Audition Flashcards
(112 cards)
1
Q
sound def
A
information carrying system
2
Q
noise def
A
random unwanted data that corrupts signal
3
Q
transverse wave
A
movement is at right angles to propagation (e.g. waves in wateR)
4
Q
longitudinal
A
movement is parallel to propagation (e.g. sound)
5
Q
sound physics
A
air molecules uniformly distributed, vibrating object disturbs uniformity,
6
Q
how does an object vibrate
A
objects have elasticity, a force pushes it -> mass spring system
7
Q
mass spring system
A
mass (of object)
spring (stiffness of)
=frequency of vibration
8
Q
what mostly determines dif sounds
A
physical properties (relating to spring stiffness).
you can hit a drum with dif forces but will sound the same.
9
Q
speed of sound in dif mediums?
A
solid -fastest bc molecules tightly packed
liquid
gas - slower
10
Q
Speed of sound in air dependent on
A
temperature of gas
molecular weight of gas
11
Q
how to change freq and amplitude
A
more displacement = higher amplitude
more stiffness = high freq
more mass = low freq
12
Q
sine wave equation and parameters
A
A * sin (2π f t + ψ)
t = time A = amplitude f = frequency ψ = phase
13
Q
features of a sine wave
A
2π = 360deg (one full sine wave)
frequency in units of Hertz (one full cycle)
14
Q
pure tone vs real objects
A
a pure tone gives a single sound, defined by one sinusoid, but most objects vibrate at dif freq as a combination of pure tones
More formally: they vibrate at a Fundamental resonant frequency (F0) and at
harmonics of that frequency: 2 x F0, 3 x F0,
4 x F0, and so on.
15
Q
harmonics
A
multiples of the fundamental frequency
16
Q
timbre
A
Perceptual quality of sound related to freq spectrum and amplitude envelope
17
Q
spectrum? envelope?
A
frequency spectrum: complex patterns of vibrations across many spectrums
amplitude envelope: responsible for dif sounds, has an onset, a steady state and an offset.
18
Q
Fourier Analysis
A
F proposed that any signal can be describes as the sum of a family of simple sine waves
Therefore, we should be able to decompose complex sounds into three parameters: amp, freq, phase
19
Q
range of hearing examples?
A
30 dB SPL - library
140 dB SPL - gunshot
20
Q
how to decompose a signal using fourier analysis?
A
First, can try to pick a sinusoid waveform that best fits the sound wave.
Second, can pick another sinusoid, with higher frequency.
Third, can combine the two, to replicate the wobbly sinusoid as best as you can
Repeat until you get it really really close.
21
Q
what is white noise?
A
random noise that has a spectrum that is flat, therefore contains all possible frequencies of sound, and all these frequencies have the same amplitude
22
Q
limitations of a strict Fourier understanding
A
based on the assumption that signals go on for an infinite time to generate sounds.
But our sensory systems process signals over a finite time window with limited frequency bands, so strict fourier analysis is not biologically realistic.
23
Q
how does sound travel through air?
A
- a vibrating object causes air molecules to vibrate
- increase in conc of molecules in one place (compression) and reduced elsewhere (rarefaction)
- high-pressure region moves to low pressure region, causing sound to travel through the air
24
Q
what is the inverse square law?
A
as sounds moves -> amplitude decreases
the amplitude of vibration is decreased in proportion to the square of the distance
25
so, amplitude changes with distance, what about freq?
affects high frequencies more than low frequencies, as they are absorbed into the surrounding air.
E.g. crack of lightnigh when near (low freq AND high freq sound) and like a low rumble of thunder when far (mostly low freq as the high has been absorbed)
26
db SPL
decibel sound pressure level
27
what is a decibel?
a ratio value of sound pressure relative to the lowest pressure we can hear
28
minimum sound level detectible for average person?
20 micropascal
29
break down of the ear into parts
```
outer ear (pinna and meatus)
middle ear
inner each (cochlea)
```
30
(1) how do vibrations move into the cochlea?
- vibrations channelled by outer ear to the ear drum
- ear drum moves due to interior vs exterior pressure differences
- Ear drum and ossicular chain act as an impedence matching device
- translates vibration into larger amplitudes and movement of the oval window moves the perilymph fluid in the cochlea
31
what is the cochlea
a snail-like structure in the inner ear divided into three parts:
- scala vestibuli
- scala typani
- scala media
32
ossicular chain
3 small bones:
- the malleus
- incus
- stapes (hits against oval window)
33
why do we need an impedance matching device?
how does it work?
-cochlea is filled with fluid (perilymph) that has a much higher impedance than air, so sound needs to be amplified
- ear drum area much bigger than stapes; concentrating same force over much smaller area
- hinge like chain provides leverage to amplify
34
what is the acoustic reflec?
muscles of ossicles have a reflex to disengage the stapes during very loud sounds (but is a bit slow)
35
measurements of cochlea
2.5 turns in spiral
3.5cm in length
2mm diameter
36
detail on 3 parts of cochlea
scala vestibuli and tympani are inter conneted at the apex (top) of the cochlea (at the helicotrema) and are filled with perilymph
scala media is much smaller, filled with potassium and ion rich fluid endolymph
37
(2) what happens to the BM as the the fluid moves?
- perilymph fluid in cochlea cannot compress, so the pressure causes the round window of cochlea to bulge out, and oval window push in
- movement causes vibration of the BM at different locations (tonotopy)
38
dif locations of tonotopy and why?
low freq sounds displacement at apex (wide and floppy)
high freq sounds cause maximum displacement at the basal end (narrow and stiff)
39
how is tonopy achieved?
- varying stiffness of the BM
| - Inertia of the mass of perilymph
40
organ of corti
- located in scala media
- auditory hair cells are on the BM side
- BM moves -> shearing motion between base of organ of corti and tectorial membrane = movement of stereocilia of the hair cells
41
two types of hair cells
- inner
| - outer
42
sterocilia
on top of the hair cells, form form rows connected by tip links.
43
how to move from vibration to voltage?
- movement in one direction --tip links cause open ion channels in the stereocilia
- the endolymph in scala media has K+
- K+ moves into cell and depolarises it
- release of glutamate
- triggers action potential in spiral ganglion cells
44
spiral ganglion cells
have long axons to the auditory nerve (the VIIIth cranial nerve) to the cochlea of the brainstem
45
technical term for vibration to voltage
the vibration is transduced into an analogue electrical signal
46
'place theory'
by Helmholtz -> says that freq of sound indicated by place along cochlea where activity is highest
47
early evidence for place theory?
- Von Bekesy drilled small hole in cochlea, inserted silver particles, then saw the movement of these particles with a microscope.
- By applying different pure tones, he was able to see how specific frequencies create maximum displacement at specific places on the BM.
- evidence that BM acts like a mechanical frequency analyser
48
frequency and fluid dynamic to BM
low freq sounds see less resistance from perilymph fluid, and move through more fluid to apex.
high freq see more resistance, travels through less fluid to the basilar membrane.
49
issues with Helmholtz's place theory
1. better technology (e.g. lasers) have found more precise measurements
2. pitch perception cannot be explained by place alone, as the timing of nerve firing plays a role.
- neurons can't continuously fire at high frequencies over 1000 Hz as they need time to recover
- however, a group of cells can, known as phase locking
50
who introduced phase locking? problems?
known as the 'volley principle' (Weer and Bray, 1937)
when neighbouring cells track frequencies together
only works up to frequencies of 5,000 Hz, at which point sound becomes strange
51
electronic speakers basic filters
- highpass: removes/attenuates below a certain cut-off
- lowpass: removes/attenuates freq above a certain point
- bandpass: upper and lower cut off point
- band stop: attenuates freq between the cutoff points
52
which filter is relevant to our auditory system?
band pass - our hearing system is a bank of overlapping bandpass filters, covering the range of audible frequencies 20-20,000 Hz
53
fletcher
- fletcher (1940) classic experiment on masking (when one sound is less audible bc of another sound)
- took band pass filtered noise and presented a single pure tone in the middle of it.
- as the width of the band pass filter increased, performance got worse (listeners needed more sound of the pure tone to hear it)
- but once the width exceeded a certain amount, performance did not get worse but plateau.
- this bandwidth was called the 'critical bandwidth', and indicates the width of the auditory filter (and the pure tone passes through only one filter)
54
classic experiment on shape of auditory filter?
- Vogten (1974) used found by psychoacoustic tuning
- presented a pure tone with a weak signal (10db above detection threshold) with masker
- varied the freq content of the masker
- tested at a range of frequencies
- found that the shape of an auditory filter is more like a U shape
55
problems with vogten (1974) work?
Filters may actually be wider as participants might have used more than one channel (known as off frequency listening)
Happens because auditory filters are asymmetric and have steep high-frequency slopes
56
problem with fletcher (1940) work?
assumes auditory filters are rectangular
57
outer hair cells function
- actively amplify the sound
- when stimulated they contract to pull tectorial membrane with them
- have connection to a singular ganglion cell to pool responses from several hair cells
58
what does the outer hair cells amount of amplification depend on
- it amplifies low intensity sound a lot (50db) but as the intensity increases not so much.
- this is why it is ACTIVE amplification, it is non-linear
59
outer hair cells super protein thing
-motor protein (PRESTIN) contract over 70,000 times per second (demonstrated by Ashmore on the hairs of guinea pigs)
60
this active process of amplification
allows it to add new freq information! weird
makes the relationship between input and output curved (a compressive nonlinearity)
61
otoacoustic emissions
sound freq in output that weren't in input
62
damage to outer hair cells
reduces active amplication to make hearing hard
63
inner hair cells function
- help with fast transmission of sound info to brain
- each cell has synaptic connections to around 20 ganglion cells
- Successive hair cells differ in frequency by 0.2%
64
ganglion cells three types
- high spontaneous rate axons respond to low intensity sounds, but they at ~ 40 dB SPL
- medium and low spontaneous rate axons, do not respond until 20-30 dB SPL, but saturate at ~80 dB.
65
axon response gains?
v different, some respond better to low, some better to high
therefore, axon is most responsive to a characteristic frequency
66
action potentials are stochastic events
i. e. there is a degree of randomness to them
i. e. when the acoustic stimulus is at its max amplitude, the neuron fires, but not every time!
there is simply a high CORRELATION between the firing rate of individual neurons and the stimulating sound frequency
67
is the cochlea active or passive?
The movement of the BM of a live healthy cochlea is greater and sharper than found in von Bekesy's cadavers. Further evidence for an 'active mechanism' which amplifies and sharpens the tuning of the auditory filters.
68
WHY are action potentials stochastic events
they need time to recover, can't continuously fire
69
place hypothesis vs temporal hypothesis
Pitch related to place of max response on BM
Predicts tuning curves not affected by freq above 4 Hz (no change in performance)
Pitch related to time intervals between action potentials (i.e. temporal firing rate).
Predicts phase locking breaks down above 4Hz (performance gets worse)
70
early evidence for temp hyp
Moore (1973) asked people to distinguish between two pure tones of similar freq.
- best in range 0.5-2kHz.
- much worse after 5kHz (logarithmic)
71
later evidence for temp hyp
Attneave & Olson (1971)
Melodies above 4 kHz do not give rise to a clear sense of pitch
Therefore, temporal coding may be critical
72
pitch related to
frequency
73
loudness related to
amplitude
74
pure tone form
sinusoidal
75
harmonics meaning
an overtone accompanying a fundamental tone at a fixed interval, produced by vibration of a string, column of air, etc.
76
what is fundamental freq determined by? how to show this?
present complex tone where F0 is 100Hz, F1 is 200Hz, etc
listeners would say this complex tone matches a 100Hz pure tone
i.e. pitch is determined by fundamental frequency
77
fundamental freq evidence for temporal hyp
present complex tone where F0 is 200Hz, F1 is 300Hz, etc
listeners would say this complex tone matches a 100Hz pure tone
300 is not an integer of 200, so auditory system resolves this, this is called the PITCH OF THE MISSING FUNDAMENTAL
78
why does the pitch of the missing fundamental not work with place theory
because listeners hear a pitch which is not seeing maximum displacement of the BM
79
what are the two loudness hypotheses
firing rate vs. number of neurons
80
what are Psychophysical loudness matching experiments?
adjust the amplitude of a 1000 Hz pure tone so it matches the loudness of a test sound
81
what do psychophysical loudness matching experiments show
Psychophysical loudness matching experiments reveal ‘iso-loudness’ contours (that graph thing)
shows loudness is a function of both frequency and sound pressure level
82
comparing pitch and loudness
The relative contributions of firing rate and number of neurons firing may well change at different frequencies, just as the ability to sense pitch may depend on both place and rate type codes for
different frequencies
83
from cohlea nucleus to...
- superior olivary complex (SOC) in brain stem
- inferior colliculus
- medial geniculate nucleus
- primary auditory cortex
84
contrast to visual system in integrating two ears?
visual system does not integrate info until it reaches the cortex
85
SOC role
superior olivary complex (SOC) integrates info from both ears
known as interaural exchange
86
inferior colliculus
involved in localisation
87
medial geniculate nucleaus
connects two
88
primary auditory cortex
is tonotopically organised in a way that mirrors the BM
89
cues for localisation
- Interaural intensity difference (IID)
| - Interaural timing difference (ITD)
90
where is the middle of the head
intersection between interaural axis and midline
91
how to construct a sphere
around this middle point
92
how to describe objects from this point
azimuth
elevation
distance
93
distance between ears in adult
18 cm apart
94
acoustic shadow
sound blocked by the head, head reflects/diffracts the sound
95
magnitude of shadowing depends on
the frequency of the
sound relative to the size of the head
high frequency sounds get reflected easily by
obstacles (LARGE EFFECT), while low frequency sounds get diffracted around it (low effect)
96
ITD evidence
Lord Rayleigh used two tuning forks to show that timing differences alone were enough to create impression of a sound coming from one location.
97
rayleigh theory
duplex theory (1907)
listeners use two dif cues for difference frequencies.
timing differences are = low frequencies
intensity differences = high sound frequencies
98
why duplex theory?
timing dif bad at high frq because of ALIASING problem (high freq ambiguous and brain doesn't know which sound peaks match between the ears)
level dif bad at low freq because there is little acoustic shadowing
99
what degree of accuracy?
differenced for ITDs and IID
psychophysical experiments using headphone presentation
1 degree of azimuth accuracy, which translates to
10μs (microsecond) time dif
0.5dB dif
100
how can we compare cues in study
can be combined together or put in opposition
101
which cue is more powerful when in opposition?
depends on freq
102
what do we use in normal life?
combination of both
limited evidence for individual differences in people.
103
What are the neural mechanisms that support sound localisation?
- work in barn owls
- different portions of the superior olive complex
- this is the first place where signals from the two ears merge
104
what are the dif portions of the Superior Olive Complex
Medial Superior Olive:
-timing differences
-individual cells appear to prefer particular delays
between the left and right ears.
Lateral Superior Olive:
- intensity differences.
- Contrasts the firing rates of inputs from the left and right cochlear nuclei.
105
how are humans so responsive to tiny time differences?
- Jeffress (1948)
- axon connections between left and right ear
- sound travels through delay lines at a finite speed
- coincidence detector fires when it receives inputs from both left and right ears
- if sound if from straight ahead, meets in the middle, if from left or right, meets there.
- shows azimuth
106
support for Jeffress?
used ITD tuning curves on owls
107
evidence against/sharpening Jeffress thoery
McAlpine (2005)
- sharply tuned neurons for birds (owls)
- broadly tuned neurons for mammals (gerbils)
The code for birds is a place code because the ITD is indicated by firing of neurons at a specific place.
The code for mammals is a distributed code because the ITD is determined by the firing of many broadly tuned neurons working together.
108
mammal distributed code
-visual system signals different wavelengths by signalling the pattern of response of three different cone pigments
109
ways to deal with ambiguity
- move head
| - head related transfer function
110
head related transfer function
- pinna cues, the shape of outer ear
- shape of our heads and upper torso
Diffractions/reflections from these create interference effects. They modify the sound spectrum in a way that is unique for each direction.
111
pinna cues study 1
Batteau (1967) gave people recordings where microphones were in casts of a listeners ears.
They perceived the sound as from coming from a location away from the head.
112
pinna cues study 2
- Gardner and Gardner (1973)
| - showed that azimuth discrimination was much poorer if the pinnae were filled to remove their filtering effects.
