Directed Sound Psycho-structures

p. 210

In the pages that follow, the reader will find an example of a creation directed by a computer. It is a method intended to found what its author calls “Directed Sound Psycho-Structures”: in the disturbing world of reactionary fiction, one might imagine that the dirigisme induced by these structures would make it possible to impose on listeners, in groups or alone, a precise psychological conditioning, by pouring into space judiciously chosen and assembled sounds, sounds to which a kind of “quantum” of action is attached, whose psychic colouring is known in advance.

The authors of the book and the publishers have left the creator of the Psycho-Structures full and entire responsibility for his account. Technically and, to a certain extent, artistically, here is one of the most interesting principles for using computing for aesthetic ends. On the other hand, one could dwell at length on the dangers that an abusive application of such conditioning methods would represent: this is not the only case in which the computer threatens to provide unscrupulous people with a means of action which, like Aesop's tongues, can become the best or the worst of things (readers will have glimpsed some of them in the second part of this book, and will find others in the last): in this respect, the present chapter may serve as an exemplary warning. It nonetheless retains its full interest on the technical level, the only one on which we wish it to be placed.

IV. Directed Sound Psycho-structures

The fundamental aim of the Directed Sound Psycho-Structures is to create sound signals and to organise their appearance in time. The organisation and choice of these sounds rest on both musicological and physiological data. The sound sequence thus generated then has the advantage of being pleasant not only to the ear but also to the psyche, and thereby of being able to give an affective colouring to the nature of the reactions it provokes in the listener.

As the method that precedes their realisation calls on precise knowledge of acoustics, physiology and musicology, we thought it necessary to present either the quantities or the results that we were to use in drawing up the program.

The physical quantities associated with the method

Acoustic signals.

Today it has become possible to arrive at a highly objective and exhaustive representation of a sound signal, whatever its origin (musical sound, noise, p. 211 speech...). The musical acoustics group of the acoustics laboratory of the Faculté des Sciences de Paris has developed particularly advanced equipment that can transform any sound signal into its corresponding object.

The sound is truly photographed, as a landscape might be.

The dimensions linked to this object are:

— pitch: frequency in hertz; intensity: sound level in decibels; duration: in seconds or milliseconds; thanks to which one can grasp in a simple and visual way: the evolution of the dynamic plane; the evolution of the spectral plane; the evolution of the frequency plane.

The devices for obtaining these quantities are, for pitch, the electronic tuner and, for the evaluation of pitch, the sonagraph, the latter making it possible to represent on a single plane the complete temporal form of the sound object.

These means of study, developed by the Groupe d'Acoustique Musicale, make visible notions that, before their use, remained subjective.

On the other hand, the equipment is fully developed and the knowledge of signals complete, which makes it possible to synthesise sounds with any desired timbre by means of a sound synthesiser: the icophone.

Psycho-physiological data.

It would take too long to describe the methods that have yielded these data, but by way of example here are some of the results that we shall have to take into account later.

a) The limit of the ear's fineness of perception is about 1/25 of a semitone, that is, one savart. This limit is, moreover, closely tied to the preceding sound conditions.

b) On the other hand, it may be accepted that the maximum number of items of sound information delivered per second should not exceed 15. It seems that exceeding this number is impossible, that the perception of an overloaded sound message becomes “smoothed out” in the listener's awareness of it, and that it thus loses its power of expression.

c) Let us also note the existence, in the listener, of distinguishable categories of possible memorisation, of varying importance from one subject to another according to his aptitudes and culture, but recognisable in all:

— instantaneous memory (of the order of a few fractions of a second), as if the effect of a note were still shining for that short duration after its disappearance. p. 212 — remanent memory (of the order of a few seconds), in the same way that a flash of lightning leaves an afterimage on our retina.

These two kinds of memory are effective for the listener who, if he asks himself what he has just heard, will find his consciousness responding both microscopically and macroscopically to the structures of the sound edifice he is listening to; they have only a very brief existence, immediately “covered over” by the other sound messages that follow in the hearing of the work, and despite their ephemeral character they are of decisive importance in grasping the musical phenomenon.

— Finally, a global memory, which is not always at our disposal when we would like to call upon it, but which enables us to recognise a work we have already heard, as if the effect linked to it were due to traces, imprints, patterns engraved in our brain which, retraced when the same work is heard again, would recall themselves to our consciousness.

d) How could one fail to mention also these observations reported by A. Danielou in his Traité de Musicologie Comparée:

“If we carefully analyse the nature of the intervals currently used by musicians, whatever the system to which they refer, we discover a strange law: the human ear can clearly grasp and exactly reproduce intervals only if no prime number higher than 5 is found in their internal structure. Thus 15/8 (that is, 5 × 3/2³, the major seventh) and 256/243 (2⁸/3⁵, the limma) are recognisable intervals, whereas 8/7 (the maximal tone) and 7/4 (the harmonic seventh) are not, although the ratios they represent are distinctly simpler. The ear's capacity to differentiate intervals is, moreover, generally limited to seven octaves.” Although it does not concern us immediately, let us also quote from the same author:

“From the point of view of musical expression, we can observe some remarkable properties of numbers when we discover that all intervals of the form 3ⁿ/2ˣ will always appear brilliant, active, masculine, while intervals of the form 2ˣ/3ⁿ will be gentle, passive, feminine.”

e) Let us note, finally, the essential difference between perfectly measurable physical quantities — pitch, timbre, intensity, duration — and the way in which we feel them. One and the same sound object may, depending on its immediate or preceding context, be apprehended in a completely different way. The same duration may, depending on the compositional structure that precedes it, seem to us either very short or very long. p. 213 The perception we have of the notion of time through listening to a musical work is very far from the one we have through our everyday experiences in the world of the visible. It perhaps comes closer to the conception suggested to us by the theory of general relativity, in which time is not marked on an undeformable ruler but on an elastic one.

Certain physical conditions are capable of shortening or lengthening the intervals representing the same quantity of time, just as, for the human mind, certain external stimuli — music among them — are capable of altering its appearances for us.

Everything happens as if, in certain cases, the dimension of time were closer to a ratio than to a simple quantity.

Everything happens as if at certain moments we could apprehend the sensation linked to it through two different approaches:

A first quality of “time” would be noted from the reference trihedron that we adopt in our ordinary experience of the world; let α be this trihedron and S_α the seconds associated with it.

And a second quality of time would be linked to a reference trihedron borne by our psyche. Let Ω be this trihedron and S_Ω the seconds linked to it.

Thus, during an intense affective emotion, or above all during an emotion of an aesthetic order, there would no longer be any effective correspondence, no longer a simple equality S_α = S_Ω. Let us imagine that the duration of an aesthetic emotion is measured in the first reference trihedron: we shall note, for example, x s_(α), the time measured by an external observer. Let us now imagine the time that the subject experiencing the aesthetic emotion believes he can attribute to it: let y s_(Ω) be the measure of time on his psychic reference trihedron.

In the absence of aesthetic emotion we shall have:

x s_(α) = y s_(Ω)

and the more intense the emotion, the more x s_(α) will differ from y s_(Ω).

In fact, these two notions would be apprehended simultaneously, and the hiatus corresponding to them in our mind would be responsible for, and indicative of, a quantity of emotions, together with the triggering of all the associated psychic phenomena. Everything suggests that this point of view can be adopted and taken as a starting model, at least for “measuring” — or more exactly approximating — the effective emotional value of a sound structure. p. 214 f) Let us state, to conclude, that the commonly accepted value of the ear's time constant lies between 5 and 50 milliseconds and that it varies according to the intensity, the frequency and the richness of the timbre.

Musical language: the two principles of its internal organisation

Historically, we observe two clearly differentiated forms of musical composition: modal and polyphonic. Whereas the first was developed by all civilisations, the second was developed only by Western civilisation.

The reason for this is implicitly contained in the nature of the scales available to musicians.

Experience shows that the meaning of a note depends on the ratio of that note to a reference frequency (the role of the tonic), and that a minute variation of this note in relation to the same tonic radically alters the sense of the feeling it evokes.

A. Danielou: “If, while continuously hearing a tonic, we wish to sing the major third in a tender and peaceful mood, we intone the harmonic third E 5/4. But if our mood is energetic, joyful, sensual, we shall inevitably intone the third of the cycle of fifths, E 81/64”.

The interval and the quantum of emotion it releases are inseparable. The Hindu or Chinese musician, who had these categories of fine and precise intervals at his disposal, could develop a horizontal composition (melodic line) associated with a reference frequency (tonic). The composer's role was then to organise the appearance of these intervals in time.

But the Western musician, who for his part had only a limited number of intervals that were not even significant (a consequence of the well-tempered scale), thus found himself unable to develop horizontal composition alone. Little by little, however, he found a solution by imposing the meaning of a note through a synchronous vertical composition by means of chords. As a consequence, polyphonic music p. 215 was bound to develop, since it became meaningful through its compositional dialectic and its internal structures.

The reader now understands why polyphonic music needs to be listened to in order to produce emotions: it indeed requires the attention of the listener, who must make a considerable mental effort to understand in which direction the chord influences the determining note that carries the emotion.

Musical language and information theory

Instead of considering musical composition in the usual terms (melodic or harmonic development, cadence, modulation...), we shall show how information theory can be adapted to describe it, in an entirely different formalism.

Calling as it does on intellectual mechanisms, the polyphonic musical work must keep the listener's attention alert. To achieve this, the composer can play on two distinct quantities: on the one hand his compositional dialectic (semantic information), on the other the nature of the instruments he brings in (aesthetic information).

In both cases, taken separately, he must avoid banality as much as excessive originality. And everything happens as if he had to move between these two extremes: banality, which, though intelligible, lacks interest, and originality, which, if pushed too far, becomes unintelligible; both of these forms lead to weariness in the listener.

Ravel, for example, in the Boléro, compensates for an absence of semantic originality with a constantly renewed aesthetic originality. Indeed, each time it returns, the theme loses its semantic originality, but the aesthetic information is kept at a high level through the diversity of the instruments used. Bach, on the other hand, in the Art of Fugue (Mirror Fugue XIII), develops above all the semantic information: the same theme appears more and more complex each time it returns, while the instrumentation remains the same. Everything happens as if the composer were striking a compromise between these two kinds of information — semantic and aesthetic — with the aim of maintaining a sufficient quantity of unpredictability in the listener's mind. This quantity of unpredictability, linked to the number of signs that carry it, leads to the notions we mentioned: banality, originality, information, redundancy.

Such may be the approach to polyphonic music, a music which finds p. 216 its means of acting on man through the setting in motion of “intellectual” processes linked to listening to it.

But if information theory can be applied to polyphonic music, it is on the other hand far less well suited to modal music. For it is no longer so much a dialectic that is responsible for its action, but above all the quality of the intervals it contains, which, let us repeat, are significant in themselves.

Thus the structuring of sound space that I advocate by means of the Directed Sound Psycho-Structures is in fact a compromise between the two existing musical forms. A compromise centred on the use of intellectual processes developed in close connection with fine, judiciously chosen intervals. In other words, I intend to bring together the causes of action linked to the power of the only two existing forms of music, which until now had never met.

Measuring psycho-physiological reactions as a function of the sound signals emitted

The various points addressed in the preceding paragraphs now give the reader sufficiently precise notions of the quantities we are going to handle, with the twofold aim of describing musical language in terms the computer can assimilate and of describing the method that leads to the realisation of the Directed Sound Psycho-Structures.

We have all, at one time or another, been seized by an emotion provoked by hearing a musical work.

We can therefore recognise that sounds assembled according to a certain arrangement can trigger reactions of a psycho-physiological order.

The musical fact can be divided into three distinct parts:

1) organisation in time of a sequence of particular categories of signals;

2) transport of these signals by means of sounds which themselves carry a new category of signals linked to their physical nature;

3) reception by the physiological ear.

On the other hand, we recall the example already cited of the harmonic third C 1/1, E 5/4, to which a tender and peaceful feeling is attached. It is difficult to accept that such a feeling is intrinsically contained in this interval; on the other hand, one can conceive that the C carries p. 217 a certain signal and that the E carries another: the ear first perceives these signals and the psyche then integrates them. It would be this integration that transforms these signals into an aesthetic message.

The signals carried during the hearing of a musical work then fall into three categories:

— category I signals, linked to the semantics of the composition of sounds in time;

— category II signals, linked to the physical nature of the sounds;

— category III signals, linked to the value of the sound intervals.

The musical phenomenon being thus described, we see that it comes down to this diagram:

TRANSMITTER: Computer, Order manipulator, Semantic information, Category I signals. CHANNEL: Aesthetic information, Physical nature of sounds, Category II signals →; Carrier of cat. I signals →; Sound intervals, Category III signals →. RECEIVER: Physiological ear →. REACTION: Psychic integration ⇢ Aesthetic message.

Let us now consider category I signals alone. Organising the appearance of such signals in time can be reduced to two aspects: either the signals are emitted one after another, or they are emitted in successive “blocks”. Here we recognise the two principles of modal and polyphonic musical composition. Information theory can be adapted to the vertical structure and to the horizontal structure independently of each other.

If we temporarily forget that these signals are to be carried by sounds, we can apply whatever treatment we wish to them. This amounts to considering them as abstract signals.

One can then easily imagine the mathematical treatment which, by means of the computer, makes it possible to arrive at the equivalent of a strip on which points would be marked, each point having the value of a signal. An appropriate handling of information theory having determined p. 218 the appearance of these points, the semantic content of the strip will thus be perfectly known as regards its quantity of information.

The immediate aim being to test the qualities of the psycho-physiological reactions associated with the appearance of perfectly quantified signals I, II and III, the preceding strip takes on its full value if it is transformed into a signal generator. To do this, one need only plot a time scale on its x-axis and a frequency scale on its y-axis, then couple this strip to an optical reader, itself connected to a sound generator (or else work directly on magnetic tape).

Consider the following experimental set-up:

Information Theory / Computer → “Semantic” strip (Signal I) → Optical reader → Signals I, II, III → Sound generator → I, II, III → Receiver: Physiological ear.

The method resulting from such an overall principle is remarkably flexible in use. Thanks to it, and to the way in which we have posed the musical fact in its entirety, the reader will see that we are in a position to test and measure the relative importance of the psycho-physiological reactions attributable to the categories of signals I, II or III. For, although they are closely linked, the relative importance of some can be reduced at will in favour of a single one. We are even able to study, for each category of signal, the influence of its descriptive parameters taken separately.

A - We wish to study signal category I.

This therefore means reducing the influence of categories II and III. To make the influence of signal II negligible, the sound generator will work at constant timbre (the aesthetic information will indeed tend towards zero).

As for signal III, working on broad frequencies will eliminate the very source of its effect.

B - We wish to study signal category II.

Signal III is treated as in A.

Signal I has a minimised role: either the semantic information is infinite (total unintelligibility) or it is zero (complete banality). p. 219 C - We wish to study signal category III.

Signal I is treated as in B.

Signal II is treated as in A.

By way of example, here are some of the studies that can be carried out:

— Classification of Psycho-Physiological Reactions (P-P.R.) according to the semantic patterns proposed, and vice versa.

— With constant semantic content: study of the influence of the pitch of sounds on P-P.R. (It suffices to slide the y-axis scale of the optical reader.) Study of P-P.R. as a function of the quantity of signals delivered per second (by simple dilation or contraction of the time scale).

— From a strip specially designed in the horizontal semantic sense: measurement of the exact influence of a drift of the tonic on the P-P.R.

— Role of the signals according to the memories on which they rely (immediate and remanent memory); influence of the intensity of the sounds.

Through systematic work of this kind, one is able to pin down precise qualitative results relating P-P.R. to signals I, II or III. But since sound language calls on all three categories of signals together, they must be studied without departing from a structuralist point of view, that is, knowing that the relations uniting these different kinds of signals may have a greater effect on P-P.R. than the signals themselves. That is why, at a second stage of study, they must be mixed together and the quality of the associated P-P.R. observed. (Precise experiments have shown the existence of reversals in the values of P-P.R.).

As for the measurement of P-P.R. as a function of the signals, it can be approached from several angles:

— recording of the breathing rhythm;

— recording of the heart rate (plus blood pressure);

— electroencephalogram tracing;

— general agitation;

— assessment of the times S_α and S_Ω.

Ideally, for the study to be exhaustive, these measurements would be made on subjects who are awake and on subjects who are asleep.

In their final phase, the Directed Sound Psycho-Structures may resemble the diagram shown, subject to a few modifications not of principle but of detail. p. 220

BASIC DIAGRAM OF THE DIRECTED SOUND PSYCHO-STRUCTURES Elements of the diagram: “manipulation console”; “COMPUTER”; three chains leading from the computer: (1) “electromagnetic exciter” (gong, triangle, bell, cantor [?]) and “microphone”; (2) “magnetic tape” and “reader”; (3) “plotter”, “graph strip”, “bank of photoelectric cells”, “complex frequency generators”, “mixer”; then “amplifier”; then, in the “PLACE TO BE CONDITIONED”: “remote control”, “amplifier”, “distributor”, “loudspeakers” (HP) and “environment sensor (crowd density, agitation, etc....)” connected back to the manipulation console. p. 221 Experiments carried out personally with very limited material means, or with the assistance of musicians and composers, have led to several concrete results. Let us cite the most significant of them:

— Semantic information is all the better conveyed, all the more perceptible, and therefore all the more effective, when it rests on sounds with simple timbres — synthetic sounds are particularly suitable. One can play in very unexpected ways on intensity gradients. The parameters that can carry this information, singly or together, are: rhythm, intensity, frequencies and slight differentiations of timbre.

— Aesthetic information is all the more effective when the partials and harmonics of the signals it uses are evolving; it varies with the organisation of the sound spectrum and its own originality. The presence of partials among the harmonics, and the energy they carry over time, are decisive for P-P.R.

— Information linked to the intervals of the signals is all the more perceptible when they lie between 500 and 3,000 Hz.

— Immediate and remanent memories can be modified as regards their duration of action, and in notable proportions.

Semantic information can be centred on one or the other of these memories, or on both. The nature of the P-P.R. depends closely on the memory targeted through the sound micro- or macrostructures used.

— The means that produce a state of tension, or of well-being and relaxation, have already been parameterised.

— A particular semantics centred on mechanisms of anticipation associated with recurrent reasoning is under study and appears very promising for producing a feeling of well-being.

— The localisation of sound sources by the listener can be of decisive importance for the nature of his reactions. Discrete sound sources distributed in space, whose relative emission intensities in energy vary over time but which each broadcast the same sound message, quickly trigger a feeling of gentle well-being.

These results, obtained in a very rudimentary way, are already enough to leave no further doubt about the effectiveness of the Directed Sound Psycho-Structures in their final stage.

Compared with the traditional conception of music, the Directed Sound Psycho-Structures propose numerous innovations: p. 222 — The compositional dialectic is neither artificial nor hemmed in by a conservatoire mentality, but dictated by the quality of the psycho-physiological reactions expected from it. Its purpose lies in its action on our innermost self, which is not incompatible with an aesthetic feeling.

— The work thus generated will have neither beginning nor end, while never repeating itself.

— It will not need to be listened to, only heard.

As for the possible applications that these Directed Sound Psycho-Structures offer, they are many: first, in the forms described above, as an experimental tool, but also in many others.

Every innovation has positive and negative aspects. The Directed Sound Psycho-Structures are no exception to this rule, and it must not be concealed that their use presents certain dangerous possibilities. This duality appeared only along the way (whereas the initial aim was to have a Computer / sound-production chains / terminals system able to “colour” certain places artificially, public places in particular), and the examples of application that follow have been chosen precisely to underline this ambiguity.

The Directed Sound Psycho-Structures (P.S.S.D.) could be introduced in many places, even in the Paris métro. One could indeed think of installing the necessary loudspeakers in the corridors, in the stations and perhaps even in the carriages, connected to the central computer from which the sound information associated with a feeling of relaxation would be sent out.

One could also envisage using these P.S.S.D. operationally, for the purpose of group “handling”. The airports of tomorrow, which will see 500, 1,000, 1,500 passengers disembark within a few minutes, must receive them quickly enough not to resemble a métro platform at rush hour. Until now nothing could make passengers naturally hurry to leave. Why not introduce gradients of unpleasantness or pleasantness in time (sound signals) and in space (influence of the localisation of discrete sound sources) to guide groups from the least pleasant places (tension) to the most pleasant places (relaxation)?

The reception halls of large companies might also be suitable for such P.S.S.D. Later studies might decide on the diffusion of particular states in offices, in correlation with the work being done.

In music therapy, the practitioner could have at his disposal an experimental tool remarkable for the flexibility and precision of its handling. p. 223 In an entirely different field, a theoretical device has already been devised to visualise the structures released into space. Starting from this realisation associated with the P.S.S.D., one can quite easily imagine a total spectacle in which they would be combined with a coloured vision of space and with choreography (the dancers would improvise to the P.S.S.D.). The receptivity of the spectators would be all the greater since the dancers would be under the same sonic influence as they are.

Finally, let us point out that it is possible to integrate data coming from outside (crowd density, various movements...). One would then have a feedback-controlled system, that is, one that would act on an environment and draw directives from it in the form of information acting on the parameters of the program.

As we have seen, the Directed Sound Psycho-Structures rest on a thesis arising from a multidisciplinary convergence.

This undertaking could only be envisaged with the help of the computer and the possibilities of manipulation, information processing and work design that can be achieved through it. All the facts, all the syntheses were conceived in this direction from the outset.

The use made of the computer can be schematised in five functions:

I — The computer as worker, a rapid “mixer” of data.

II — The computer as innovator, which makes it possible to observe the variations of the final result as the parameters of the basic program are varied successively or simultaneously (scanning of the “possibles”).

III — The computer as manipulator of chance (limited use).

IV — The computer as integrator of data coming from outside.

V — The computer as “creator”, which makes it possible to give works of art new dimensions, to make them evolve and develop according to the environment surrounding them and the reactions it provokes.

Are the Directed Sound Psycho-Structures a work of art? In the sense of our time, I can answer: yes, provided they are used only for aesthetic ends.

Only the instrument, the intermediary between the artist and the finished work, has changed in nature: it is no longer the brush or the score, it is here the computer.

Jean-Marc PHILIPPE.