Use LEFT and RIGHT arrow keys to navigate between flashcards;
Use UP and DOWN arrow keys to flip the card;
H to show hint;
A reads text to speech;
152 Cards in this Set
- Front
- Back
Amplitude
|
Maximum variation in air pressure
|
|
Pitch
|
The perception of frequency
|
|
Cycle
|
1 complete variation of air pressure
|
|
Frequency
|
the rate at which cycles occur
|
|
Human perception of sound range
|
16-20Hz to 20,000Hz
|
|
Speech
|
the medium through which we transmit linguistically encoded thoughts
|
|
2 Properties of all communication systems
|
Meaningful and ecological
|
|
4 Properties of some communication systems
|
Reciprocity, learned through interaction, arbitrary, compositionality
|
|
3 Properties particular to human language
|
Displacement, open-endedness, duality of patterning
|
|
Sound
|
A disturbance of the air that we sense through our auditory system
|
|
Acoustic waveform
|
record of sound-producing pressures over time
|
|
Periodic sounds
|
waveforms that repeat at regular intervals
|
|
2 types of periodic sound
|
Simple and complex
|
|
Simple sound
|
sine wave and pure tones - make up complex sounds
|
|
Complex sounds
|
combination of 2+ simple sounds repeating at regular intervals
|
|
How to calculate standard pitch ovtaves
|
multiply the frequency by 2x to go up 1 octave
|
|
Phase
|
timing of a waveform relative to some reference point
|
|
Fundamental frequency
|
the main repeating pattern of a complex waveform
|
|
White noise
|
the most complex sound possible - doesn't occur naturally
|
|
Harmonic
|
Any whole number multiple of the fundamental frequency
|
|
power spectrum
|
amplitude x frequency plot of simple components of a complex wave
|
|
Aperiodic sounds
|
sounds with no regular repeating pattern
|
|
Transient sounds
|
An aperiodic sound - sudden pressure fluctuations not sustained over time
|
|
IPA
|
International phonetic alphabet - set of symbols designed to provide a universally understood system for transcribing speech sounds
|
|
Phoneticians
|
transcribe speech sounds using the IPA
|
|
Consonants
|
a sound made when the vocal tract is constricted - defined by POA and MOA and voicing
|
|
POA
|
Place of articulation - where the closure happens
|
|
MOA
|
Manner of articulation - how the closure occurs (quickly, slowly, fully)
|
|
5 MOA
|
Stop, fricative, affricate, approximate, lateral
|
|
Stop
|
air stream is completely obstructed: m,b,t,g,
|
|
Nasal Stops
|
air is prevented (stopped) from going out the mouth but is allowed to escape through the nose: m, n, ng
|
|
Fricatives
|
When air is pushed through a narrow vocal tract and air flow is turbulent: s, sh, chi, th
|
|
Approximate
|
When the vocal tract is slightly narrowed but not enough to make the air turbulent
|
|
Articulators
|
part of the vocal tract used to form speech sounds - 7 kinds
|
|
7 articulators
|
Bilabial, labiodental, dental, alveolar, palato-aevolar, palatal, velar, glottal,
|
|
voiced
|
pronounce with vibration of the vocal cords: z,b,v,d,g
|
|
unvoiced
|
pronounced without the vibration of the vocal cords: s,p,f,t,k
|
|
2 tpyes of speech recording
|
Digital and analog
|
|
Digital speech recording
|
a discrete signal where the number of decimal places are always limited - loose info
|
|
Analog speech recording
|
a continuous signal where time and amplitude are represented by numbers that have a theoretically infinite number of places after the decimal
|
|
2 steps in analog to digital conversion
|
Sampling and quantization
|
|
Sampling
|
breaking up a sound wave into time chunks - need at least 2 samples per sine wave
|
|
Nyquist frequency
|
the highest frequency component that can be captured with a given sampling rate
|
|
Aliasing
|
A distortion in the signal that occurs when the continuous signal completes more than 1 cycle between successive digitized samples
|
|
What transport is used by the Na-glucose cotransporter in the SMALL INTESTINE??
|
Symport
Na OUT, Glu IN |
|
Quantization
|
measurements of amplitude - need to determine how accurately you want to measure it. more points = smoother curve
|
|
Quantization noise
|
Quasirandom noise introduced because of the step-like pattern of digital wave forms
|
|
Accuracy of quantization
|
number of bits used to store the waveforms - 16 bits is standard for human speech
|
|
Auto-correction pitch tracking
|
overlaying sucessive periods of a wave form there is correlation - can be used to find f0
|
|
Pitch-halving
|
when a computer is told to restrict pitch range 2 pitch periods are mistaken for 1
|
|
Pitch-Doubling
|
When a computer is told to restrict pitch range a 1/2 pitch period is mistaken for a full pitch period
|
|
Formants
|
Strong harmonics/resonant frequencies
|
|
IPA
|
International Phonetic Alphabet used to transcribe speech sounds of any language
|
|
What motion causes sound?
|
Opening and closing of the vocal cords
|
|
What makes 1 vowel different from another vowel?
|
Quality
Not pitch or loudness |
|
2 types of non-periodic sounds
|
aperiodic/transient and fricative
|
|
Andrew Carnegie
|
tried to convert all english spelling to IPA - spelling reform - didn't work
|
|
Gestural economy
|
tendency to use the same gestures in many different sounds ex. d and n have the same gesture but have different MOA
|
|
parallel transmission
|
speech can encode and transmit information about several linguistic events in parallel - very efficient
|
|
consonant
|
sound made with an obstruction of the vocal tract - defined by MOA, POA and voicing
|
|
Vowel
|
Sound made with no obstruction of the vocal tract - can be defined by pattern of their 1st 2 formants
|
|
What are vowels generally louder than consonants
|
because there is no constriction of air flow. more air = more amplitude
|
|
VOT
|
Voice onset time - time from release of a stop to onset of voicing
|
|
aspiration
|
when air rushes through the glottis without vibration - produces 'h' sound
|
|
Why can we never hear unmodified vocal cord sound?
|
Because by the time it reaches the lips it has been shaped by the vocal tract
|
|
Resonance
|
the way in which objects vibrate when they have been set into motion
|
|
Segmentals
|
the vowels and consonants of speech
|
|
Suprasegmentals
|
information superimposed on top of speech segmentals
|
|
Source-filter model of vowel production
|
source of sound = vocal cords, filter = vocal tract. the independence of source and filter allow changes in amplitude and pitch to produce the same vowel
|
|
Diphthong
|
A vowel whose qualities change from start to finish
|
|
Monophthong
|
A vowel whose qualities don't change from start to finish
|
|
How to characterize vowels
|
By F1 and F2 or POA
|
|
Consonants
|
Sounds made with an obstruction of the vocal tract. Defined in terms of MOA, POA and voicing
|
|
Manner of Articulation
|
Stops, Nasal Stops, Fricatives, Affricates, Approximates,
|
|
Stops
|
MOA in which the airstream is completely obstructed
Ex. M, B, T, G |
|
Nasal Stops
|
MOA in which the air is prevented from going out the mouth but is allowed out the nose. Ex. N, M, NG
|
|
Fricatives
|
A MOA where the vocal tract is narrowed, which leads to turbulent airflow - a hiss sound close to white nosie
|
|
Approximate
|
MOA where the vocal tract is only slightly narrowed
|
|
Affricate
|
MOA that involves 2 places of articulation, followed by a fricative.
|
|
Place of Articulation
|
Bilabial, Labio-dental, Dental, Alveolar, Palato-Alveolar, Palatal, Velar
|
|
Articulator
|
Parts of the vocal tract that are used to form speech sounds
|
|
2 Modes of voicing
|
Unvoiced and voiced
|
|
2 types of speech recording
|
Analog and digital
|
|
Digital speech recording
|
a discrete signal where the number of decimal places are limited - speech signal = dotted line
|
|
Analog speech recording
|
A continuous signal where time and amplitude are represented by numbers that are theoretically infinite
|
|
2 steps in analog to digital conversion
|
Sampling and quantization
|
|
Sampling
|
1st step in analog to digital conversion. Decide number of points on the time axis - need at least 2 per cycle
Sampling must be 2x the cycle rate |
|
Quantization
|
The 2nd step in analog to digital conversion. Deciding how accurately amplitude is to be measured. More point = smoother curve
|
|
Nyquist frequency
|
In sampling the highest frequency component that can be captured given the sampling rate. = 1/2 the sampling rate
|
|
Aliasing
|
a distortion in the sound signal that occurs when a continuous signal completes more than 1 cycle between successive digitized samples
|
|
Accuracy of quantization
|
The number of bits used to store the waveform samples. For speech 16 bits is standard
|
|
Quantization noise
|
Quasirandom noise generated from quantization because digital noise is step-like rather than a curve
|
|
Auto-correction Pitch tracking
|
When overlaying sucessive periods there is a high degree of correlation which can be used to find f0 of a waveform
|
|
Pitch-halving
|
when restricting the pitch range on a computer 2 pitch periods are mistaken to be only 1 and the f0 is halved
|
|
Pitch-doubling
|
when restricting the pitch range on a computer 1/2 a pitch period is mistaken for 1 and f0 is doubled
|
|
Formants
|
resonant frequencies of the vocal tract. formants change each time vocal tract configuration changes
|
|
What makes 1 vowel different from another vowel
|
Quality: the frequencies and amplitudes of the component waves
|
|
3 components used to describe a sound wave
|
Amplitude, frequency and phase
|
|
Power spectrum
|
amplitude vs. frequency plot of the simple sine wave components of a complex wave
|
|
2 types of sound
|
Periodic and non-periodic
|
|
2 types of periodic sound
|
Simple and complex
|
|
2 types of non-periodic sounds
|
aperiodic/transient and fricatives
|
|
Transient/aperiodic sounds
|
sounds resulting from sudden fluctuations in air pressure that are not sustained over time
ex. balloon popping or stop consonants /b/ and /p/ |
|
Andrew Carnegie
|
tried to convince the president that there needed to be a spelling reform to IPA
|
|
Gestural Economy
|
tendency to use the same gestures in many different sounds
|
|
IPA Transcription
|
qualitative description of speech done without relying on orthography
|
|
2 ways of describing speech sounds
|
IPA Transcription and Acoustic recording
|
|
Gestures
|
Speech tasks (ex. closing the lips) that are coordinated in a certain way to produce speech sounds
|
|
Parallel trasmission
|
In speech when gestures are continuous and overlap (co-articualted). A very efficient way to transmit information
|
|
Why are vowels generally louder than consonants?
|
Because there is no contriction in airflow which leads to more air and higher amplitude
|
|
3 possible orderings during a nasal stop
|
1. Closing the vocal tract at the velum
2. Lowering of the velum to let air through 1 --> 2; 2 --> 1 or 1 and 2 at the same time |
|
Diacritic
|
IPA symbol used in narrow transcription
|
|
VOT
|
time from the release of a stop to the onset of voicing. Can be positive, negative or zero
|
|
Vowel production
|
Source: vocal cords vibrating in a periodic fashion. f0 = number of glottal pulses per second
|
|
Why can we never hear the unmodified sound of vocal cord vibration
|
By the time vocal cord sound reaches the lips it has been modified by the vocal tract
|
|
If we could hear unmodified sound of the vocal tract, what would it look like?
|
An f0 with a 2nd, 3rd, 4th… harmonic. The lower harmonics will have greater intensity that the higher harmonics
|
|
Resonances
|
the way in which an object will vibrate when its been set into motion
|
|
Segmentals
|
the vowels and consonants of speech
|
|
Suprasegmentals
|
information superimposed on top of speech segmentals
|
|
Source-filter model of vowel production
|
noise generating Source = glottis and filter that attenuates and amplifies harmonics = vocal tracts. Both are independent of each other
|
|
What are the vertical striations in a spectrogram?
|
opening/closing of the glottis. 1 cycle = 1 strip
|
|
Independence of source and filter
|
the rate of vocal cord vibration is independent of the configuration of the vocal tract - f0 and harmonic may change, but the formants remain the same
|
|
How are vowels characterized?
|
By their f1 and f2 values (and sometimes by f3)
|
|
4 ways of crowding f/f2 space
|
1. Rounded vowels
2. Diphthong vs. Monophthong 3. Lax vs. Tense 4. Vowel length |
|
Vowel in english that is characterized by f3
|
the vowel in 'bird' - 1st 2 formants look like the verb in 'hood'
|
|
Why do the vocal folds get blown apart only to be sucked back together (2 reasons)
|
1. Bernouli Effect
2. Elasticity |
|
Velar pinch
|
in Velar sounds (g/k) f2 and f3 are pinched together at the ends
|
|
What makes voices uniques?
|
Environment and anatomy
|
|
What do voiced sounds look like in a spectrogram?
|
vertical striations
= vibration of vocal cords |
|
What do bilabial sounds look like in a spectrogram?
|
formants are generally lowered
|
|
What do velar sounds look like in a spectrogram?
|
f2 and f3 pinched
|
|
What do stop sounds look like in a spectrogram?
|
gap in the pattern, followed by a burst of noise
|
|
What do nasal sounds look like in a spectrogram?
|
formants similar to a vowel, but lowered
|
|
What do approximate sounds look like in a spectrogram?
|
Formants have similar energy to a nasal but move (inclined or declined)
|
|
Prosodic information
|
another term for suprasegmental information
|
|
Prosody
|
the melody and rhythm of language - stress and intonation. The first thing our auditory system can perceive
|
|
Why use spectrograms?
|
Allow the clear visualization of formant structure of speech and show formants over time
|
|
What does tensing the vocal cords do?
|
increase the pitch by making the vibrate faster - increase sub-glottal pressure and increase the Bernouli effect
|
|
Breath group
|
section of an utterance between 2 breaths
|
|
What does increasing sub-glottal pressure do?
|
causes the vocal cords to open and close more quickly, which leads to an increased f0 (pitch) and louder sounds
|
|
Intonation
|
The use of melody to change the meaning of groups of words (in english) and words in other languages
|
|
What does an increase in vocal fold tension do?
|
Increase pitch and sub-glottal pressure
|
|
3 Possible Rhythm classes of languages
|
Stress times, syllable times and mora timed
|
|
Stress time languages
|
Temporal organization is around the stressed syllable - stressed syllables are said at regular intervals
|
|
Syllable timed languages
|
Temporal organization is around the syllable - each syllable is said at regular intervals.
|
|
Why are psycholinguistic studies not done on just written language anymore?
|
not everyone who communicates can read
Not all languages are in written form Not all information is transmitted in writing |
|
4 Effects of Prosody
|
1. segmenting words from speech
2. disambiguating syntactic structure 3. Discourse structure 4. emotional prosody (happy, sad, sarcastic) |
|
Segmenting words from speech
|
Done with prosody. word and segment duration are effected by the placement in prosodic strucutre
|
|
In prosodic structure monosyllabic words are longer/shorter than polysyllabic words?
|
monosyllabic words are longer
|
|
Slips of ear
|
errors is sensitivity to stressed syllables that blur word boundaries
|
|
Words at the end of phrase boundaries
|
Words at the end of phrases are lengthened
|
|
Parity
|
achieved when the intended thoughts of one persons are understood by another person
|