Scheduled maintenance: Thursday, January 26 from 6PM to 7PM PST
Upgrade to remove ads
Only $35.99/year

EXAM II review set: 1

Terms in this set (93)

frequency
1/period
period
1/frequency
fundamental frequency
the slowest/longest period; the lowest frequency
-all complex sounds can be mathematically analyzed as compositions of these sounds
-the higher the ____________, the more sparsely spaced the harmonics
Fourier's theorem
any periodic signal is composed of the summation (or superposition) of multiple sine waves with particular relative temporal alignment (phases)
harmonics
the component freq that compose a certain sound
power spectrum
useful for showing the freq composition (spectrum) of a sound; x = freq, y=amp
outer ear
the parts of the ear you see outside the head PLUS auditory caal
pinna
cone-shaped soft skin and cartilage outside ear
sound localization
where a sound is coming from (outer ear useful for this)
conductive hearing loss
when middle ear fails to conduct sound energy
inner ear
- a labyrinth
- located inside temporal bone
cochlea
provides some analysis of the composition of the sound wave that has now been translated into neural signal for the brain
basillar membrane
separates fluid filled cavities in the cochlea
-tonotopic organization
tonotopic organization
-describes the organization of the basillar membrane
-acts like a spectogram
base of the basillar membrane
- it is a part of the basillar membrane that is narrow and stiff
-responds to high frequency
apex
-it is a part of te basillar membrane that is wide
-less stiff and responds to low frequencies
sensoneural hearing loss
when the cochlea (inner ear) or auditory nerves are impaired - describing the resulting hearing deficit
eardrum
-the tympanic membrane
-separation between outer and middle ear
-exquisitely sensitive
middle ear (cavity)
- made up of ossicles- smallest bones in body
- the hammer, anvil, and stirrup
-increases the pressure from the eardrum of the outer ear to the inner ear by 20x
Corti Organ
our hearing sense organ that sits on the basillar membrane
-sensory cells are called hair cells (consisting of projections called stereoclia)
voicing
for ______, the fundamental frequency is the rate of vocal fold vibration (one cycle of vocal fold is one opening and one closing of the vocal folds)
source-filter model
describes the 3 components of spoken language
-source, filter, medium
source
-in the source-filter model
-it is the acoustic energy, primary generated by voicing, which yields a complex, periodic sound with harmonic structure
filter
-in the source-filter model
- is determined by the shape of the vocal tract that adjusts to strength of the different frequency components found in the source (s)
medium
-in the source-filter model
- a ________ through which signals travel (usually air)
spectograph
plots time on x-axis, freq on y-axis, and amplitude (darkness)
independence of the source and filter
the size/ shape of a person's vocal tract is independent of the rate of the vocal fold vibration
F1
-"height" of the vowel plane
-the higher the vowel on the vowel chart, the lower its _____
F2
- "back/front" of the vowel plane
-the higher the _____, the front-er the vowel
amplitude
sound pressure level of the fluctuations
periodic
presence of regular vibrations (created by the voicing behavior of the vocal folds)
-for voiced, regular repeating fluctuations are seen in the microphone signal captured for a sound wave (on your eardrum)
waveform
graphically and numerically records air pressure fluctuations over time - the sound wavve
resonant frequency
-the natural frequency
-a natural object's frequency that fits well within their part structure
formants
strong frequency in speech
-we are generally concerned with the 3 lowest frequencies
coordination
timing of 2 articulatory movements (using additional superscript/ subscript symbols called diacritics)
types of systemic variation
includes nasalization, VOT, devoicing, and aspiration
VOT
the time from the release of a stop closure to the onset of voicing
+, -, zero
devoiced
happens when consonants or short schwa vowels occur after a voiceless stop
aspiration
explains why "bad" and "pad" are not necessarily differentiated by voicing in English (both may start with voiceless closure)
parisimony
the simpler theory is preferred over the complicated one if they both explain things equally well
scientific method
observation, theory, prediction, experiment, revision, theory
falsifiability
no scientific theory can conclusively be declared true in an absolute sense bc there is always the possibility that tomorrow the theory will be deemed untrue
predictions
are tested with experiments
-if they are borne out, the theory is supported
-if not, the theory is falsified
independent variable
what the experiment purposefully manipulates
dependent
the effect of the experiment's manipulation
reaction time
speed of the response
confounding variable
random uncontrolled interference
-when undetected, leads to incorrect conclusions
noise
all sources of uncontrolled influences on the dependent variable of an experiment that aren't confounded with an independent
condition
based on how many harmony levels of an independent variable is tested (ex. for 2 = static noise present/not present; OR 3 = no static/ moderately static/ very static)
between subjects
each participant is tested under one condition only. One group of participants is tested under condition A, a separate group is tested under condition B, and so on
-regarding experiment groups
within subjects
each participant is tested under each condition. The conditions are, for example, "device A", "device B", etc. So, for each
participant, the measurements under one condition are repeated on the other conditions
-regarding experiment groups
central tendency
quantitative representation that captures the similarity among a set of data points
mean
interpreted as an approximation of what the measurement ould have been if each point w/o noise and w/o individual variation not part of the underlying measured process
dispersion
set of ways of qualifying the degree of variability in data (ex. standard deviation)
normal distribution
-a bell curve
- typical of processes in nature
probability
diff in 2 means (if less than .5%, the diff is statistically insignificant)
motor theory of speech perception
the idea that perception process involves representations of the motor command structure for the gestures that produced the acoustic signal
-one of the many theories in invariant cues
cues in consonant perception
for a given consonant, the contour of the formant transitions changes depending on the following vowels
speech perception
- may involve working backwards
- computing a representation of the gesture that would have produced the formant patterns in question, and perceiving a speech sound based on the computed score gesture
ratio
computing the _____ of formants is a possible way of extracting reliable properties of particular vowels
vocal tract size/shape
even if the filter of two sources are the same, the sound will still be diffferent
-twins with identical vocal tracts sound diff (accents, etc.)
speech processing
some scientists believe that this involves simultaneously evaluating several possible interpretations of the acoustic signal and selecting the most coherent one
McGurk effect
describes a situation in which the visual info a listener gets from watching the mouth of a speaker alters the listener's perception of what the speaker is saying
-vision affects the perception of speech bc visual info fundamentally affects the way we process auditory info
top-down processing
the breaking down of a system to gain insight into its compositional sub-systems
-an overview of the system is formulated, specifying but not detailing any first-level subsystems; each subsystem is then refined in yet greater detail, sometimes in many additional subsystem levels, until the entire specification is reduced to base elements
bottom-up processing
the piecing together of systems to give rise to more complex systems, thus making the original systems sub-systems of the emergent system
- a type of information processing based on incoming data from the environment to form a perception
- in speech: cognitive operations that build more abstract representations of language from less abstract ones
categorical perception
referring to the fact that we normally perceive speech sounds categorically as one sound or another, but not as a blend of sounds or as an intermediate or indeterminate sound
-implies that perceivers pay attention to the variability that matters
-perceivers have good between-category discrimination, but poor within-category discrimination
forced-choice ID
-subjects hear a CV syllable and have to report hearing [pa] or [ba]
-proves that the curve of speech perception is categorical, not continuous
- considers labeling facts
labeling facts
show that something different happens in the judgements of speech sounds depending on where along a cue's continuum an acoustic diff falls - there is a clear category boundary in this respect
category boundary
point at which perspetion switches from one speech sound to another; VOT boundary, for voicing differs slightly depends on place of articulation
the ABX discrimination experiment
confirms that syllables that are labeled equivalently are also not disciminable
infants
________ perceive speech categorically but have to learn the specific categories of a given language (bc a category - "aspirated" or "unaspirated" in one language may not make a different in another
-IN FACT, ______ perceive all the contrasts that languages make use of (which slowly go away); thus, adults can only perceive those of the language they speak
phonological development
while the details differ slightly for consonants v. vowels, w/in the first year of life infants learn to selectively perceive only the contrasts in their language; can only perceive categories in the language they grow up in
compensation for coarticulation
perceptual adjustment that shifts the category boundary in a way that compensates for the acoustic effect of coarticulation
-thus also proves that the way speech is perceived reflects the facts about the way it is produced
Gonaong effect
a sound that would normally be percieved as voiced is percieed as non voiced bc of top-down word knowledge
-further evidence for top-down process; shows that the abstract knowledge of words affects lower-level process of perception
-ex. listeners hear [dep] but since that is not a word in English, they perceive [tep] "tape"
-the VOT boundary had shifted slightly in order to make a word rather than a nonword
duplex perception
dual modes of perceiving speech sounds
-experiment: playing 2 diff parts of a speech sound in the right and left ear; the perceiver ends up perceiving both a speech sound AND a non speech sound
transitional probabilities
syllable-to-syllable predictability used by infants to segment continuous sequences of syllables into word-like units by observing that certain syllables always occur together (whereas other don't)
headturn test
• Conditioned: the image of toys appear after each sound change
• The baby Turns head when they hear the sound change in anticipation of the toy appearance
• If they do, they perceive sound change; if they don't, they do not perceive sound change
word segmentation
how to figure out when one word ends and the next word begins
segmentation strategies
these only work when an infant has already:
1) segmented sufficient #s of words to be able to use words themselves
2) segmented words in order to figure out stress patterns
3) aware of phonotactic constraints
-these help solve the infant bootstrapping problem
interactive process
it is believed that word recognition involves multiple levels of representation in which representations at each level interact with each other, and where interaction may occur between competing representations with a single level
parallel process
word recognition:
the representation of many words become active as a listener processes a particular auditory signal
bottom-up
_________ ____ processing OR signal-driven processing:
the info used to construct representations is determined entirely by the speech signal; small compositional levels determined by larger compositional levels (ex. a new word starts when the previous word ends)
-the Ganong effect shows that this process could not be the only thing, since phonological units are not perfectly recoverable from the speech signal
top-down
_____ _______ processing: representations at one level influence (ie. words) representations at smaller compositional levels (ie. phonological units)
-(word knowledge to phonemes/gestures to formants)
phoneme restoration effect
produced when a section of the speech signal corresponding to a speech sound (ie. [s]) is removed and replaced by white noise (static) or a cough
-demonstrates our ability to recognize words in degraded signals; top-down processing
close shadowers
those who, under the phoneme restoration experiment, responded even before the entire word was uttered, while others repeated back a corrected version
cohort model
_________ consist of:
cohort: all the words your mind is considering that match the bottom up signal so far (at any given moment)
-bottom up: gradually removes from the cohort words that do not follow the progression of the target word
-HOWEVER: the first phonological unit of a word is derived from top-down processing
word frequency
a measure of how frequently an average person would hear a particular word in everyday speech
- if a listener is trying to decide btwn 2 words being to potential target, the more frequent one is more probable to be the target
phonological neighborhood
words a listener knows that are "close" to the target word; for our purposes, there are words that differ in one phonological unit
neighborhood density
refers to how many phonological neighbors a target word has
lexical decision experiment
on average: participants responded to more frequent words about 90 milisecs faster than infrequent ones
cross-modeling priming
shows that frequency plays a role even earlier, when there are still multiple words that correspond to the signal processed so far, any of which could be the actual word that was spoken
- probe (seeing the word - say, "dog") is partially activated by hearing the prime (hearing a prompting word such as "bone" )
effects of phonological neighborhood
recognition is better if the phonological neighborhood is sparse rather than dense bc, for ex, bring a low-freq word:
having many phonological neighbors and having highly frequent neighbors makes recognizing words in noise DIFFICULT
competitiors
what we call phonological neighbors who compete for selection as the recognized word
Sets found in the same folder