554 Chapter 16
cally, as shown in Fig. 16-135, or recreated by an artifi-
cial head system.15,16,17
Reflections and diffraction of sound at the upper
body, the shoulder, the head, and the outer ear (pinna),
as well as resonances caused by the concha and ear
canal, are mainly responsible for the transfer character-
istic. The cavum concha is the antechamber to the ear.
The spectral shape of the external ear transfer function
varies from person to person due to the uniqueness of
people and the dimensions of these anatomical features.
Therefore, both artificial heads and their mathematical
models are based on statistical composites of responses
and dimensions of a number of persons.
All of these contributions to the external ear transfer
function are direction sensitive. This means that sound
from each direction has its own individual frequency
response. In addition, the separation of the ears with the
head in between affects the relative arrival time of
sounds at the ears. As a result the complete outer-ear
transfer function is very complicated, Fig. 16-136, and
can only be partially applied as a correction to the
response of a single or even a pair of microphones. In
the figure, the base of each arrow indicates reference
SPL. The solid curves represent the free-field (direct
sound) external ear transfer function, while the dashed
curves represent the difference, at each direction, rela-
tive to frontal free-field sound incidence.
Artificial heads have been used for recording for
some time. However, the latest heads and associated
signal processing electronics have brought the state of
the art close to in the ear (ITE) recording, which places
microphones in human ears.
The KU 100 measurement and recording system by
Georg Neumann GmbH in Germany is an example of a
high-quality artificial head system, Fig. 16-137. Origi-
nally developed by Dr. Klaus Genuit and his associates
at the Technical University of Aachen, the artificial
head, together with carefully designed signal processing
equipment, provides binaural recording systems that
allow very accurate production of spatial imaging of
complex sound fields.
The head is a realistic replica of a human head and
depends on a philosophy of sound recording and repro-
duction—namely, that the sound to be recreated for a
listener should not undergo two transfer functions, one
in the ears of the artificial head and one in the ears of
the listener.
Fig. 16-138 is a block diagram of a head microphone
and recording system. A high-quality microphone is
mounted at the ear canal entrance position on each side
of the head. Signals from each microphone pass through
diffuse-field equalizers in the processor and are then
available for further use in recording or reproduction.
The diffuse-field equalizer is specifically tuned for the
head to be the inverse of the frontal diffuse-field
transfer function of the head. This signal is then
recorded and can be used for loudspeaker playback and
for measurement. The headphone diffuse-field equal-
izers in the Reproduce Unit yield a linear diffuse-fieldFigure 16-135. Transfer function of the left ear, measured 4
mm inside the entrance of the ear canal, for four angles of
incidence (straight ahead, to the left, straight behind, and to
the right).
0°90°180°90°10 dB200 500 1k 2k 5k 10k 20k
Frequency HzRelative leveldBFigure 16-136. The human external-ear transfer function.Dependent on sound incidence
Shoulder
Hg(f)
Torso
Ht(f)
Head
Hh(f)
Pinna
Hp(f)InputSignal
+Independent of sound incidence
Prp(t)Ear Canal
He(f)Cavum
concha
Hc(f)Legend
Prp(t) Output signal dependent on sound incidence
Prpff(tt) Output signal equal to the sound pressure at human ears
rp Reference plane 4 mm inside the ear canal’s entrance
ff Free-field sound field
H Complex transfer functionPrpff(tt)