The Voice (1962) — Alfred Tomatis

In-depth article by Alfred Tomatis (1962) devoted to the human voice considered as audio-vocal servomechanism. In nineteen dense pages, illustrated with thirteen figures, he there unfolds his complete theory of phonatory regulation — self-information loop, directing ear, transcerebral transfer, time adapter and gain adapter — and draws from it the consequences for the understanding of stammering, dysphonia, vocal ageing, auditory scotomata and musical audition. This presentation constitutes, before the publication of his synthesis works, the most complete statement of the servomechanical physiology of the voice according to Tomatis.

The Voice

by Dr Alfred Tomatis

Offprint, pp. 225-243 (1962).

Introduction

The voice offers such a power of evocation, of transcendence, that it can bring into play, in its study, very diverse disciplines. For it to be, a marvellous conditioning was required, made up of a psycho-physiological accommodation bringing into operation a highly differentiated servo-mechanism. To study the voice is to attempt to grasp the intimate play of the elements which ensure its emission, its sustainment and its monitoring.

I. — The audio-vocal servomechanism

The vocal act proceeds from a cyclic schema: a cortical act commands the centre of phonation, which actuates the organs of phonation; the emitted sound propagates through the air, reaches the ear, which transmits it to the auditory centre; the latter, by self-information, informs in its turn the cortex. The loop is thus closed and allows the permanent self-regulation of the emission.

[Fig. I — Servomechanical loop: cortical act → centre of phonation → organs of phonation → air → ear → auditory centre → self-information → cortex.]

The first stage of this monitoring is the ear. It is on it that the whole of the regulation rests. Any alteration imposed upon this receptor immediately entails an alteration of the vocal gesture.

II. — Asymmetry: the directing ear

Far from being symmetrical, the auditory system is lateralised. One ear exercises a directing role: the right in the right-hander, the left in the left-hander. It is on this ear that the regulation of phonation principally rests.

[Fig. II — Self-information circuit in the right-hander: right ear (R.E.) → left auditory centre (L.A.C.) → centre of phonation (C.P.).]

When the directing ear is the ear on the same side as the phonatory laterality, or when the roles are inverted, the circuit is unbalanced and the regulation is altered.

[Fig. III — Unbalanced circuit: dominant left ear in a right-hander — the pathway no longer normally passes through the contralateral auditory centre.]

The experimental suppression of directing audition — by plugging, masking or delaying — immediately entails significant disorders of phonation.

III. — The time adapter

Regulation supposes a very short transcerebral transfer delay. For the French syllable, whose average duration is of the order of 0.15 second, this delay must remain appreciably lower, without which the return information arrives outside useful time: phonation engages blindly. This is precisely the situation of experimental delayed feedback: one delays, by electronic means, the return of sound to the subject’s ear, and one observes that stammering appears as soon as the delay crosses certain limits.

The experiments of Lee, Black, Azzo and Azzi have made it possible to identify these thresholds:

delay T < 0.05 s: no notable disturbance;
0.05 s < T < 0.10 s: provoked stammering;
0.10 s < T < 0.20 s: suppression of the directing relay; the subject stammers or loses the thread of his emission;
T > 0.20 s: the voice continues normally, the auditory return no longer being taken as information for regulation.

[Fig. IV — Diagram of the experimental delay between input E and output S, highlighting the critical zone 0.10-0.20 s where regulation breaks down.]

Pathological stammering reproduces, without external apparatus, this dysfunction. It translates a defect of directing audition: the ear no longer informs in time. The placement, in the circuit, of a time adapter allows the correction of this delay and, often, the disappearance of the symptom.

[Fig. V — Servomechanical diagram with time adapter and gain adapter interposed on the return loop.]

IV. — The gain adapter

To time corresponds a second magnitude: gain. Phonation supposes not only that audition return in time but also that it return at a suitable sonic level. Too loud, it invades the circuit and falsifies regulation; too weak, it leaves the emission without monitoring information.

Two receptors C₁ and C₂, representing the two ears, share this monitoring. C₁, whose path is short, arrives very rapidly; its appreciation regulates intensity, timbre and allows monitoring of integration. C₂ brings a complementary factor — a time factor, a delay — which causes a third dimension to appear: sonic relief.

[Fig. VI — Simultaneous representation of the two paths: E → S with C₁, C₂ and the auditory centres A.C.₁, A.C.₂ and the centre of phonation C.P.]

[Fig. VII — Temporal profiles of the input E and of the two receptors C₁ and C₂, showing the temporal advance of C₁ over C₂.]

An imbalance of the two receptors — excessive increase of the gain of one or the other — risks giving too much importance to the signal, which then takes the direction of the regulation.

[Fig. VIII — Anatomical diagram with crossed sensory pathways between ears and auditory centres, and direct pathways of lesser importance, illustrating the mutual counter-reactions of C₁ and C₂.]

Sonic relief is owed in part to the presence of two ears; but one can perceive it from a single ear, with less ease, nevertheless with certainty — for besides the crossed sensory pathways (each ear linked to the opposite auditory centre), there exist direct pathways of lesser importance which ensure the mutual counter-reactions of the two receptors.

Without entering into mathematical detail, one can write — for the formula linking the gain G and the time T of the receptors in inverse sense — that there exists a law of the type G·T = constant. The rupture of the gain or the rupture of the time of one of the receptors implies close disorders: no gain, no relay; too much delay, the same. The deaf person who listens to himself too loudly thus encloses himself in a diverted self-information, and tips towards mutism.

V. — Volume, integration, analysis

Beyond regulation, audition operates two simultaneous functions: a global integration which decomposes and monitors, and a more detailed analysis, finer, which bears on a broad sonic band and demands a longer time. It is by the combination of these two functions, and by their differential play, that the perceived dimensions of sound are born: intensity, timbre, height, and what Tomatis calls volume — the contribution of a third dimension, relief.

From this system flows naturally the characterisation of the auditions of the singer: tenor, baritone, bass each correspond to a specific mode of listening, to a particular pass band, to a particular type of self-monitoring.

VI. — Normal audiogram and musical audition

The clinical diagnosis rests on the audiogram. With the help of an audiometer — a sort of electronic tuning fork whose intensities can be dosed at will and which allows the appreciation of the effective affinity for certain ranges of audition — it is possible to characterise in an individual a singer’s ear.

The audiograms of musicians draw a particular silhouette: zone lying between 500 cycles/second and 2,000 cycles/second (approximately middle C and the flute’s C). It is possible to distinguish in it a surplus of audition. The normal audiometric curve is meant to be linear: it is one of the standards adopted.

[Fig. IX — Normal type audiogram: linear curve from 125 to 8,000 Hz.]

[Fig. X — Audiogram of the type musician: ascent in the 500-2,000 Hz zone, slope of 6 decibels/octave at minimum, that is, a gain of 10% of auditory sensitivity per octave.]

[Fig. XI — Auditory deficiency characterising an inaccurate reproduction.]

[Fig. XII — Diagram of a perfectly musical audition.]

To be a musician, in the broadest sense of the term, that is, to hear and to reproduce accurately, does not necessarily imply that one can emit sounds of quality. There exists in audition, or more exactly in the auditory sensitivity of the subject, a zone of appreciation of the quality of the sound to be emitted. This zone, sensational, allows at will the superposition, the filtering, or the encoding of virtually all the auditory information indispensable to the reproduction of any sonic quality.

Audition and register answer to the same problem: the different modes of monitoring, the different diaphragmatic openings — effectively of different apertures — determine the pass bands inherent to each mode of listening (tenor, baritone, soprano, mezzo, contralto, etc.). In summary, to be a musician is simply to have, when listening, an auditory detector abnormally precise in its response curve and especially endowed with a power of emergence for such or such sound.

In another order of ideas, experimentation has allowed the establishment with precision of the bands of selectivity proper to the different nationalities, their characteristics, and we have in fact been led to suspect that there existed adaptations of audition to the ambient environment — notably to the impedances of the air. The plastic factor retained is not acoustic resonance; it is indeed damping. It is in the acoustic bath of the environment — resonance and damping, more or less strong — that the ear must adapt and exploit to the maximum the physico-acoustic qualities of it.

VII. — Ageing, noise, vocal scotomata

But this audition, this auditory permeability of which we have just spoken, can it not be modified? To be sure, under the influence of numerous factors — notably with age — the auditory field narrows, the band of self-monitoring is reduced to the good perception of low frequencies, and the voice faithfully follows the ear which rigidifies. But it is not only age that denatures the ear: there is noise, the most recent of social scourges, which destroys, ages audition, dislocates its auditory curve, mutilating the algos, reducing the perception of low frequencies, blocking relations with the exterior, losing the permeability of information, reducing the quality of self-information, putting the subject out of tune with the ambient environment.

This destruction so specific, the testimony of the aggressiveness of noise upon audition, is translated into a disorganisation of the auditory curve which ceases to be linear, to allow the appearance of a V-shaped notch called the auditory scotoma, whose locus settles directly towards the highs, at the level of 4,000 hertz. Experimentation has revealed that this auditory scotoma was systematically accompanied by an alteration of vocal timbre, which is characterised by the appearance of a vocal scotoma at the same frequencies as the diagram obtained by practising spectral analysis — that is, the harmonic analysis of the voice.

[Fig. XIII — Spectral curve of the voice showing the vocal scotoma at 4,000 Hz, alongside the audiogram of the subject revealing the same notch.]

Conclusion

We were thus led to think, then to verify scientifically, that every vocal image has its corresponding audiometric image on the scale of frequencies. We were then able, in our experiments, to affirm that “a subject emits only the sounds he is capable of hearing”. The observation, of great importance, allows us, depending on the vocal spectrum gathered, to know with certainty the manner of hearing of the emitting subject. With the help of a recorder, we can always, in the laboratory, become aware at any moment, in the course of a career, of the auditory behaviour of an individual — better still, of the very sound of his vocal emission.

Masters of this intimate element, proper to the monitoring of each subject, we have been able, subsequently, thanks to sets of electronic filters, to modify at will the audition of the subject and thus to witness an instantaneous transformation of his voice, his vocal gesture, his attitude, his behaviour.

The profound modifications, conditioned reflexes, that such procedures engage would deserve to be studied more thoroughly, but would take us too far in this article. We shall retain from them only the considerable enrichments they can bring to the sociology of language, to the study of the life and mutation of words, to human relations, to all information.

It is in this framework that we may inscribe the study of the psycho-physiology of music as a language exploiting a sonic information.

It was too brief, perhaps a little rash, to broach the study of the human voice at the beginning of an article that could only be a summary, and of which each of the headings brought into play would demand a much wider development. We nonetheless remain persuaded that this overview will have at the very least allowed an introduction to current research work on experimental phonology and the broad spectrum of its applications, notably in matters of re-education.

Dr Alfred Tomatis

Source: Tomatis A., “La Voix”, offprint, 1962, pp. 225-243 (internal pagination 5-23). Digitised document from Alfred Tomatis’s personal archives. Thirteen figures referenced (Fig. I to Fig. XIII) accompany the original text.