Compression of Audio Signals to MPEG and Dolby Digital - Digital Video and Audio Broadcasting Technology - page 147

8.3 Psychoacoustic Model of the Human Ear
In the following section, the process of audio compression will be dis-
cussed. Redundancy reduction (lossless) and irrelevance reduction (lossy)
lower the data rate of the original audio signal by about 90 %. Irrelevance
reduction relies on the psychoacoustic model of the human ear, which es-
sentially goes back to Professor Zwicker, former holder of a professorship
for electroacoustics at the Technical University of Munich. This type of
reduction is based on what is referred to as perceptual coding. This means
that audio components which are not perceived by the human ear are not
transmitted.
Let us first have a look at the anatomy of the human ear (Fig. 8.3., 8.4.).
The ear consists of three main parts: the outer ear, the middle ear, and the
inner ear. The outer ear performs the functions of impedance matching,
sound transmission over air, and acts as a filter with a slight resonance
step-up in the region of 3 kHz. It is in the same region, i.e. from 3 kHz to
4 kHz, that the human ear exhibits its maximum sensitivity. The eardrum
or tympanic membrane converts sound waves to mechanical vibrations,
which are transmitted via the malleus, incus and stapes to a membranous
window leading to the sensory inner ear. The air pressure must be the
same, ahead of and behind the eardrum. This is ensured by a tube connect-
ing the region behind the eardrum with the pharynx; the tube is called the
Eustachian tube. Everyone knows the problem of pressure building up in
the ear when climbing large heights. By swallowing, the mucous mem-
brane in the Eustachian tube provides for pressure compensation.
In the inner ear we find the organ of balance, which is made up of sev-
eral liquid-filled arches, and the cochlea. The cochlea is the actual hearing
organ (organ of Corti) by which sound is directly perceived. If the cochlea
were to be uncoiled, the sensors for the high frequencies would be found at
its entrance, then the sensors for the medium frequencies, and at the end of
the cochlea would be the sensors for the low frequencies.
The cochlea consists of a spiral canal in which lies a smaller membra-
nous spiral passage that becomes wider from the front to the rear. On the
inner membrane rest the frequency-selective sound-collecting sensors from
which the auditory nerves extend to the brain. The auditory nerves trans-
port electrical signals with an amplitude of approx. 100 mV pp . The repeti-
tion rate of the electrical pulses is in the order of 1 kHz. The information
contained in this rate is the volume of a tone at a given frequency. The
louder the tone, the higher the repetition rate. Each frequency sensor com-
municates with the brain via a separate neural line. The frequency selectiv-