Psychoacoustics 45
voice differently from how other people hear the voice.
While not receiving much attention in everyday life, it
might be sometimes very important. For example, an
experienced voice teacher often asks a student singer to
record his or her own singing and playback with an
audio system. The recording will sound unnatural to the
singer but will be a more accurate representation of
what the audience hears.
3.2.3 Ear Canal
The ear canal has a diameter about 5 to 9 mm and is
about 2.5 cm long. It is open to the outside environment
at the concha, and is closed at the tympanic membrane.
Acoustically, it can be considered as a closed pipe
whose cross-sectional shape and area vary along its
length. Although being bended and irregular in shape,
the ear canal does demonstrate the modal characteristic
of a closed pipe. It has a fundamental frequency of
about 3 kHz, corresponding to a quarter wavelength
close to the length of the ear canal. Because of this
resonant frequency, our hearing is most sensitive to a
frequency band around 3 kHz, which is, not just by
coincidence, the most important frequency band of
human speech. On Fig. 3-4, the number 5 curve shows
the effect of the ear canal, taking the eardrum into
account as well. As can be seen, there is an
approximately 11 dB of gain at around 2.5 kHz. After
combining all the effects of head, torso and neck, pinna,
ear canal and eardrum, the total transfer function is the
curve marked with a letter T on Fig. 3-4. It is relatively
broadly tuned between 2 and 7 kHz, with as much as
20 dB of gain. Unfortunately, because of this resonance,
in very loud noisy environments with broadband sound,
hearing damage usually first happens around 4 kHz.
3.2.4 Middle Ear
The outer ear, including the pinna and the ear canal,
ends at the eardrum. It is an air environment with low
impedance. On the other hand, the inner ear, where the
sensory cells are, is a fluid environment with high
impedance. When sound (or any wave) travels from one
medium to another, if the impedances of the two media
do not match, much of the energy would be reflected at
the surface, without propagating into the second
medium. For the same reason, we use microphones to
record in the air and hydrophones to record under water.
To make our auditory system efficient, the most impor-
tant function of the middle ear is to match the imped-
ances of outer and inner ears. Without the middle ear,
we would suffer a hearing loss of about 30 dB (by
mechanical analysis^6 and experiments on cats^7 ).
A healthy middle ear (without middle ear infection)
is an air-filled space. When swallowing, the eustachian
tube is open to balance the air pressure inside the middle
ear and that of the outside world. Most of the time,
however, the middle ear is sealed from the outside envi-
ronment. The main components of the middle ear are the
three ossicles, which are the smallest bones in our body:
the malleus, incus, and stapes. These ossicles form an
ossiclar chain, which is firmly fixed on the eardrum and
the oval window on each side. Through mostly three
types of mechanical motions—namely piston motion,
lever motion and buckling motion^8 —the acoustic energy
is transferred into the inner ear effectively. The middle
ear can be damaged temporarily by middle ear infection,
or permanently by genetic disease. Fortunately, with
current technology, doctors can rebuild the ossicles with
titanium, the result being a total recovering of hearing.^9
Alternatively one can use devices that rely on bone
conduction.^10
3.2.4.1 Acoustic Reflex
There are two muscles in the middle ear: the tensor tym-
pani that is attached to the malleus, and the stapedius
muscle that is attached to the stapes. Unlike other mus-
cles in our bodies, these muscles form an angle with
respect to the bone, instead of along the bone, which
makes them very ineffective for motion. Actually the
function of these muscles is for changing the stiffness of
the ossicular chain. When we hear a very loud
sound—i.e., at least 75 dB higher than the hearing
threshold—when we talk or sing, when the head is
touched, or when the body moves,^11 these middle ear
muscles will contract to increase the stiffness of the
ossicular chain, which makes it less effective, so that
our inner ear is protected from exposure to the loud
sound. However, because this process involves a higher
stage of signal processing, and because of the filtering
features, this protection works only for slow onset and
low-frequency sound (up to 1.2 kHz^12 ) and is not effec-
tive for noises such as an impulse or noise with high fre-
quencies (e.g., most of the music recordings today).
3.2.5 Inner Ear
The inner ear, or the labyrinth, is composed of two sys-
tems: the vestibular system, which is critical to our
sense of balance, and the auditory system which is used
for hearing. The two systems share fluid, which is
separated from the air-filled space in the middle ear by