Intensity indicated by darkness
Formants
Frequencies at which sound intensity peaks occur
Formant transition
Rapid shifts in frequency preceding or following formants
Manner of articulation
How a speech sound is produced by the interaction of articulators
Place of articulation
Locations of articulation during speech sound
Phoneme
Smallest unit of sound that, if changed, would change the meaning of a word
What are the three causes of speech acoustic signal variability?
Coarticulation
Sloppy pronunciation
Individual differences
Coarticulation
Overlap between the articulation of neighbouring phenomes
Categorical perception
When stimuli existing along a continuum are perceived as divided into discrete categories
Vocal onset time
Time delay between when a speech sound begins and when the vocal chords begin vibrating
Phonetic boundary
VOT at which the percept of a sound changes categories (e.g., “da” to “ta” as VOT passes 40 ms)
Multimodal
The involvment of multiple sense in determining speech perception
Categorical perception experiment
Test to discern a sound’s phonetic boundary
Discrimination test
Part of a categorical perception test, in which the VOTs of two stimuli are simultaneously raised until they are on opposite sides of the phonetic boundary
McGurk effect/audiovisual speech perception
The effect of visual perception on the perception of speech sound
Phonemic restoration effect
When an obscured phenome of a word is restored in the perception of the word
Shadowing
Experimental technique in which listeners repeat aloud what they hear through earphones as they hear it
How do listeners decode speech into words and meanings?
Previous knowledge of words and meanings to perceive words in sentences
Speech segmentation via knowing transitional probabilities
Experiential learning (“pop-out” effect) to preceive degraded speech
Speech segmentation
Process of decoding words from continuous acoustic signal (perceiving breaks between continuous words)
Transitional probabilities
The chances that one sound will follow another sound
Transitional learning
The process of learning about transitional probabilities from an early age
Noise-vocoded speech
Experimental technique in which speech signal is divided into frequency bands and then noise is added to each band
“Pop-out” effect
Ability to hear previously unintelligible words
Broca’s area
Area in left frontal lobe associated with language grammar and structure
Broca’s aphasia
Condition caused by damage to Broca’s area associated with difficulty producing speech (slow, labored, and jumbled). Difficulty understanding speech whose meaning depends on word order
Wernicke’s area
Area in tempora lobe associated with speech recognition
Wernicke’s aphasia
Condition caused by damage to Wernicke’s area characterized by inability to produce or understand meaningful speech (more profound than Broca’s aphasia, despite production of fluent “nonsensical” syllables)
Word deafness
Extreme Wernicke’s aphasia characterized by complete inability to recognize words
Voice cells
Neurons in “voice area” - section of STS - that respond most strongly to human voices
Describe the two streams identified in the dual-stream model for speech perception
“What” (ventral) pathway - anterior auditory cortex to frontal lobe. Responsible for recognizing speech
“Where” (dorsal) pathway - posterior auditory cortex to parietal lobe to motor cortex. Responsible for connecting speech syllables to motor movements
Phonetic features
Cues associated with the unique motor movements used to produce a phenome - connected to specific neurons
Motor theory of speech perception
Theory proposing perception of speech depends: first, on activation of motor mechanisms; and second, on activation of additional perception mechanisms
Audiovisual mirror neurons
Neurons observed in monkeys that activate both when performing an action that produces a sound and when hearing the sound of someone else carrying out the same action
Define and describe an example of experience-dependent plasticity
A change in the brain’s ability to respond to specific stimuli that occurs as a result of experience
Age-dependent reduction in ability of infants to discriminate sounds of foreign langauges
Social-gating hypothesis
Theory of language learning postulating that learning is “gated” by social brain (social interaction necessary for learning)
Motor theory of perception study (Liberman et al.)
Using TMS, slimulated different brain areas associated with different articulators; responding improved for phonemes that corresponded to stimulated area
