The dynamic audiometer
A founding text by Alfred Tomatis published in September 1953 in the Bulletin du Centre d’Études et de Recherches Médicales de la S.F.E.C.M.A.S. (Société Française d’Étude et de Construction de Matériel Aéronautique Spécial), where he directs the Medical Research Laboratory. Tomatis here sets out the principle of the dynamic audiometer: an instrument capable of measuring the real value of a subject’s hearing in the presence of a background noise, by contrast with the classic audiometer which artificially isolates the ear from the sound world. A major methodological break: audiology ceases to consider the ear as an abstract organ and replaces it within its normal conditions of operation.
BULLETIN OF THE CENTRE D’ÉTUDES ET DE RECHERCHES MÉDICALES OF THE SFECMAS — September 1953
THE DYNAMIC AUDIOMETER
by Dr Tomatis
Deputy Director of the Research Laboratory
of the SFECMAS
Audiology, as its name indicates, has as its principal aim the determination, with the utmost rigour and precision, of the auditory behaviour of individuals.
As regards its field of application to therapeutics, it presents itself as a most precious means of investigation.
Indeed, the statistical procedures applied within audiology have made it possible to establish a number of general rules which most often allow the practitioner to classify the audiogram of the deaf subject examined in a well-defined category, and to draw useful conclusions therefrom as to the treatment to be prescribed or the surgical interventions to be envisaged.
Moreover, it is now possible, thanks to audiometry, to anticipate a deafness not yet declared, and to prevent it by remedying in due time the deficiencies that cause it.
Thus the development of the techniques of audiology (classic audiometry and ever more numerous tests) marks very important progress in the detection of hearing disorders.
But in this domain, whose possibilities assert themselves more strongly each day, much research remains to be done.
There is no question that for some years great efforts have been made and that more than encouraging results have been obtained.
We are nonetheless obliged to observe that, at present, this science is only beginning to free itself from the framework of the necessities that gave rise to it.
Until now, attention has simply been devoted to undertaking anatomical research on the ear, to calibrating this ear in more or less arbitrary circumstances, to seeking to establish relations between the alteration of a given band of the sound spectrum and a given modification of the auditory system, and finally to setting the problem of hearing within a far too narrow framework (which is, moreover, perfectly normal at this preliminary stage).
But audiology must now cross the limits hitherto imposed upon it, which restrict its field of application to the ear proper, isolated from the surrounding sound environment. Let us not forget that the role — indeed, the primary destination — of the ear is to ensure a link between the outer world and the individual.
Consequently, audiology as we conceive it concerns not only the transmission of sound vibrations but also the circumstances in which these vibrations are captured, and the totality of their repercussions on the individual, and finally their detection by the brain.
Now, we hold that audiometry, as currently conceived and practised — that is to say, by the determination of minimum thresholds of perception — does not make it possible to obtain the real value of an individual’s hearing, in other words, the true response curve of their auditory circuit under the normal perceptual conditions to which they are accustomed.
When we examine an audiogram, we deduce that the individual, isolated from all noise, perceives a given frequency normally and has a normal deficit for another; but we cannot in any way ascertain how the ear behaves in everyday life.
We simply obtain a particular response curve under circumstances that practically never exist in ordinary conditions, save in certain cases of fairly accentuated deafness.
Indeed, when an audiogram is established, the subject being examined is plunged into a profound silence. The intensity at which they perceive the various frequencies sent into one ear or the other is noted on a graph.
But this is a purely quantitative examination and one that allows no precise conclusion about their normal auditory behaviour.
We are keen to repeat what we have already stated at the beginning of this article, namely that we in no way at present contest the validity of tonal audiometry as it is currently practised.
To derive statistical laws from the examination of audiograms, it is absolutely necessary that these be established in rigorously identical circumstances, and it is evident that among these circumstances silence represents that which is simplest to achieve and which offers the minimum risk of error.
This is why this method was the first to come into being. Its simplicity and its absolute character, allowing comparative studies, made it natural that it should impose itself, and it lies at the basis of the present development of otology.
But we nonetheless think that this aspect of audiometry is only a means of diagnosis, an effective procedure of screening and research resting on solid though somewhat arbitrary bases, and which does not provide a ground for drawing valid conclusions about the state of hearing, since it isolates the ear from its real setting.
Our aim, therefore, will be to seek to determine the auditory behaviour of an individual under the normal circumstances of their existence.
Physiological optics having a certain advance over audiology, we shall draw a comparison between such of their elements as present common features, in order to make our thought clearer.
When an ophthalmologist examines the visual acuity of an individual, they seek to know how the eye of that individual reacts under normal conditions of visibility. And the lighting environment of the examination room will preferably consist of white light — that is, a mixture of all the frequencies of the visual spectrum — at an intensity to which the individual is accustomed.
Under such conditions, the examination will enable the ophthalmologist to draw valid conclusions on the everyday visual behaviour of the subject examined.
Let us now imagine the individual placed in a dark room (absence of any visual frequency) and noting from what light intensity they become able to perceive luminous patches of different colours — that is, of frequencies varying within the visual spectrum — presented to them successively. We shall thus obtain a “sensitivity curve of the eye” to the various frequencies, but this curve will in no way give us the possibility of judging the visual behaviour of the individual in everyday life, of their astigmatism, hypermetropia or any other anomaly of their vision.
It is exactly the same with regard to audiometry.
The classic audiogram is only the sensitivity curve of the ear to the various sound frequencies in the absence of any background noise.
These various observations led us to study an apparatus capable of furnishing information of a more real, more concrete value about hearing, and which allows, during the examination, the creation of sound circumstances tending to replace the ear in its normal domain of operation.
Diagram of the dynamic audiometer
[Fig. 1 — Front panel of the dynamic audiometer: two symmetrical channels (A — frequency adjustment of channel C1 / a — frequency adjustment of channel C2; B — calibration at frequency 0 of C1 / b — calibration at frequency 0 of C2; C — variable intensity adjustment C1 / c — variable intensity adjustment C2 + general switch; D — fixed-point intensity adjustment C1 / d — fixed-point intensity adjustment C2; E — switch C1 (C2 or Masking) / e — switch C1—C2 Masking; F — press-switch C1 / f — press-switch C2 or Masking; K — 90/110 dB switch; L1 — 110 dB light signal; L2 — 110 dB light signal; μA — central micro-ammeter].
We have called our device the DYNAMIC AUDIOMETER because it enables us to obtain, in a manner of speaking, a value of hearing; and since one no longer abstracts from all external disturbances and the excitation due to the pure frequency is superimposed on those resulting from the background noise, the ear reacts in this case as it is accustomed to in everyday life.
It is known that the audiometers currently in use consist principally of a calibrated low-frequency generator, providing sinusoidal acoustic vibrations spread from 128 c/s to 12,000 c/s, and of a set of perfectly calibrated attenuators allowing each frequency to be transmitted at a sound level varying in a known manner from –10 to +100 dB for air conduction.
As regards bone conduction, the range of frequencies extends from 128 c/s to 4,096 c/s, and the intensity range from –10 to +60 dB (zero decibel indicating by convention the level at which a normal ear perceives each frequency in the absence of any background noise).
The more advanced audiometer we have built consists essentially of the following elements.
Two LF beat-frequency generators G1 and G2. Let us recall briefly the principle of operation of such a generator. A fixed HF oscillator of frequency NA and a variable HF oscillator of frequency NB are coupled to a detector-mixer C. Within C we shall therefore have the frequencies:
NA, NB, NA + NB, NA – NB
NA, NB, NA + NB are HF and consequently of no interest; they will be eliminated at the output of C by means of a low-pass filter. By contrast, one may adjust oscillators A and B so that the differential frequency NA – NB is LF, which the filter, suitably set, will let through. If oscillator B includes an adjustable tuning condenser, it becomes possible, by the sole manipulation of C1, to vary the frequency NA – NB and to make it cover the whole band of audible frequencies.
[Fig. 2 — Schematic principle: Fixed HF oscillator (NA) → detector-mixer C ← Variable HF oscillator (NB); output of C → low-pass filter → LF (NA – NB)].
The generators G1 and G2 are independent of one another. Each can provide a sinusoidal voltage of frequency adjustable between 32 c/s and 17,000 c/s.
For each of them, a calibration device has been provided which, by an action independent of the tuning C/F, allows the variable oscillator to be adjusted so that the frequency heard corresponds exactly to that indicated on the dial. For this it suffices to perform an adjustment at frequency zero. If the micro-ammeter, by its maximum deflection at frequency zero, indicates the absence of impedance between the two self-inductors measuring the LF current, the maximum deflection towards zero of the micro-ammeter signifies that LF is negligible and that one is at frequency zero.
A background-noise generator G3 supplies a complex sound composed of numerous audible frequencies, often called “white noise” by analogy with white light. In most devices, recourse is had to white noise inside a neon tube to generate noise. We have preferred to use another procedure.
G3 consists of a high-gain amplifier whose input voltage is reduced to the hiss of a high-value resistor placed in the grid of the first amplifying valve. The hiss thus produced is amplified normally and supplies the desired background noise.
A switch makes it possible to select at will either of the generators, or to group them two by two, or all three together.
The amplification of the frequencies issuing from each of the generators is independent, and a mixer adds the currents that generate at the level of the earphones perfectly defined sounds or perfectly identifiable noises (case of pure tones emanating from G1 and G2; by contrast, for G3, the oscillation produced having a complex form, the indication supplied by the micro-ammeter is not entirely rigorous and gives only an approximate value of the emitted power).
Let us now examine the possibilities of this device.
First of all, it makes it possible to obtain an intensity variation by decibel from –5 dB to +90 dB in the case of normal use.
Indeed, to each generator corresponds a fixed-point switch allowing variations of 5 dB and 6 dB between zero and 90 dB.
Furthermore, with each of these switches is associated a potentiometer that allows progressive variations from –5 dB to +5 dB to realise the intensity variations considered, variations which are all on the dial of the previously calibrated micro-ammeter.
In the case of great deafness, a switch makes it possible to obtain a greater intensity and to rise to 110 dB. In this case, and for the frequencies for which this is necessary, one uses 4 fixed points 72 dB, 78 dB, 84 dB and 90 dB, which are increased by 20 dB each.
Thanks to this possibility of progressive and continuous intensity variation, we have been able to ascertain that, contrary to a fairly widespread opinion, the ear is able to perceive intensity variations equal to a sub-decibel, and that such a difference is sufficient to establish a hearing threshold with certainty.
The audiometric examination will begin with a recording of the classic tonal audiogram, which will furnish us with what we shall agree to call henceforth the linear value of the ear. For this recording, we shall use generator G1 alone. Then, we shall seek to obtain the dynamic audiogram properly so called by simultaneously sending to the ear a background noise and the pure frequency emitted by G1.
Experience has shown us that there is interest in establishing both audiograms at the same time, noting on a single graph the linear threshold and then the dynamic threshold relating to a given frequency.
The noise of conversation in general has an intensity of 35 dB, which corresponds to an ordinary sound environment.
But one may plot several curves corresponding to different intensities of background noise, which will give us the dynamic value of the ear for each of these intensities.
In particular, it will be of interest to plot the dynamic curve of an individual’s ear by injecting a background noise whose intensity will be that to which they are accustomed by virtue of their occupation, for instance.
These audiograms benefit from a reduced margin of error, owing to the precision in the establishment of the thresholds.
Furthermore, the ease with which one can vary the frequency continuously makes it possible to “sound” certain bands that arouse the examiner’s curiosity by some anomaly.
Thanks to this device, still using G1 or G3, it becomes easy to determine the saturation thresholds of the ear.
The restricted number of dynamic audiograms available to us for the moment has unfortunately not yet permitted us to establish precise hypotheses on the behaviour of the ear in noise. Nevertheless, the first comparisons carried out tend to prove how the ear reacts in a wholly different way when it is no longer perfectly isolated from the outside.
Using generator G1 alone, it is easy to determine the selectivity curves of the ear at different intensities.
Finally, thanks to this set of physiological values, it becomes possible to revise a number of tests of auricular fatigue and to create new ones.
The study we have just carried out of this new apparatus enables us to anticipate the very many services it will be able to render in the domain of audiometry. We propose to make several modifications to it to render it still better adapted to a role that will be, above all, to place the ear under examination back within its normal conditions of operation.
We thus intend to complete the transmission arrangement by adding a bone vibrator that will excite the bony cortex at the same time as the earphone transmits the vibrations to the eardrum, since in the domain of dynamic audiometry it is essential not to separate the two modes of conduction.
The fixed adjustment will enable us to distribute the sound intensity supplied by each of these vibration generators in accordance with reality. We shall thus obtain the overall dynamic value of the ear.
We also intend to add to this device a magnetic reading head, which will make it possible to carry out far more thorough research on the relations existing between the fundamental frequencies of a complex noise (forge, boilermaking, aircraft engine, for example) and the audiograms, classic and dynamic, of individuals regularly subjected to such noises.
The classic audiogram will be established normally.
As for the dynamic audiogram, it will be established as usual, but replacing the background noise habitually supplied by G3 with the playback of a recording taken in a sound environment identical in nature and intensity to that to which the subject being examined is accustomed.
We shall thus have three sources of comparison: the classic audiogram, the dynamic audiogram taken under circumstances corresponding to reality, and the sound spectrum of the transmitted noise (which we shall probably have analysed).
We are convinced that by this comparison we shall obtain interesting results in the domain of occupational deafness, as well as precious indications of the possible means of remedying it.
We hope to have given here an idea of what one is entitled to expect from this new device and of the principle that made it necessary. And we believe that it may soon enrich with new chapters this entirely new science that is audiology.
Source: Tomatis A., “L’audiomètre dynamique”, Bulletin du Centre d’Études et de Recherches Médicales de la S.F.E.C.M.A.S., September 1953, pp. 76-86. Bulletin directed by Dr J.-R. Rounon. Digitised document from the personal archives of Alfred Tomatis. Transcription respecting the pagination and the particularities of the original (typography, diagrams).