Lecture 4

Question 0

The time of arrival difference between the two ears is

Answer 0

a function of the angle of
incidence of the sound wave. If the sound comes from directly in front, it arrives at
both ears at the same time. If it comes more from the left, it arrives first at the left
ear by a direct path and then at the right ear by a direct path up to the face and then
it wraps around the face. So the extra time it takes to get to the right ear is the time ittakes the sound to travel the straight path after the first dashed line to the face and
then the curved path around the face.

Question 1

Space: the two concepts

Answer 1

Localization: in space
Lateralization: headphones

Question 2

The ITD varies from

Answer 2

0 to about 600 microseconds (0.6 milliseconds) as the angle
varies from straight ahead (or behind) to completely to one side. This temporal cue
works best for low frequency sounds whose wavelength is bigger than the diameter
of the head. For higher frequencies there is phase ambiguity because you don’t know how many cycles of the wave there are between the two ears in order to make sense of the phase (time) difference.

Question 3

Differences in level between the two ears are greatest for high frequencies whose?

Answer 3

wavelengths are smaller than the diameter of the head, because the head casts an
acoustic shadow for these wavelengths, whereas longer wavelengths simply wrap
around the head. Note that the ILD is only about 8 dB at 1000 Hz and is negligible at200 Hz.

Question 4

The minimum audible angle is?

Answer 4

the smallest change in angle of incidence that a
listener can detect from a given position in space. Note that this angle is much
smaller for sounds straight ahead than for sounds more to the side. The bump just
below 2000 Hz in the curve for the solid circles (a sound straight ahead) is due to the
fact that ITDs play a stronger role at lower frequencies up to about 1500 Hz and ILDs
play a stronger role at higher frequencies above 2000 Hz. Neither is very good around1800 Hz, hence the increase in minimum audible angle.

Question 5

It only takes a difference of about ___ dB to move theimage from the middle all the way to one speaker or the other?

Answer 5

6

Question 6

There are sets of positions that give exactly the same values of ITD and ILD, and they
should theoretically be impossible to distinguish if these were the only cues we used. Call edit he come of confusion. However, we don’t confuse them, so there must be another source of information
about localization.

Answer 6

This information comes from the pinnae which create bumps and dips in the
spectrum of the sound, that change in frequency as a function of position in azimuth and elevation. This is a particularly important cue for resolving the cone of confusion and perceiving the elevation of a sound, as well as front-back differences. In the
figure, the elevation is given in the 5 rows, from -30° to +30°, and three different
values of azimuth are shown in the columns for the right and left ears. Each curve
shows the filtering characteristic of the ears for a given position. For example, in the
upper right corner of each panel, we see that the higher frequencies are more
attenuated than the lower frequencies. Also note that the curves are not identical forthe two ears. The dashed line in the right columns shows how a prominent dip
evolves with change in elevation. Pinna cues are also responsible for what has been
termed facial vision in blind people: the feeling that, as one approaches a wall, there
is a change in sensation referred to the skin of the face. However, if the bumps and
dips of the pinnae are filled up with putty, leaving only a straight path to the auditory canal, this sensation disappears, and blind people are not as good at navigating
through space under such conditions.

Question 7

Distance perception in audition depends on which three main factors?

Answer 7

Level, direct/reverb ratio, filtering…
Distance perception in audition depends on three main factors. The sound pressure
decreases proportionally to the distance between the listener and the sound source.
So if you double the distance, you decrease the pressure by 6 dB. Also, as a sound
source moves further away from a listener in an environment where there are variousobjects around that reflect sound, the ratio between the sound level coming straight
from source to the listener and the sum of all the sound levels from the same sound
bouncing off of all the objects (the reverberation) decreases. Finally, molecules in the air can absorb sound waves that are small (high frequencies), so sounds that are
further away have more and more absorption of high frequencies, or filtering. So
sounds further away sound softer, less bright and feature more echoes.

Question 8

Masking

Answer 8

The capacity of one sound to reduce the intensity of another sound, even to zero.e

Question 9

Equerry selectivity

Answer 9

The ability to process frequency information in small bands, frequency bands.

Question 10

masking curves characteristics?

Answer 10

the curve is fairly symmetric at low levels, but
becomes more and more asymmetric at higher levels. The slope of the
masking curve on the low-frequency side remains constant, but the slope onthe high-frequency side gets shallower and shallower.

Question 11

Ingle fibre tuning curve characteristics?

Answer 11

s there is a lot of noise,we have a harder time selectively analyzing different sound components. Frequency selectivity much more difficult

Question 12

lower-
frequency maskers have more of a masking effect than do higher-frequency maskers.
True false

Answer 12

True

Question 13

Masking pattern of complex tones.

Answer 13

Note that on a logarithmic scale similar to theone along the basilar membrane, the harmonics get closer and closer together as the harmonic rank increases, and are less and less distinguished in the
masking pattern, reflecting the fact that they are not separately resolved in the basilar membrane excitation pattern. Note also that the degree of resolution
depends on the level. There is less resolution at high levels due to the spread
of the excitation pattern around each frequency

Question 14

The phon scale?

Answer 14

This diagram shows the level one would have to set for a pure tone of a given
frequency to be as loud as a 1 kHz pure tone at a particular level (the value in dB SPL
is shown for each curve at 1 kHz). To make a glissando of a pure tone across all of theaudible frequencies, it would have to follow one of the curves to be perceived as
having a constant loudness. That is why they are called equal-loudness contours. This loudness is given a number in units called phons that corresponds to the physical
level in dB SPL of the reference sound at 1 kHz. So a 1 kHz tone at 40 dB is said to
have a loudness level of 40 phons. To have a loudness level of 40 phons, a 100 Hz
tone would need to have a physical level of 52 dB SPL. These two points on the 40
phon curve are shown with red Xs. Note that the slope of the curve on the low-
frequency end is much steeper at low loudness levels than at higher loudness levels.
This means that loudness grows faster with level at low frequencies than at medium
and high frequencies. For example, to go from 10 phons to 70 phons at 2 kHz, one
has to increase the level by 60 dB, but at 50 Hz, one only has to increase the level by
38 dB.

Question 15

Because of the equal-loudness functions being flatter at higher levels than at lowerlevels, a recording made and equalized at high levels would?

Answer 15

sound very different at low levels (the lower fuzzy curve). Indeed you might not hear
the very low frequencies at all. This effect is compensated for on some home audio
units by having a “loudness” button that compensates for the difference by boosting
the low and high frequencies. This is an example of applications of psychoacoustics inthe audio industry.

Question 16

The Sone scale

Answer 16

This diagram shows the normal function of loudness for a 1000 Hz pure tone. It is
obtained by asking people to attach a number to the perceived sound level. If
another sound has twice the loudness, it should have a number that is twice as big.
This is called a ratio scale in psychophysics. Notice that there is a linear relation
between the log of the rating and the level in phons (which correspond to dB SPL for a 1000 Hz pure tone). For equal-loudness levels above 30 phons (30 dB SPL), the
equation of the function is a power function, i.e., the mapping of the physical level
onto the perceived magnitude involves an exponent (0.6 in the case of loudness). Forweaker equal-loudness levels, the slope of the function is higher. We will derive a
scale like this for noise bursts in Lab 2. However, in the lab we’ll use a different
paradigm with a constant reference sound presented just before each tested sound.
The number is fixed for the reference at 100 and the ratio of loudness of the test
sound to the reference sound determines the number assigned to each. In this kind
of paradigm, the exponent of the loudness function is 0.3.

L = loudness in sones 
K = .01 (constant) 
P = sound pressure in µPa

Question 17

Ow does the basilar membrane descriminate loudness?

Answer 17

certain auditory nerve
fibres fire at a higher rate, but more fibres would fire because adjacent regions of thebasilar membrane would be stimulated as the level increased.

Question 18

El coding and temporal synchronization of nerve fibers.

Answer 18

at lower levels, firings of the fibre tend to increase near the
peak of the waveform, but the average rate doesn’t change much. As the level
increases further, both the firing rate increase and the tendency to fire at the peak ofthe waveform increases up to about 60 dB. Beyond that not much changes in either
firing rate or firing synchrony. Across a whole set of fibres being stimulated by this
sound, there would be a tendency for their activity to synchronize.

Question 19

Dependence of loudness on duration.

Answer 19

This diagram shows physical level on the top, specific loudness in the middle and
relative total loudness on the bottom. Tones of several durations are shown. The
middle panel demonstrates that when a tone is turned off, it takes some time for its
loudness to decay away. This is a kind of sensory persistence effect, like you get when you see a brief flash of light in the darkness and it slowly fades away. The bottom
panel shows that the sensory information for loudness also takes time to accumulate,as shown by the slope of the curve at the left of the diagram. Very short tones do not last long enough for the total loudness to attain its maximum. So even though they
have the same maximum physical level, shorter tones below about 100 msec seem
less loud than tones lasting longer but having the same maximum level. This is the
psychoacoustic basis for agogic accents in organ performance. You can’t play the
notes louder on an organ, so to accent them you must play them longer.

Question 20

Dependence on loudness on spectral spread?

Answer 20

The diagram on the left shows what happens to the loudness of a band of noise when the bandwidth changes keeping the total level constant. Two examples of high- and
low-level noise bands are shown on the right. The dark blue line is the centre
frequency and the light blue rectangle indicates the critical bandwidth. The curved
dashed line indicates auditory threshold. The loudness stays constant until the
bandwidth of the noise exceeds the critical bandwidth, and it then increases with
bandwidth for levels above 20 phons. This demonstrates that the energy in each bandis being summed across critical bands. At the lowest level, the wider band noises start to have very little level in each band and so fall below threshold, which is why the 20-
phon curve starts to descend above a 1000-Hz bandwidth.

Question 21

Time varying loudness?

Answer 21

the different bands are louder or
softer at different times and that the specific loudness in each band fluctuates overtime. The total loudness is shown at the top. Notice that the louder portions
correspond to the vowel portions of the signal.