Transcript Pitch - Department of Psychology

Psy280: Perception
Prof. Anderson
Department of Psychology
Audition 1 & 2
1
Hearing: What’s it good for?


Remote sensing
Not restricted like visual field

Can sense object not visible
2
Hearing: The sound of silence

A tree in the forest


One hand clapping



No physical signal, no perception
Separate physical quantity from perceptual
quality
Sound is the perceptual correlate of the physical
changes in air pressure


Physical signal but no perception
Or water pressure when under water
John Cage’s 4:33 No. 2, 1962
3
What are the physical attributes
associated with sound?

Loudness


Amplitude or height of pressure wave
Pitch

Frequency of times per second (Hz) a pressure wave
repeats itself
4
What is sound quality?

Pure tones



Single frequency (f)
Rarely exist in real world
Complex tones



More than one f
Due to resonance
Air pressure causes reverberations


E.g., tuning forks
E.g., Plucking the A string on a guitar


Fundamental frequency 440 Hz (cycles/s)
Harmonics




Reverberations at multiples of the fundamental
E.g., 880, 1320
Creates fullness of complex sounds
Timbre is the relative amplification of harmonics
5
The human ear

Outer ear




Focusing of sound
Resonance amplifies 20005000 Hz range
Converts from air to
mechanical vibration
Middle ear

Amplification




Fluid denser than air
Focus vibrations onto
stapes/oval window
Increased leverage from
ossicles
Inner ear

Sensory transduction



Physical to neural energy
Fluid pressure changes
Bending of hair cells
6
Auditory sensory
transduction: The inner ear

Cochlea



3 layers
Cochlear partition


Contains organ of corti
Organ of corti



Coiled and liquid filled
Cilia (hair) cells
Between basilar and
tectorial membranes
Transduction

Movement of cilia
between membranes
7
Auditory transduction

Bending—>physical energy
Converted to neural signals

Bend one direction —> depolarization



More likely to fire AP
Other direction —> hyperpolarization

Less likely to fire AP
8
Auditory pathways
QuickTime™ and a
GIF decompressor
are needed to see this picture.
9
Audition: What and where

What is it?
*Pitch
 Identification



Surprisingly, little is
known beyond speech
Where is it?

*location
10
What: Pitch

How does neural firing signal different
pitches?
1) Timing codes
 2) Place codes

11
Pitch: Temporal coding


Idea: Diff f’s signaled
by rate of neuronal
firing
Hair cell response



Bend one direction —>
depolarization
Other direction —>
hyperpolarization
Result?

Bursting pattern of
neural response
related to frequency of
oscillation
12

Problems with temporal
coding
Problem: A single neuron can’t fire at the rate necessary to
represent higher f tones



Solution: volley principle



E.g., 1000-20,000 Hz (i.e., 1000-20000 per second)
Max neuron firing rate: 500-800 per second
No single neuron represents f
Coding across many neurons with staggered firing rates
Evidence: Phase locking
Diff neurons respond to
diff peaks
 Not every peak
 Pool across multiple neurons to
represent high f’s

13
Pitch: Place coding


Related to doctrine of specific
nerve energies
What is pitch?


Activation of different places in
auditory system
Frequency specific

Tonotopy




Owl
brainstem
Cochlear
Brainstem
Cortical
Stimulate these regions
 Should result in pitch
perception
Human auditory cortex
14
Place coding starts in cochlea

Von Bekesy studied
basilar membrane in
cadavers




Observed traveling waves
Diff frequencies (f) result
in waves w/ diff envelopes



Base more narrow and
stiffer
Apex wider and more
flexible
Higher f: Peak closer to
base
Lower f: Peak closer to
apex
Thus, f related to “place”
where peak fluctuation
15
Frequency tuning:
Neural place coding

Tonotopic arrangement of hair cell nerves


Diff nerves innervate diff parts of basilar membrane
Allows for “place” code for frequency
Frequency tuning curves of
single hair cells
16
Complex tones:
Fourier decomposition


Basilar
membrane acts
as f analyzer
Breaks down
complex f inputs
into constituent
pure tone
components
17
Auditory masking: Evidence
for cochlear place coding

Auditory masking



Presence of certain tones
decreases perception of
nearby tones
Similar f result in greater
masking
Asymmetry in spread of
masking


400 Hz mask
Increases threshold
for 800 more than
200 Hz
Consistent with basilar
vibrational overlap
E.g. 400 Hz mask overlaps
more with 800 than 200 Hz
18
Mystery of the missing
fundamental

400 Hz fundamental plus harmonics
(800, 1200, 1600, 2000)


What if remove fundamental f (400Hz)?



Perceived pitch doesn’t change!
Hence: The missing fundamental
Problem for place coding


Sounds like 400 Hz pitch with complex timbre
No direct stimulation of 400 Hz on basilar
membrane
Harmonic structure determines perceived pitch
Not what is present on basilar membrane
 What we hear is not what the basilar membrane tell us, but what
19
our brain does

f
What does Barry White sound
like on the telephone?


Telephone carries 3003400Hz
Typical male voice


Barry white



Fundamental f = 120 Hz
30 Hz?
Can’t speak to Barry on the
telephone?
Missing fundamental allows
us to hear “virtual” pitch of
voice
20
If its too loud your too old

Db (SPL) scale


Attenuated low and high f
relative to midrange
High volume



Loudness varies with f
Low volume


Pain and pleasure
Audibility curves


Loudness doubles about
every 10 db at 1000 Hz
Less frequency attenuation
Low volume sounds muddy
 Mostly mid range
I like my music loud
Each curve represents equal
loudness
21
Otoacoustic emissions:
Talking ears

Ears don’t only receive sounds, they make
them!



Occur spontaneously and also in response to
sound


It like your ears are talking back!
Created by movement of outer hair cells (ohc)



Discovered in 1978
Tiny microphones
Part of auditory sensitivity is movement of ohc to
change region specific flexibility of basilar membrane
Allows tuning curves to be so narrow
Hearing impairments often start with loss of ohc
function
22
Auditory localization

Where is the sound coming from?




Distance
Elevation (vertical)
Azimuth (horizontal)
Localization not nearly as precise as vision



Localization within 2-3.5 degrees in front of head
20 degrees behind head
Suggests important role of vision

Tunes auditory localization
23
Why is is auditory localization
not obvious?

Vision


Stimulate different photoreceptors in eye
Audition
No such separation of sounds sources on
sensory surface
 Sources combine to equally stimulate ear
receptors

24
Why have two ears?

Two aural perspectives on the world

Like vision, can be used to get different
sound pictures of environment

Binaural cues

The disparities between ears is used for
localization
25
Azimuth

Interaural (between ears) Time
Difference (ITD)





Air pressure changes are very slow
relative to speed of light
ITD at side = max 600 µS
ITD at front = 0
Can induce perception of location by
varying ITD using headphones
Interaural Level (intensity) Difference
(ILD)


Amplitude decreases w/ distance
Head casts sound/acoustic shadow



Reduced amplitude due to reflection
Measure w/ tiny microphones
f dependent

Greater shadow for higher f
26
Elevation



ITD/ILD not very useful
Use spectral cues
Frequency information
can result in different
perceptual qualia




Monaural: f serves as
signal for pitch
Binaural: f serves as
signal for location
Pinna differentially
absorb f
Result: Notches in
frequency spectra
Above
Level
Below
27
Distance

At close distances (< 1 meter)




ILD can discriminate near and far
At very close distances ILD is very large (e.g. 20 Db)
But what’s that going to do for us?
At far distances

We are very poor judges for unfamiliar sounds



Use sound level for familiar sources
Frequency: Auditory atmospheric haze



Suggests that sound serves as signal for visual search
Absorption of high f
Sound muffled
Auditory parallax

Sounds move faster across ears at near relative to far
distances
28
Brain basis for localization
Sound to right

ITD detectors



Brainstem: Superior
olivary nucleus
Primary auditory cortex
Coincidence detection



Neurons fire maximally
when signals arrive at
same time
Thus: “coincidence”
Axonal distance create
input delays
Sound to left
29
Auditory scene analysis

How do we segregate different sounds being
produced by many sources simultaneously?

How do we tell what frequencies belong to what
source?


E.g., Cocktail party
Don’t perceive an unorganized jumble of frequencies



Not simply high vs low f
Most f ranges overlap
How do we segregate information as belonging to
distinct auditory objects?
30
Principles of auditory
grouping



Like gestalt visual principles
Auditory stream segregation
Similarity
Timbre
 Location
 Pitch
 Time
1 stream

2 streams
31
Auditory-visual interactions:
Location and pitch

Visual capture of sound

Location: Ventriloquism
effect
Pitch: McGurk effect






QuickTime™ and a
Cinepak decompressor
are needed to see this picture.
“Ba”
“Va”
“Tha”
“Da”
Visual information is
integrated with audition

Creates fused auditory visual
perception
32
Auditory-visual interactions:
Location and pitch

Auditory experience is much more than
pressure level changes
33