ling313 midterm

Fundamental frequency

the rate at which the complex waveform pattern repeats

Harmonics

Complex periodic waves broken down into simple waves, each of these has a frequency that is a multiple of the fundamental frequency

Relationship between fundamental frequency and harmonics

Harmonics is a multiple of the fundamental frequency

Calculating F0 and harmonics based on another

Frequency = harmonic number X f0

How is volume measured?

cm cubed

Nyquist frequency

The highest-frequency component that can be captured within a given sampling rate

Sampling rate

The number of times per second that we measure the continuous wave in producing the discrete representations of the signal

Calculating relationship between sampling rate and nyquist frequency

Nyquist frequency is always half the sampling rate

Broadband spectrogram

Prioritizes temporal information by sampling at smaller time intervals, formants are visible

Narrowband spectrogram

Prioritizes spectral information by sampling at larger time intervals, harmonics are visible

Pressure measured in

cm H20

Cm H20

The amount of pressure needed to support a column of water

Boyle's Law

Assuming temperature does not change volume and pressure are inversely proportional, ie. pressure goes up and volume goes down and vice versa

Source-filter theory

A theory of speech acoustics in which the vocal tract is analyzed as a series of band-pass filters

Source

The origin point of the noise

Filter

That which shapes the noise

Glottal source

Provides a complex (quasi-) periodical signal through modes of vibration by modulating an air stream, produces the fundamental frequency and an infinite number of harmonics

High-pass filter

Let high frequencies go through

Low-pass filter

Let low frequencies go through

Band-pass filter

Let frequencies within a range go through

Examples of sources

Vocal fold vibration, F0, harmonics

Examples of filters

Vocal tract, resonant frequencies of tube, formants

Natural resonance

All objects oscillate at greater amplitude in response to certain frequencies

Resonance is dependent on

Size, shape, and what the object is made of

Low resonant frequency

Big, wide, long or heavy objects vibrate slowly

High resonant frequency

Small, narrow, light or short objects vibrate quickly

Four key processes in articulation

Articulatory process, phonation process, oro-nasal process, airstream process

Articulatory processes

Part of the phonological loop that repeats sounds or words to keep them in working memory until they are needed: place and manner of articulation

Phonation processes

Actions of the vocal folds: voiced vs. voiceless, voice modality

Oro-nasal processes

Refers to airflow directed through either the oral or nasal cavities: nasality

Articulation process

Place of articulation and manner of articulation position the vocal tract into a position that shapes airflow, they create conditions for sound generation and sound shaping

Airstream process

Air movement provides the power to make noise in speech, the power that allows us to make noise

The most basic air mechanism

Pulmonic egressive

How are pulmonic egressive sounds created?

Modulating the flow of air coming out of the lungs

3 airstream mechanisms

Lungs (pulmonic), glottis (glottalic), tongue (velaric)

Pulmonic sounds

[u↓] or [s:↓]

Forming pulmonic ingressive

1) oral closure

2) lungs expand and pull air into the supra-glottal cavity, decreasing pressure

3) oral closure is released

4) result: air flows inward

Glottalic egressive (ejective) sounds

[p' t' k' q']

Glottalic ingressive (implosive) sounds

[ɓ ɗ ɠ ʛ]

Velaric sounds

[ʘ | ! ǁ]

Forming velaric sounds

1) velar closure and an anterior closure further forward in the oral cavity

2) lower the tongue body, while maintaining a sealed pocket of air; increasing volume, decreasing pressure

3) release anterior oral closure

4) release velar closure

Can clicks be nasalized?

No because you cannot lower the velum to get the air out

Can glottallic egressives (ejectives) be nasalized?

Yes because the velum is not being used

Longitudinal wave

Particle motion on the same axis as the direction of wave travel, sound waves

Transverse wave

Particle motion perpendicular to the direction of travel

Periodic sounds

Regular repeating patterns in waveform, always voiced sounds, quasi-periodic (vibration is not constant), vowels, approximants, and nasals

Aperiodic sounds

No clear repeated pattern, "shaggy" or "hairy" appearance, noisy acoustic percept, no periodicity so cannot measure F0, fricatives, devoiced sonorants, voiced fricatives (when periodic and aperiodic combine)

Transient

Sudden pressure fluctuations, stops

Impulses

An idealized transient where there is one pressure fluctuation at a single point in time, don't occur naturally (e.g. door slamming)

F1 related to

tongue height (high-low1)

F2 related to

tongue frontness/backness (B-L2)

Base

Closer to bone chain, thick end that responds to high frequency components, voicing, periodic structures

Apex

Thin end, responds to low frequencies, turbulence, transient/aperiodic sound waves

Low frequencies travel _ than high frequencies

further

How do frequency and intensity interact?

Loudness, sometimes measured in so-called sones

Pitch: steps perceived as equal

100 Hz -octave-> 200 Hz -octave-> 400 Hz -> 800 Hz

Pitch: steps are not perceived as equal

100 Hz -octave-> 200 Hz -5th-> 300 Hz -4th-> -> 400 Hz

How do we deal with non-linearity?

Transformations based on cochlear space, equivalent rectangular bandwidths, mel scale, bark frequency scale

Formula for tube closed on one end and open on another

Fn = (2n-1)c/4L

Formula for tube closed on both ends

Fn = nc/2L

Laminar airflow results in

Vowels

Laminar airflow

• Volume of lungs decreases at a certain rate (this volume change/movement is the initiator velocity)

Constriction near a point of maximum velocity ___ the formant frequency

lowers

Antinodes

Points of maximum velocity/minimum pressure

Nodes

Points of minimum velocity/maximum pressure

Constriction near a point of maximum pressure ____ the formant frequency

raises

can vowels be voiceless

yes

apical

tip of tongue

laminal

blade of tongue

describe manner of articulation in terms of the nature of a constriction

MA describes how the airstream is constricted in the vocal tract (type and degree of constriction made by articulator)

difference between fricatives and approximants

fricatives force air through narrow opening vs. wide → approximants are smoother w/o turbulence

what makes tap/flaps and trills unique compared to other pulmonic consonants

brief, rapid contact (vibration) rather than sustain closure or narrowing

coronal stops

made by closing tongue tip/blade on the upper front surface of the hard palate

tap production + example

up and down movement

e.g. potty

flap production + example

passing movement from behind

e.g. party

what is the black area in static palatography images

the point of contact (for dental sounds, teeth will have darkness)

uses of cardinal vowels

• set of reference vowels

• provide framework for comparison for vowels of languages (e.g. a vowel halfway between CV 8 and 9)

nasals in spectrogram

high F1, low F2

vowels in waveform

peaks

3 components of sound

source, medium (air), receiver

why is phase important

because waves combine

speed of sound

35000 cm/s

what happens if Nyquist frequency isn’t met

aliasing: misrepresentation of a signal

Ruben’s Tube

frequencies that resonate in a tube set up a standing wave (these are the waves that fit in a tube)

average vocal tract volume

3000 cm3

creaky vs breathy voice

• creaky (laryngealization): vocal folds pulled tightly, no vibration

• breathy: vocal folds loose, vibration

continuous vs discrete signals

• continuous: analog (sound waves, vinyls, photos from real camera)

• discrete: digital (recordings of sound)

standard bit rate

16 bit

wavelength =

c/f

cycle and period

cycle is each repetition of a pattern, period is duration of the cycle

F =

1/T (sec)

phase

measured in degrees, difference between 2 particular states in the same sound wave

aerodynamic requirements for voicing

subglottal pressure must be slightly greater than supraglottal pressure to get air flowing fast enough

temporal vs spectral frequency

• temporal = rate of repetition

• spectral = range of frequency a signal contains (spacing between adjacent harmonics)

role of basilar membrane in frequency perception (base and apex)

• basilar membrane is in cochlea; nonlinear

• base is thin end: responds to high frequency components

• apex is thick end: responds to low frequencies

why are standard spectrograms not reflective of cochlear spacing + alternative

• they plot frequency linearly

• alt: cochleagrams and neurograms

• use perceptual scales like Mel and Bark

key feature of a rhotic

lowering of F3 —> amount lowered depends on manner of producing

nasals vs laterals

• look very similar

• but formant spacing is wider in laterals than nasals because nasal tube length is longer (smaller tube = higher freq = multiples spaced out further)

• nasals have more reduction of amplitude (larger tube configuration)

sones

perceived loudness

how does rounding lips affect vocal tract tube

makes it longer