Perception Chapter 11: Hearing and Listening

Perception Chapter 11: Hearing and Listening
Audiogram or Audibility Function (AF) for normal human young adult. Note: peak
sensitivity 2-5K frequency range.
Audiogram or AF: a graph indicating the threshold intensity of varying frequencies. A
normal audiogram show declining threshold intensity from 20hz to about 2,500hz
(where it usually bottoms out). Then a steady increase from about 5,000 to 10,000;
with a very steep increase from 10,000 on.
Afs: ages and species comparisons
Top shows different
Bottom shows
different ages for
humans. Note: as
one gets older
“contraction” of
range occurs, but
most pronounced
at high frequencies
increasing age
increases threshold
intensity for higher
Hearing loss: Conduction deficit
Conduction loss:
raised threshold
levels for entire range
of frequencies due to
malfunction in outer
ear (can be as simple
a ear wax build-up or
more serious like
severe rupture in
tympanic membrane)
or middle ear (which
usually involves
reduced conductive
abilities of ossicles)
Otitis media:
childhood ear
Otosclerosis: gradual
immobilization of
stapes (overgrowth).
Hearing loss: Sensory-neural
Affects response on
Basilar membrane.
Usually more restricted
range of frequencies
affected. No differences
in bone vs. air
conduction. Loud sounds,
drugs, presbycusis,
trauma to BM all
potential causes.
Hearing loss: treatments
Hearing aids: electronically amplify incoming auditory signal, usually boosts
speech range more relative to other frequencies
Cochlear implants: electrodes implanted direction in cochlea to directly stimulate
Auditory masking and the Critical Band
Auditory masking: The extent to which background
noise of varying frequencies (broadband noise) can
serve to block or mask the perception of a signal.
Ex: trying to listen to the radio (signal) with and without
the engine running (masker).
The paradigm for studying this works in the following
1) a subject attempts to detect a signal of a certain
frequency, in the presence of a noise masker.
2) the masker is broadband noise which is varied in two
ways (1) location of the center frequency of the noise
and (2) the extent of the range of frequencies
surrounding the center frequency.
Results indicate:
1) masking is asymmetric: lower frequencies mask
higher frequencies more effectively than vice-versa
2) masking only occurs within a certain range of
frequencies surrounding the center frequency: called the
critical band.
Two reasons for this:
1) response of basilar membrane to lower and high
frequency stimuli
2) response characteristics of frequency tuned neurons
in auditory nerve.
Auditory masking and the Critical Band
Why is masking asymmetrical and why is there a critical band? Two reasons for
1) response of basilar membrane to lower and high frequency stimuli
2) response characteristics of frequency tuned neurons in auditory nerve.
Loudness Perception
Magnitude estimation: this is a
procedure where a certain intensity
level of sound (say 60dB) is assigned an
arbitrary number to describe its
loudness. Subject is then presented a
series of sounds which vary in their
intensity. Subject assigns a number to
describe the loudness of the sound
relative to the standard sound.
Results using ME demonstrate that the
relationship between sound intensity
and perceived loudness is not a 1 to 1.
Loudness tends to increase more slowly
than then does intensity (bottom line,
open circles on graph)
Doubling intensity increases perceived
loudness only about 60%. To double
loudness, about a 10dB (or tripling)
increase in intensity is required
Equal loudness contours
Loudness Matching: this is
where a certain frequency
tone is presented at a certain
dB level (for example 1000hz
tone at 60dB). Then a second
tone at a different frequency is
presented and subject adjusts
intensity level until it is the
same loudness.
Equal loudness contours: a
family of curves representing
the required intensity of
varying frequencies such that
those frequencies match the
perceived loudness of a
1000hz tone at a certain dB
level. For example, an ELC for
20dB, indicates the dB level
required for each freq, such
that it sounds as loud as a
1000hz tone at 20dB. All freqs
along curve sound as loud as
1000hz tone at 20dB.
Neural basis of loudness perception
Two mechanisms:
Which nerve fibers are active?
How widespread is activity on BM?
Louder sounds will activate Lower spontaneous activity
fibers; and will involve more fibers in total along BM
Discriminatory power: usually 1-2dB change in intensity is
required to reach JND
Hi spon activity
Med spon activity
Lo spon activity
Pitch perception
frequency: usually
lowest frequency
produced by vibrating
body that accounts for
pitch perception. For
If a musical instrument
plays the note B3 at
around 220hz, it will
create a pattern of
overtones consisting of
440,660,880 etc. The
perceiver will hear a low
220hz note, with certain
timbre qualities.
The missing
what if
overtones are
What will the
person’s pitch
Missing fundamental
Despite absence of fundamental, pitch perception remains the same –
person perceives fundamental based on overtones.
How is this possible?
Place theory: big problem because there is no displacement at the place
of the fundamental frequency.
Frequency: yes -- the firing rate is determined by the differences between
overtones (harmonic spacing) in complex sounds.
Notice -- because overtones are multiples, it will be low freqs that require
frequency matching by neurons, higher tones will have place, thus this fits
in nicely with freq=low, place=hi, freq+place=midrange.
Localizing sounds: Where is it coming from?
Two cues discussed earlier:
Inter-aural intensity differences and inter-aural time differences
Intensity differences vary by frequency; greater difference for higher
frequencies. Shadowing effect weaker for lower frequencies (left).
Localizing sounds: Where is it coming from?
Inter-aural time differences vary with location (left). However, there is a zone on
either side of the head where neither ITD nor IID specify a unique location in
space: Cone of confusion (right).
Duplex theory: ITD used for low frequencies (1,000Hz below); IID used for high
frequencies (4,000Hz above). 2-4K Hz most difficult to localize.

similar documents