|
Hi, I think the first to publish a measured HRTF in a peer-reviewed
journal is probably Francis M Wiener and Douglas D Ross, JASA 18,
401-408 (1946) "The pressure distribution in the auditory canal in
a progressive sound field". In this study, not only did they
measure the free-field transfer function, and recognize the
amplification caused by the ingenious physics, they also addressed
the significance of the recording point and demonstrated that the
sound transmission along the ear canal does not depend on
direction (in the frequency range considered). A supplementary account was given by Francis M Wiener in 1947 JASA 19, 143-143 (1947), which was titled "On the diffraction of a progressive sound wave by the human head". This study further demonstrated the dependence of angle of incidence of the sound wave. It has always surprised me, how few citations these papers have, comparing to later reports of HRTF measurements. The understanding of the diffraction by ear, head and body
predates these publications, thus it is not unlikely, that the
term head-related has been used casually at this time. There are
several relevant studies of human localization preceding Wiener
and Ross' measurements, which all relate the transmission
properties incl. diffractive characteristics of the ear and head
to the nature of human localization. Later publications, incl.
e.g. the work by Edgar A. G. Shaw and colleagues, focused on the
external ear characteristics, and the term external-ear transfer
functions was also used for transfer functions, which we today
would probably refer to as head-related transfer functions. H.-J. Platte and P. Laws used the term head-related stereophony
(Kopfbezogene stereophonie) in their paper in Rundfunktechniches
Mitteilung, 22, 22-26 in 1978, and H. Wilkins in Acustica 26,
213-221 (1982). The term kopfbezogene (head-related) was probably
already widely used in the tonmeister community, referring to the
binaural recording technique. My guess is that the scientific
communities used the term in presentations and discussions, but
that the lack of adequate definition required written
communications to be more accurate in what transfer functions were
reported. Still, today, I don't find that the term head-related transfer function has an equally robust definition, as e.g. the free-field transfer function or the progressive wave definition. Strictly speaking, a free-field transfer function is measured in conditions, where the wave impinging on the individual is plane. We try to establish that in the lab with speakers at distance, thus any study that investigates the limits of this, or study distance dependence in the transfer functions within 1-2 meters, measure distance (and source) dependent transfer functions. The literature has many accounts of "distance dependent head-related transfer functions", which wouldn't make sense, if we accept the plane progressive wave definition. But because the transfer function is measured in much similar conditions as the free-field transfer function, and since it is something "head-related" it makes sense to refer also to such transfer functions as "head-related". This also suggest that the term head-related transfer function is not unequivocal, and has maybe never really been coined. I have also come across presentations that refer to the "earphone
head-related transfer function", which is an even vaguer
definition for the in-situ measurement of an earphone's response.
For the majority of the HRTF scientific community, I wouldn't
expect this to be an acceptable term, but it is intuitive to refer
to this as a head-related (it is measured on the head, and its a
transfer function). I guess that head-related is intuitively and de facto a term,
which we can't assume to be used only for the progressive plane
wave transformation of the ear, head and body, which most of us
consider to the head-related transfer function. Yet it seems still
necessary to be accurate, whenever using the term. Many published
"HRTFs" have arbitrary levels, and cannot be the free-field
pressure ratios, although they display some of the distinct
features (peaks and dips). Some don't even approach 0 dB at low
frequencies, which the real free-field transfer function always
do. And graphics showing the HRTF always seem to show a source
close to the head (trying to save space in the illustration no
doubt). Best, Dorte. On 18-02-2018 15:31, Stephan Paul
wrote:
|