Module 7: Sound Theory and Sound Practice

Sound Theory

What we hear from the screen is not an image of the sound, but the sound itself which the sound camera has recorded and reproduced again …. there is no difference in dimension and reality between the original sound and the recoreded and reproducted sound.

— Béla Balázs The Theory of Film

…. in a photgraph, the original is as present as it ever was.  Sound can be perfectly copied….. the record reproduces the sound.

—- Stanley Cavell The World Viewed

And it is true that in the cinema —- as in all talking machines — one does not hear an image of the sounds, but the sounds themselves.  Even if the procedure for recording the sounds and playing them back deforms them, they are reproduced and not copied.

—- J. L. Baudry The Apparatus

There is no ontologiacl difference between hearing a violin in a concert hall and hearing it on a sound track in a movie theater.

—- Gerald Mast Film/Cinema/Movie

Auditory aspects, provided that the recording is well done, undergo no appreciable loss in relation to the corresponding sound in the real world: in pronciple, nothing distingushed a gunshot heard in a film from a gunshot heard in the street.

—- Christian Metz Aural Objects

….. it is never the literal, original “sound” that is reproduced in the recording, but one perspective on it, a sample, a reading, of it.

—- Alan Williams Is Sound Recording Like a Language?

Revealing its mandate to represent sound events rather than to reproduce them, recorded sound creates an illusion of presence while constituting a new version of the sound events that actually transpired.

…..[recordings have] only partial correspondence to the original event.

—-Rick Altman “The Material Heterogeneity of Recorded Sound”

That a gunshot seems to sound the same in the different acoustic spaces of the street and the inside of the cinema is a deception…..

….. familiarity has dulled the capacity to recognize the violence done to sound be recording.

— Tom Levin “The Acoustic Dimension”

Original/Copy: Sound Theory?

Despite their obvious disagreement, these two sets of quotations illustrate clearly that within the study of sound recording, especially as it functions in the cinema, the relationship of original sound represented sound is a (perhaps the) fundamental issue.  I have separated the groups according to the position each takes in regard to the original sound and its copy are irrelevant —- some, Metz, on what appear to be functional grounds others on what could very broadly (and with caution) be called “ontological” grounds.

It is significant that the apparent struggle between these two opposed theories is not carried out solely within the field of film theory or within academic context in general, but within the realms of “practice” and “basic research” as well. A surprising amount of self-conscious theorizing is exhibited by sound technicians in and around the film industry with regard to the problems, goals, and status of the representations they produce.  In effect, there is a theoretical struggle at work in the debates among technicians over “proper” technique and its relationship to the form of realism they define as constitutive of the film medium. Indeed, this struggle is indicative of the contradictions within the representational system itself.  Although it would be foolish to suggest the “theory” plays the same role in the discourse of researchers and technicians as it does in academic discourse, in each it plays as an important role, providing an intellectual framework which allows the participant s in the discourse to perceive, understand, and evaluate representations as instantiations of some more general function or phenomenon.

Addressing perceived problems in such theories.  Tom Levin attempts to perform a “critical analysis” of the sound apparatus, “….in order to understand how sound [like the image]… is transformed in the process of its reproduction” (1984, 56).  Strengthening his objection to theories which ignore these transformations, he maintains that it is critic’s task to reawaken the “capacity to recognize the violence done to sound by recording” (66).  On the need for critical analysis, I am sure we can all agree.  However, the assumption that there exists a sound which is pure and present, which exists unviolated or wholly prior to the “transformations,” a firm ground which is more real than its recording, may itself require some critical analysis.  Similar assumptions underlie the Williams and Altman quotations.  All assume a standard sound which is transformed or violated in recording, which is “digested”, pre-digested,” “blunted,” “spit out,” “neutralized,” “conventionalized,” and “homogenized” by the technology “itself,” and necessarily by any application the technology (Levin1984, 67).  For all three, recording results in an “essential transformation,” a certain “partiality,” or a “loss” of the original sound.  It is importance of the notions of the “original” and a single “technology itself” for theories of sound representation which I will address in this essay.

All three theorists stress the materiality of sound events, understanding by the term “sound” the “entire volume of vibrating ait” (Williams, 53—4) within the space of production.  By posting such a definition, Williams indicated that every sound is spatiotemporally specific, or in a broad sense of the term, historical. Given that a sound is inseparable from the time and space of its production, each sound becomes an essentially unrepeatable event—- an event distinguishable from all others.  Levin makes the point of specificity most forcefully in the following quotation:

….if a sound is understood as a volume of vibrating air, then assuming for a moment that sound reproduction were absolutely flawless …. such a “perfect” reproduction of sound waves in a different volume would effectively constitute a different sound (62)

Figure 7.0 University of Colorado Boulder ATLAS Classroom

Figure V7.0 The Clintons in Dallas 2017. Shot on iPhone 7

That a gunshot seems to sound the same in different acoustic spaces…. is a deception. If differences remain unnoticed, this is a function of a socially constructed auditory practive which emphasize the similarity of such sounds in order that they be understood (i.e. linked to a common source) by the hearer. (1984, 62)

Metz’s claim that “nothing distinguishes a gunshot heard in a film from a gunshot heard in the street” (29), evinces a rather different understanding of film sound.  Metz is surely aware that the experience of hearing a gunshot in the street hardly resembles hearing one at the movies.  In fact, as Altman points out, they resembles hearing one at the movies.  In fact, as Altman points out, they resemble each other in one respect only —- in name — an important point, since one cannot even say that the source of the sounds is the same.  Anyone who has ever attempted to post-sync a gunshot for a film can tell you that there is apparently indicating is that in are dozens of acceptable substitutes — many of them more acceptable than an actual gunshot.  What Metz is apparently indicating is that in a functional sense, or in a narrative sense, what matters is the ability to identify the source of a sound.  In other words, the sound event must be legible— it must be recognizable across a series of different contexts as a culturally defined signifying unit.  Against the definition of sound as an essentially unrepeatable event. Metz describes sound as an eminently repeatable and intelligible structure.

For Metz, it would be fair to say that sound, insofar as it is recognized as sound, is always already a social/semiotic phenomenon, not a purely sensual or phenomenal one.  It is arguable that the very designation “sound” implies not just any sonic phenomenon, but one which already bears the traces of its social pertinence.  In other words, the simple fact that something can be identified as a sound, as a particular type of sound, necessarily involves some social or cultural dimension — the very possibility of its having an identity presupposing its social character.  Metz rightly comes under fire for reducing the variety of film sound phenomena to a simply narrative function, but he comes very close, in my opinion, to describing the actualities of Hollywood sound representation — a point I will demonstrate at length below.

While it is easy to understand how every sound is unique.  Metz’s assertion that nothing distinguishes the original from the copy makes sense only in context.  A sympathetic reader might take Metz to be saying, “Given this particular form of narrative whose role is strictly defined within a particular economic system, in order to ensure maximum legibility so that they may take their place as clearly defined elements in a narrative structure.  Given these conditions, to the extent that two sounds can be recognized as deriving from the same sources and hence, of performing the same function, they are interchangeable.”

Figure V7.1 The Great Train Robbery, 1903

Levin is far less sanguine and somewhat more concise about the effects of Metz’s formulation:

Avoiding the thorny problem of whether learned perceptual habits are automatically social constructed” (let’s assumed that they are), the term “deception” is probably a bit stronger than most of us would care to use.  In fact, although Levin’s assertion that “phenomenology obfuscates ontology” when dealing with representations seems an adequate rebuke of Metz’s assertion about filmic motion, Metz’s argument about sound is precisely not a phenomenological account, but a semiotic or structural one holding that sound in the cinema is a socially produced phenomenon, as illustrated be related claims in the article “Aural Objects.” The use of the term “deception” does, however, point to the sour of stakes involved in the original versus copy or original verses repetition relationship, but to keep those stakes high, Levin and Co. must assume an identity and fullness to the original event which is irretrievably lost in the reputation.  In almost Platonic terms, there must be a decrease in both truth and being as one moves toward the copy—- necessarily assuming some “real” which was, itself unmediated.

Original Sound: Recording/Representation

By defining sound recordings as partial, transformed, or to some degree absent with respect to the original, Williams, Altman and Levin uphold this hierarchy with different degrees of rigidity.  Sensing an almost theological undercurrent to Adorno’s argument as he adopts it for his own critique, Levin points out that the “analysis is not a nostalgic ‘fall’ story, but a narrative of transformation” (66). Nevertheless, if the original/copy distinction were not one of absolute and irreversible decline, was it simply “transformation,” why would recognition of similarity between original and copy of repetition be a deception?  And further, if the identity of a sound and its repetition can only be a “function of a socially constructed auditory practice” producing a deception, there must be some non-constructed experience, a pure perception, against which to measure this as a deception.  If there were no such experience, and we were to assert, as I will, that all practices of audition are equally constructed, there would be no valid reason for suggesting that one socially constructed practice grounds the discussion and evaluation of all others.

To be sure, there clearly is a casual relationship between the “original” sound and its recording, but there is no logical or essential reason why that particular relationship should be the privileged one in theoretical discussion.  Given that these theorists stress the absolute uniqueness of every producing and auditing situation, even the “original” event is hopelessly multiple and differentiated in its ‘singular’ occurrence.  Since every two seats in an auditorium, for example, can legitimately be distinguished sonically, there can be no absolutely definable original against which to measure the transformations of a recording.  Indeed, the original sound can be posited as a theoretical absolute lonely through the dogmatic assertion that one socially sanctioned experience of the sound, its “best” presentation, is somehow “logically” the essential nature of the sound in question.  In fact, the primary ideological effect of sound recording might be the creation of the effect that there is an “original” independent of its representation.

The endorsement of symphony listening as the sine qua non  of sound theory has far-reaching consequences because it assumes that all perceptible differences are always equally relevant.  When listening to speech, for example, we register differences through categories such as “accent” and “emphasis,” but we always disregard certain differences between different performances of the same words or sentences.  Yes, there are “material” differences between these performances, and yes, it is significant that we do ignore some differences but not others, but these conventions of speech recognition can hardly be called “deceptions” in a meaningful sense.  Likewise, I would not say that a friend who had listened at home, on the radio, to a concert I had attended was deceived in believing that he had heard the same concert because I would have to say the same thing to anyone seated in a different section of the theater, either of whom may have had a “better” point of audition that I.  Is it not more of deception to believe that one of us had an experience which legitimately serves as a reference point for the analysis of the others?  We simply heard the “same” concert differently.

Figure V7.2 The Sound The Universe Makes, Janna Levin on TED-Ed

The sonic differences to which we attribute significance are always contextually determined, hence to single context provides a reference point for theorizing all others.  If the recognition of differences between a sound and its representation (or between two different instances of the “same” sound, since Levin offers this possibility explicitly) were simply a matter of recognizing differences pure and simple, there would be no reason to order originals and representations hierarchically.  By definition, hearing a recorded sound is just as unique as an experience as hearing one “live” —- a situation hopelessly complicated by the use of recorded sounds and electronic enhancement in a live performance.  Does the insistence on the original not surreptitiously and needlessly introduce a privileging of presence over mediation rather than simply illustrating transformation?

What is at stake, finally, is the importance granted to the original sound in any theory of sound representation.  In the pieces Levin refers to, Adorno is discussing symphonies (a highly conventionalized form of sound practice) and the alterations to musical structure wrought by recording and broadcasting over the radio.  Adorno defines serious music by its structure — by the specificity of its part/whole relationships, (Adorno 1945; 1941) hence, any acoustic transformation which affects the perception of that structure is, by definition, significant.  However, crucially, he is nowhere talking about the “technology itself” and what it does to sound in general.  One could easily conclude that for Adorno, there is no essential transformation of popular music through recording since the “loss of tone colour and intensity, a reduction of the overtone series, a compression accompaniment, an exaggeration of contrasts, etc.” are already structurally a part of popular music.  These attributes may, in fact, be enhanced by recording.

Sound recording is important to Adorno when, through transformation, technology impedes both the presentation and the perception conventionally felt to ve appropriate to a particular social or cultural sound practice —- the symphony.  And he seems very right.  He does not, however, say that this is only “technology’s” fault.  I feel no hesitation in suggesting that Adorno would have been just as concerned had a new fad in architecture produced concert halls that causes similar transformations of the music.  In fact, listening from the lobby of a hall accomplishes many of these same transformations.  I’m certain that Adorno would find the lobby as poor a point of the audition as the individual living room.  The question, for Adorno, is not one of “original versus copy.” A poor original is no better for him than a copy if it similarly distorts the work.  It is not a question of technology mediation, but of degenerative mediation, whatever the source.

Adorno is concerned with the perceptibility of pertinent elements, given the dictates of a particular culturally defined form— the network of interlocking cultural practices that shapes our relationship to the staging, presentation, perception, categorization, and evaluation of a particular set of sound, and therefore determines which elements are pertinent to this particular use.  However, the desire, in music listening, for the most delicate of differences to be audible is not a general model for how we deal with sounds.  Like all forms of the audition, music listening ignores certain features of a sonic event in favor of others.  In addition, serious music is normally presented in locations designed specifically to hide certain kinds of imperfections and to enhance the overall quality, further complicating the situation.  Is lush tone color truly a characteristic of the “music” of the romantic period, or is that a property of the concert halls constructed to produce or enhance this effect?  Where is the “original” here?

Figure 7.2 The Aesthetic Experience in Venn Diagram

All recordings do transform, in some measurable way, the sounds they represent, but they do not always transform them to the same degree of in the same way.  In fact, in certain situations, the resultant transformations might actually be a benefit, even for music.  It seems arguable that Adorno might even praise recordings if, once certain technical failures had been lessened, they facilitated a new, more analytical appreciation of the classics.  I can imagine that given his interest in critical analysis and in pedagogy, he might consider a recording superior to a concert situation because it isolates and makes available for repeated listening what would otherwise be a fleeting experience.  At several points, he does mention the transformations of listening in a generally negative way, but I feel fairly certain that he would not necessarily be opposed to them. They need not entirely serve any single ideology master.

So it should be clear that absolute fidelity to the “original” sound is only one obvious, possible goal for representing sound.  Now, I take it for granted that people in a movie theater are never actually fooled that they are present at an “original” sound event when it comes over a loudspeaker—-as much as institutional effect of going to the movies as it is a matter of the technology factors of sound recording.  The fact that filmgoers take a recording of a gunshot to function within a film as a gunshot is no more or less deception than taking a flat in a theater for the rear wall of a room. It is, as Levin says, a “socially constructed auditory practice” suited to a particular cultural form which is supported by a whole network of other socially constructed practices (like facing the stage, or not talking during a performance).  Altman and Williams have taught us to hear how all sound recordings are staged representations (i.e. socially constructed). For the moment, then, let us simply ask whether the decision to perform an orchestral work in an auditorium, and the ability to hear differences from performance, are somehow not socially constructed auditor practices, or in other words, conventions? Is the performance of a musical piece in an auditorium not itself a mise-en-scène of sound possessing each and every attribute that is used to differentiate the reproduction from the “original”? Ultimately I will show that the attributes claimed to be indicative of the “copied” sound is, in fact, structural necessities of any staging of a sound, even an “original” sound, and in fact are necessary prerequisites for the identification of any sound as being an “original.” All sounds, recorded or live, are equally “inscribed.”

The Technology “Itself”

The force of Levin’s critique derives from his assertion that, despite effective studies of the uses of technologies and the ideologies implicit in these uses, what still remains to be done is an analysis of the transformation produced through the technology “itself”—- that is, independent of any particular use of it.  In order to make this claim, Levis must assume the existence of “the technology of sound itself” (1984-66) as an object which can be theoretically and practically abstracted from the network of relations and practices which situate it.  It is significant, and perhaps symptomatic, that he defines “technology” as an object, rather that as a set of cultural and social practices, as I tend to do.  Through a discussion of Heidegger, Levin claims that the cinematic apparatus is part of the larger process of objectifying through the structure of representation (Vorstellung) which Heidegger’s sees as a characteristic of modernity. However, I do not believe that Heidegger’s text so easily supports the conclusion that it is the device, the technical object itself that frames the world in an objectifying manner.  Although I would agree with Levin that “irrespective of the particular content of the image, the photograph is a representation already interpreted, selected and ideologically”framed” by its very technology” (65), this does not mean that ar every historical moment and in every historical use the “very technology” is “the same.”

A Japanese interlocutor uses Heidegger’s own example of a non-Western alternative, Rashomon, to claim that even here, “the Japanese world” has been “especially framed for photography” (Heidegger, 1971).  Framed for photography and not by the camera. Levin seems correct in assuming that, as a product of the dominant structure of representation throughout Western culture, the camera will no doubt be formed to its purposes, but it is not entirely defined by them, nor is it even entirely self-defining.

Figure 7.3 Rashomon Movie Poster

The very term “photography” implies a representational image, of an existing photograph — defined as images made on light-sensitive paper (and let’s even assume, with a camera) doesn’t fit this “normal” understanding of the term “photography.” Would the technology “itself” objectify the Japanese world if, for some reason, the camera never showed the actions pertinent to the narrative? If the camera strayed toward the crew?

A symptom of the generalized technicity typical of modernity, photography “frames” reality as entirely commensurate to human understanding and as manipulable to human needs. In the vocabulary of Heidegger’s technology essay, it constitutes the world as “standing reserves.” Technology, therefore, does not “itself” enframe objects, because it, too, is something of an expression of the dominant historical configuration he refers to as Ge0stell, a positing of the world which frames it as available for human purposes,  Any relationship (not only those mediated by a technological instrument) which seeks such mastery of the world is indicative of technology be used in a particular way.

Figure 7.4 Rashomon Lost all faith in the human soul

In the case of photography (and by extension, sound recording) the production of a representational image understood as “a systematic structure, a model of formation (Gebild) devise by man serv[ing] to explain that which it represents,” seems to embody the “frame-up” that Levis decries.  Thus understood, the term “photography” refers not simply to a device, but an entire system of relating to the world of objects commensurated-to-man, implying a vast network of practices surrounding the use of the device in order to produce images that represent, and explain through representing.

When we say “photography” we refer to just a such a network of interrelated practices whose parameters are defined by standards such as “professionalism,” and so forth.  Because of this fact, when Levin refers to a recording of a symphony I can assume that such a recording was mand with microphones directed at the orchestra rather than at the audience from a distance, rather than from the bell of the third trumpet, and in a relatively unobstructed space, rather than from under a chair in the last row of the balcony.  Much of the work of “representing” a sound occurs prior to, or independent of (but compatible with) the technology which does the representing.

If the film were exposed so that only very vague shadows were produced? If the film were exposed so that only very vague shadows were produced? Heidegger is here claiming that the entire process of staging a narrative with actors for a camera, is destined for a larger audience, and objectifies the world by making it subservient to the demands of a particular representational structure inappropriate to it.  What this entire industrial mode of production (the production of the world as object for consumption in a form (the photograph) which is universalizing and trivializing in the same measure.

Levin asserts, in response, that “it is not enough to consider a variety of sound practices because no matter how radical and innovative they may be, they are ultimately put in an apparatus which spits them out again in a digested, blunted, and conventionalized form” (1984, 67, my emphasis). I hope I have illustrated, however, that there is nothing-to-be-representing, even by means of technology, arranges its objects before it in such a way as to make it representable.  So, representing music can be represented “accurately” (to choose one standard). The concert hall itself is an apparatus that “frames” music in a certain way, facilitating the appreciation of its defining features and helping to define its “essence.” The original is “formed” by the hall apparatus, and this forming is an essential first step in presenting the object (“music”) for representation. The act of representation, by selecting only certain objects or objects in a certain form, from a certain point of view, pre-structures its objects for the device.

Figure 7.5 Lotte Concert Hall

Original/Copy: Sound Practice

I have effectively polarized the sound theorists cited at the outset according to the general schema event versus structure, original versus copy, and unique versus conventional.  By turning to an early, transition era debate over the representation of film sound space I believe we can clarify the conflict a little.  From the point of view of the studios, the basic technical and aesthetic task of a film sound system was to record and reproduce precisely synchronized, intelligible speech and music in a form that could be mass-produced.  Despite the fact that the majority of early Vitaphone films were based on the synchronous recording of music (and Warners may have initially had no plans to make “all-talking” features), all sounds were functionally subordinate to speech in an important sense or were treated as species of a more general, speechlike situation, the most important aspects of a sound being its origin, identity, and comprehensibility.  the essential problem was to produce the effect of appropriate and recognizable sounds emanating from the objects on the screen, as if the sounds were, like the voice, a property of the depicted object.  However, the model of realistic representation which guided research and the goals of practice (at least as it was given expression in various technical journals) stressed absolute fidelity as the ultimate goal.

Two models of sound recording dominated, and to a remarkable degree continue to dominate thinking about sound representation.  Very broadly, these could be called the “fidelity” model and the telephone model.  The former stressed the importance of representing a spatio-temporally specific production of sound, the latter stresses intelligibility or legibility at the expense of specificity, if necessary.  A recording of an orchestra, for example, tries to preserve the cavernous space within which the sounds were produced, while a telephone, being designed with a very different social function in mind, would never be suitable for this.  However, a “fidelity” recording could very easily represent one of the “worst” seats in the house from which to experience a sound event such as a concert, while still remaining “faithful” or specific.  A telephone, in contrast, sacrificing specificity, renders speech rather clearly.  For each model, the very conception of what a sound “is” differs in important ways. The “fidelity” approach assumes that all aspects of the sound event are inherently significant, including long or short reverb times, rations of direct to reflected sound, or even certain peculiarities of performance or space.  The “telephonic” approach, not literally limited to telephones and voices, assumes that sound is internally hierarchized —- some aspects are essential while others are not. As with our theorists, the relevant terms of comparison are uniqueness versus recognizability, or event versus structure.

Figure 7.8 Vitaphone

Figure 7.7 The Fidelity Model Demonstration

Within the classical Hollywood style, there exist types of sound representations that correspond to both the “telephonic” and “fidelity” modes.  Schematically, dialogue recording tends almost uniformly, from the early thirties on, to minimize the amount of reverberations, background noise, and speech idiosyncrasies, as it simultaneously maximizes the “directness” or “frontality” or recording, and the intelligibility of the dialogue.  Even when a speaker appears to turn away, a high level of direct sound often implies that he or she is still speaking “to” the auditor because speech is understood not simply as an abstract sound, but as a sound with a specific function —- a narrative function.  This form of recording differs markedly from what Rick Altman had dubbed point-of-audition (POA) sound.

POA sound (like the POV shot) represents the experience of hearing within the diegesis, normally the hearing of a character.  Sometimes this is indicated by muffling sounds, as if hearing them through a wall, by including a Dopper effect to indicate the rapid passage of a sound source, by an increase or decrease in volume indicating the approach or retreat of the source, or perhaps by giving a clear sense of the acoustics specific to a particular location.  All of these different phenomena can ve assimilated under the rubric of what Alrman designates “spatial signature” — those indicators of the spatial and temporal specificity of sound production and reception which characterize any recording as unique.  It is quite significant that in both the case of the POV shot and the POA sound, a representation of a space taken to be “real” within the diegesis is at stake.

Figure V7.3 Apocalypse Now The Dossier

I would suggest that in Hollywood films visually represented space, when not explicitly marked as a POV shot, is primarily a marker of the narrative or enunciative level of the presentation.  In other words, its dimension, angle, framing, and so on are not understood as someone’s actual vision (narrator or character), but as a means of narrative emphasis.  Unlike the POV shot, which is marked as a unique perception in a specific, diegetically real space, a typical shot can usually be replaced by another shot representing a noticeably different view of the space — its index of substitutability is much higher since it presents narratively important information rather than perceptually specific information.  It is indicative, however, that should technicians who call for perceptual fidelity in sound and image representation seem to employ an implicit “invisible witness” model of cinematic narrative.  Such a heuristic seems obviously to encourage one to think of filmmaking as a stringing together of unique perceptions, as I will attempt to show below.

Now, while POA sound is not equivalent to sound bearing a spatial signature (since all recordings include “signature” in one way or another) it is one clear instance wherein the spatial characters of sound are allowed” to manifest themselves (often, significantly, for narrative purposes).  In general, close frontal miking of actors, which minimized reflected and indirect sound, became the norm for dialogue since, as one researcher put it, “In no case did an increase in reverberation cause an increase in articulation [intelligibility]” (Steinberg, 132). Or as another phrased a related observation, “The quality of a recording is effectively independent of the reverberant characteristics of the set if the microphone is within approximately three feet of the speaker” (Hunt, 482).

Figure V7.4 Apocalypse Now The Captian’s Mission

Together, the two quotations indicate that a “good” recording could be associated with recognizable speech sounds, and dissociated from any strict sense of fidelity.  In related studies, other researchers sought to qualify “Good Voice Quality,” or intelligibility (“articulation”), extending their (historical) standards of correctness through the apparatuses of science, bringing them to bear directly on the voices and bodily techniques of performers.  Other researchers from the period asserted that “naturalness” seemed to be equated with the presence of reflected sound, while “articulation” was associated with direct sound (Hunt, 499).  In other words, indications of spatial specificity are related simplicity of certain forms of unintelligibility or at least the possibility that fidelity and intelligibility are not necessarily related.  Although my argument here focuses heavily on speech, I would suggest that terms such as “intelligibility” have analogs in the case of sound effects as well, where something like “recognizability” or “identifiability” performs roughly the same function —- that of comprehension. A recording with a high degree of reflected sound, or some other indicator of spatial signature, is linked to sound considered as an event while closely-miked sound, with a relatively “contextless” spatial signature, is linked to sound considered as an intelligible structure—- as a signifying element with a larger structure.

According to Carl Dreher, there need not be any extreme conflict between the two, however:

Since the reproduction of sound is an artificial process, it is necessary to use artificial devices in order to obtain the most desirable effects.  For example, it is normal procedure to reproduce dialog at a level higher that the original performance.  This may enbtail a compromise between intelligibility and strict fidelity ….. (Dreher 1931, 756)

Speaking/Signing

In theorizing cinema sound, Altman and Williams generally attend to the auditing or recording as opposed to the production of sound, yet in each there nevertheless remains a certain privileging of sound production.  By stressing the unique character of every sound event=, its essential unrepeatability, we can better understand the phenomena of sound representation and, as Levin quite rightly points out, better specify exactly what material alterations occur in a particular representation of a sound, but we are simultaneously encouraged to conceive of sound along the model of speech— as the effect of a singular act and intention, the reception of which is guided by the desire for absolute presence of the “original.” What we want, it seems, is not simply to understand or to recognize this speech but to touch what is “behind” the sounds — presumably something which unsullied by transformation into the common coin of language, is more real. Understood in this way, any recording or writing of this speech disrupts its presence.

Despite recognizing a conflict between the “naturalness” and “intelligibility” of a sound recording, technicians argued repeatedly over the form of “authenticity” appropriate to sound representations.  Such conflicts are symptoms of historically changing notions of representational “reference,” and the resultant struggles had important effects on representational form, technical practice, and even devises.  Indeed, in the debates over what form of “authenticity” representations should have, we indirectly witness the struggles of professionalization, and the emergence of a new “science.”

Altman has argued that new realism can emerge only in relation to existing models of reality and of representation (1984).  A relationship of “quotation” obtains between an existing realistic mode of representation and an emergent one.  Photography, for example, “quoted” enough of the codes of painting and other pictorial arts that, while it appeared different, it nevertheless “counted” as a realistic representation of conforming to existing codes of recognition and so on, even as it transformed them.  From this point of view, it is easy to understand how there can never be an entirely “new” form of representation — there are simply far too many constraints (in the form of codes of recognition, practices of everyday life based on representations, entrenched ways of understanding representations and so on) mitigating against its acceptance.

The relationships of a new form to existing ones allow us to specify historically how and why some new form emerged, noting how it manifests its genuinely novel attributes while still conforming to the established cultural category of “representational image.” Thus, the process becomes one of the regulated appearances of a certain form of knowledge about the technology rather than the simple emergence of a radically new, but self-defining form.

In the case of the sound film, then, what would be the relevant reality codes to which a recording (and a synchronized print) would have to codes to which a recording (and a synchronized print) would have to correspond? Although technicians repeatedly make appeals to the body of the spectator and to real perception, this is surely not the only or even the primary standard of realism.  The theater, telephony, radio, and public address are all potential models with strong relationships to film sound, but perhaps the most important codes for sound come from silent cinema.  Silent cinema had already established ways of dealing with problems of uniqueness versus intelligibility, with sound “space,” and with sound source, and had already defined the roles “speech” should fulfill in a film.

Figure 7.9 Sound Studio

Bordwell has drawn our attention to how aesthetic and technical norms impinge on all Hollywood practice, by shaping the paradigmatic substitutability of various techniques in any particular film.  Techniques which perform the same, or nearly the same function in a film are termed “functional equivalents”
(Bordwell 1985a, 4-6).  So, with Boarwell’s argument in mind, we might ask what roles were preordained for sound even prior to its possibility as a synchronized effect.  Despite contemporary claims about how the revolutionary technology was to change “everything,” sound could never have been adopted if it did not fulfill or was not capable of fulfilling some function already deemed important by Hollywood standards — the short-lived effect of novelty is never a convincing enough argument for people involved in a multimillion-dollar business. As Bell technician Harvey Fletcher pointed out in 1929, from the very start, even basic research into sound recording and reproduction was guided by some very industrially determined standards.  Although apparently researching “sound” in general, Fletcher makes it clear that even fundamental research worked under the supposition that “sound” was guided by some very industrially determined standards.  Although apparently researching “sound” in general,

Fletcher makes it clear that even fundamental research worked under the supposition that “sound” was always understood as having a very specific role to play as a carrier of information.  Thus, the reproduction of speech was implicitly and often explicitly) guided by the preordained role of speech as, above all, an intelligible conveyor of information.  Although specific social uses of sound obviously must serve as a starting point for the development of any sound technology and its technical standards, the particular sound forms and uses chosen and nevertheless extremely important.  In this case, they meshed well with Hollywood needs.

In silent films, however, the situation was different.  An intertitle is affected neither by the space of production nor by the space of reproduction.  As a form of writing, it is (almost) by definition legible. For any literate audience member, then, dialogue is 100% intelligible. However, as a form of writing, intertitles, introduce another series of dilemmas.  Primary among these is their relative separation from their sources.  The voice, as writing, is separated spatially and temporally from its putative origin—- the moving lips on the screen. This formulation echoes one of the classical ways of distinguishing writing from speech.  Aside from the loss of certainty as to source, though, several other breaks intervene.  For example, intertitle cannot easily reproduce the the “grain” of the voice, regional accents, difference in pronunciation, tone of voice, rhythms of speech, inflections, and so on— all the sorts of indicators we normally use to identity speech as a unique discursive and phenomenal event.

Figure 7.13  Folly Artist Cincinnati 1945 Sound Recording Studio

Figure 7.13  Tin Metal Sheet Making Thunder in a Folly Studio

Decades of tin-sheet thunder and coconut shell hooves, however, prove that fidelity to source is not a property of film sound, but an effect of synchronization.  A gun firing on the screen accompanied by any brief, sudden, explosive sound produces the effect of source, it doesn’t require the predication of a source of the sound, although even the “correct” sound, if out of synchronization, can destroy the very grounds of the fidelity judgment.

Figure 7.14  Mary Ann Doane Femme Fatale

Figure 7.15  Mary Ann Doane Bigger Than Life Close-up and Scale in Cinema

In each case, the characteristics which mark the utterances of unique are indicated by making them visible, or more exactly, legible.  Those aspects of a sound which mark it as authentic are never simply self-evident “attributes” of the sound.  Only their inscription within a system allowing or requiring them to become perceptible gives them a semiotic import within the system.  Thus, the supposed unique attributes of an original sound become significant and, to a certain degree, perceptible as such only through their constitution signs— precisely that which is not unique.  The potential “scandal” of writing is overcome by …. more writing.

The iterability characteristic of writing is overcome by —- more writing.  The iterability characteristic of writing, which marks writing as repeatable, as quotable or citable in an infinite number of contexts, and thus as non-self-identical, is traditionally seen as disrupting the presence or plenitude of speech. (In fact, it is the intertitle’s susceptibility to being repeated without alteration in different contexts that is specifically used to make the protagonist’s speech to his several sexual conquests an inauthentic in Don Juan,  one of the first (Vitaphone films.) As I have sketchily pointed out, however, it is writing (as legible mark) which “comes to the rescue” to ensure the uniqueness of unrepeatability of the speech event.  As in the sound film, sound “fidelity” is an effect of inscription.  Ultimately, the intertitle’s status as writing in characters of sound recording in general and may extend to the “original” events as well.

Figure V7.7 Don Jua Seduction Presented by Turner Classic Movies

The task, then, is to illustrate how these silent film techniques weigh on early film practices and then, later, on the recording process, or the inscription of traces in general.  As I noted earlier, conflicts over representing sound space were settled de facto  by the adoption of the standard of close-miking and a certain “frontality.” This “solution” ran against a trend in microphone research, however, which sought to allow more and more distant miking to simulate hearing more closely (Altman 1985b).  The type of sound recording which results is relatively “contextless” or spaceless—- that is, it is a sound whose “quality,” to repeat a technician cited earlier, “is effectively independent of the reverberant characteristics of the set” (Hunt, 482). In other words, such sound is spatio-temporally unspecific and therefore in some sense more like writing that like speech.

In the difference between the name and the signature we can read the difference between the aural object and the aural event, the sound as recognizable across contexts, as iterable, and the sound as event, as unrepeatable.  But it is the uniqueness of the original sound, its historical happening, and this signalar event’s relation to its representation (its repetition) that needs to be analyzed.  The signature, in order to function as a signature, requires the notion of a fully present intention.  The idiosyncrasies of the mark are what stamp this form of writing as historical, as happening once and happening at that time specifically.  Yet no signature could function if it were not, as Derrida has pointed out vis-a-vis Austin, something like a citation (Derrida 1982).  Your “unique” signature is only valid through its comparison with some (obviously repeatable) model—- its functions as unique only through its being verified as a copy.  Signatures, which undoubtedly “work” every day as embodiments of unique intentions, are nevertheless marked already and always be a certain absence, or in other words, the event of the unique signature in marked already be repetition.

Figure 7.16 Aural Event

The “Originality” Effect

The character of this absence is what is finally at stake.  It seems as if the non-presence, the general iterability of the sound is a negative characteristic, a barrier to uniqueness, or to fullness.  At its most extreme, it seems to deny that specific events “happen” in a historical sense.  This, however, is not the case at all.  Although the repeatability of a signature seems to indicate that all signatures are somehow lacking, it is this absence which is the positive condition of their functioning.  Even a signature which appears never to have been reproduced has the possibility of repetition inscribed in it.  In essence, for something to be original, its repetition must be recognized as conforming in some way to the original—- or the recognizable in various contexts. Ideality then would depend on repetition, would be, in a sense an effect of repetition rather than a precondition for repetition (Derrida 1989). Understood in this way, the repetition, insofar as it is recognized as a repetition, creates the original.  The non-uniqueness or non-presence of the event marks both the possibility of and limitation of ideality.

The character of this absence is what is finally at stake.  It seems as if the non-presence, the general iterability of the sound is a negative characteristic, a barrier to uniqueness, or to fullness.  At its most extreme, it seems to deny that specific events “happen” in a historical sense.  This, however, is not the case at all.  Although the repeatability of a signature seems to indicate that all signatures are somehow lacking, it is this absence which is the positive condition of their functioning.  Even a signature which appears never to have been reproduced has the possibility of repetition inscribed in it.  In essence, for something to be original, its repetition must be recognized as conforming in some way to the original—- or the recognizable in various contexts. Ideality then would depend on repetition, would be, in a sense an effect of repetition rather than a precondition for repetition (Derrida 1989). Understood in this way, the repetition, insofar as it is recognized as a repetition, creates the original.  The non-uniqueness or non-presence of the event marks both the possibility of and limitation of ideality.

When listening to a friend tell a story, the emphasis is on the comprehension of words but was we concerned with our friend’s health we might pay attention to the quality of her voice– is it strained, unusually hoarse, tired? Likewise, a recording of a symphony my very will be inadequate for appreciating the sonic complexity to its fullest, but the same recording could function perfectly well if all that was required were the identification of, say the flute’s, melody in the first movement, or the recognition of a musical “quotation” from another piece.  The point in these cases, which seem to me fully “legitimate” ways of listening, is recognition and association not musical appreciation per se. 

Figure 7.19 Spatio-Temporal Specificity

Nevertheless, the sense that there is something more primordial than the trace persists.  There is nothing particularly striking about this.  The historical happening of the sound event, its spatio-temporal specificity, always appears to escape our apprehension.  Whether we’re in an auditorium, or listening to a relatively contextless, closely-miking recording, or to one which stresses the peculiarities or the room, the event, in its fullness, seems to escape.  What is the “real” of which these are the traces?

Figure 7.20a Brain Diagram Spatio-Temporal Specificity

Perhaps it is in the non-presence of the repetition, of the wearing away of the original surfaces, the decaying remains, that we recognize the possibility of the event itself.  Out of the trace, then, is born history…. classify representations as representations, recordings as recordings is a positive condition of possibility rather than a fault.  If we recognize that all realities are constructed realities, this need not open us to the abyss of absolute indeterminacy or relativism.  These “absolute” are just as unattainable as absolute presence.  We do, however, open us to the possibility of explaining specific effects of “authenticity,” or of “immediacy” precisely as effects of a general possibility of “writing,” and we can relinquish misty assumptions of the ineffable unattainable, but for some reason all-defining original.  We need not relinquish the original, the real, or the authentic, but we must recognize that these experiences and values, too, are products of historically defined conditions,s and that their emergence, like the emergence of representations of those phenomena, follows certain rules. We may well ask, then, to what is attributed or denied these qualities or “originality” and “authenticity,” and in whose interests? Representational reference is finally, as it is in the writing of history, a question of right and law, a question of morality, and a question of social ethics.

Figure 7.20b Brain Diagram Spatio-Temporal Specificity

What remains is to develop a theory of representation which does not insist on a loss of being between original and copy, a theory which does not seek to define the preferential effects of each and every representation by comparison with some logically unattainable ideal, and a theory which avoids falling into any naive sense of equivalence between “original” and representation.  Paradoxically, it is to Stanley Cavell’s quotation about photographs that we return to find an ironic statement of this. By repeating Cavell out of context, perhaps we can create a new intention for his statement.  “In the sound recording, the original is still as present as it ever was,” which is to say, just as absent.

Figure 7.21a Sound Recording, The Microphone

The Microphone’s position is almost always comprised by the camera’s needs.  It’s important, however, that the sound source be solidly within the pickup pattern of the mic.  Generally, a boom mic will be brought in from just above the camera’s frame, as close as possible without getting in the shot.  The boom operator should practice during rehearsals.  Be attentive to mic shadows — you may need to move away from a light to avoid them. Ask the camera operator for a frame line before the shot begins so you know how far in you can go.

Figure 7.21b Sound Recording, The Microphone

The ideal placement for many mics is between one and three feet in front of the person speaking, slightly above or below the level of the mouth.  If the microphone is directly in line with the mouth, it may pick up popping sounds from the person breathing into it. A windshield helps.  If a directional mic to too close, it will bring out an unnatural bass tone quality.  This is the proximity effect, which results from the particular way low frequencies interact with directional microphones.  If a microphone is too far from the subject, background sound (ambient sound) often completes with or drowns out the speaker’s voice.  Also, undesirable acoustic qualities of the recording space, such as echo and boominess, become more noticeable.

Figure 7.22 Sound Recording, On Set

Keep in mind that sound, like light, diminishes in intensity with the square of the distance (see figure 12-6 pg 450).  Thus, moving twice as far from the sound source diminished the sound to one quarter of its previous level.  If the recording level of the sound seems low, especially with respect to louder background sounds, you must get closer to the sound source and not attempt to correct the problem by turning the level way up.

Many beginners think the recordist should try to capture all sounds in a general fashion, standing back from, say a party, a conversation, or a street scene to record all the sounds together.  The results of such recordings is usually an indistinct blur.  The recordist should instead select individual sounds and get close enough to record them clearly.  If an overall mix of sounds is desired, it may be necessary to put together several distinct tracks later.  For documentary filming in noisy conditions, you may need to get closer to the subject than you feel comfortable doing.  A few experiences with scenes ruined by bad sound will help overcome shyness.

Another approach to the noisy situation is to use a lavalier mic on the subject with a wireless transmitter to get the mic in close while allowing the recordist to stand back.

Whenever you put a mic very close to the sound source it minimizes both ambient sound and the natural reverberation of round reflecting in the recording space. In some situations, a close mic sounds artificial.  For example, if the camera is filming a distance long shot and the mic is very close, the recording will lack the proper sound perspective. Although distorted sound perspective is found regularly in movies, you may not like.  To Correct this, the recordist could move further back, but at the risk of sacrificing clarity.  Alternatively, during the sound mix, you could lower the level and maybe use equalization and a little reverb to give a sense of distance to the sound (see Chapter 16).   Similarly, any missing ambient sounds can be added later by mixing in additional tracks.  These kinds of effects are usually handled better under the controlled conditions of sound editing that they are while making a live recording.  You can always add background sounds, distance effects, andequalization during sound editing and mixing, but nothing can make a noisy, echoing, or weak recording sound pleasing and clear.

See Sound Editing and Mixing Below

Sound Editing and Mixing

The Idea of Sound Editing and Mixing

(Page (625)

Sound editing refers to the process of creating and refining the sound for a movie.  Mixing is the process of enhancing and balancing the sound.  On a large production, sound editing is generally done by specialized sound editors who are not involved in the picture editing.  A sound designer may be brought in to create unique textures or effects.  On a small production, the same people may do both picture and sound editing.

Sound is often treated as an afterthought something to be “tidied up” before a project can be finished.  But sound is tremendously important to the experience of watching a movie.  An image can be invested with a vastly different sense of mood, location, and context, depending on the sound that accompanies it.  Some of these impressions come from direct cues (the sound of birds, a nearby crowd, of a clock), while others work indirectly through the volume, rhythm, and density of the sound track.  The emotional content of a scene (and the emotions purportedly felt by characters on screen) are often conveyed as much or more by music and sound design as by any dialogue or picture. Even on a straightforward documentary or corporate video, the way the sound is handled in terms of minimizing noise and maximizing the intelligibility of voices plays a big part in the success of the project.

It is said that humans place priority on visual over aural information. Perhaps so, but it’s often the case that film or video footage that is poorly shot but has a clear and easily understood sound track seems okay, while a movie with nicely lit, nicely framed images, but a muddle, harsh, and hard-to-understand track is really irritating to audiences.  Unfortunately for the sound recordists, editors, and mixers who do the work, audiences often don’t realize when the sound track is great, but they’re very aware when there are sound problems.

The editing of dialogue, sound effects, and music often evolve organically during the picture editing phase.  While the dialogue and pictures are being edited, you might try out music or effects in some scenes.  You might experiment with audio filters or equalization and often need to do temporary mixes for test screenings.  Nonlinear editing is nondestructive,  which means you can do many things to the sound and undo them later if you don’t like the effect. (A few effects do in fact alter the captured audio files— avoid until the final mix.)

Even so, sound work doesn’t usually begin in earnest until the picture is locked and the movie’s structural decisions are all made.  The kob of sound editing begins with the sound editor screening the movie with the director of the picture editor.  If you’re doing all these jobs yourself, then watch the movie with a pad of paper,  Every scene should be examined for problems, and to determine where effects are needed and where a certain feeling or quality of sound is desired.  If music is to be used, the composer (who will create music) and/or the music supervisor (who helps find existing music) are brought in as well.  Deciding where to place music and effects is called spotting. 

The sound editor then begins the process of sorting out the sound.  The audio is divided into several different strands or tracks.  One set of tracks is used for dialogue, another for effects, another for music, and so on. Portioning out different types of sounds to separate tracks makes it easier for the mixer to adjust and balance them.  This process is called splitting tracks. Effects are obtained and other sounds are added, building up the layers of sound this is track building.

When all the tracks have been built, a sound mix is done to blend them all back together. Enhancing the way the tracks sound in the mix is something called sweetening. The mix may be done in a studio by a professional mixer, or it may be performed by the filmmaker using an editing system or other equipment.  A professional mix studio (also called a dub stage or mix stage) is designed with optimal acoustics for evaluating the audio as it will be heard by audiences.  Through more costly, studio mixes are preferable to mixes done in the editing room or other multi-use spaces that may have machine noise, poor speakers, or bad acoustics.  For theatrical and broadcast projects, it’s essential to mix in a good listening environment.  Also, professional mixers bring a wealth of experience about creating a good sound track that most filmmakers lack.

After the mix, the sound track in recombined with the picture for distribution.  Different types of distribution may require different types of mixes.

The Sound Editing Process

Like everything else in video and film, sound editing has been transformed by new technology.  Where there was one a standard path through the process, there are now many paths and types of equipment.  How you do sound editing depends on the project, how it was shot, your budget, and how you plan to distribute the finished movie.

The distinctions between the capabilities of a nonlinear editing system, a digital audio workstation (DAW), and a mix studio have broken down somewhat.  For example, some NLEs have a very good sound editing capabilities.  A DAW, which is a specialized system for sound editing and mixing, might be just another application running on the same computer as the NLE,  And a mix studio might be using the same DAW, but with better speakers in a better listening environment.

For most video projects, the production audio (the sound recorded in the field) is recorded in the video camcorder along with the picture.  For some video shoots and all film shoots the audio is recorded double-system with a separate audio recorder.  For workflows involving double-systems sound, see Chapters 14 and 15.

Here are some workflow used for sound editing and mixing.

Projects Shot on Video

BASIC NONLINEAR EDIT

The simplest, cheapest way to do wound work is to just do it on the NLE being used for the offline edit.  Audio is captured originally at high quality and can be used for the finished product.  Sound editing and mixing are done on the system.  Audio is exported as a finished sound track, either with the video or separately (to be remarried to the picture after the online edit is done).  This method is best for simple and/or low-budget projects and is usually not a good idea for complex mixes.

Figure 7.23 Non-linear Edit

LINEAR VIDEOTAPE EDIT

Offline editing is done with worktapes, followed by online edit with original camera tapes.  Simple sound mixing can be done during the online session. More complex audio sweetening is done by stripping the studio from the online edit master and laying it over to a DAW.  Effects and music are added, then the mix is done.  Audio layback is then done to put the final mixed track on the master tape.

NONLINEAR EDIT; TRACK WORK ON DAW; MIX IN STUDIO

Picture editing and basic sound work are done on a nonlinear editing system.  Audio tracks are exported to a DAW for final track work, which gives maximum control over the process.  Mixing is done in a mix studio, with good speakers and an optimal environment for judging the mix.

Other Recording Issues

Bass Filters

Many recorders, mixers, and microphones are equipped with filters that reduce the level of low-frequency sounds,  These filters are variously balled bass cut, bass roll-off, high-pass, and low-frequency attenuation (LFA). Some filters cut off bass fairly sharply, often diminishing them 12 dB per octave; thus the filter might reduce 150 Hz somewhat, 75 Hz quite a bit, and 37 Hz almost entirely.

Filtering, also called equalization, is done to minimize the low-frequency rumble caused by such things as wind, traffic, machinery, and microphone-handling noise.  The low-frequency components of these sounds.  If low-frequency sounds are very loud, the recording levels must be kept low to avoid over-recording, and this impairs the sounds that you really care about, such as voices.

Figure 7.27  Mico Pop Filer Hero

Figure 7.28  Effect of the Roll Off Filter

Figure 7.29 Mico with Screen in Studio

Figure 7.30a,b Polar Pattern Diagram

Microphones and recorders may have a two – or three-position bass roll-off switch.  The first position (sometimes labeled “music” or “M”) provides a relatively flat frequency response with no bass filtering.  The next position (“voices” or “V”) provides filtering response with no bass filtering. If there is a third position, it rolls off bass starting at an even higher frequency (see figure 10-21 pg 394 The Filmmakers Handbook). Do test recordings with the filter to judge its effect. On some mics, the “M” position is optimal for most recordings, and the “V” position should be used only for excessive rumble or when the mic is very close to someone speaking.  Sometimes the third position removes so much of the low end that recordings sound very thin and hollow.  For this reason, it’s usually best to avoid using a camcorder’s “wind” menu setting.

There are two schools of thought on filtering bass: One is to filter as needed in the original recording: the other is to hold off as much as possible until postproduction.  The first school argues that the low frequencies will be filtered out eventually, and a better recording can be made if this is done sooner rather than later.  This must be weighed against the fact that frequencies rolled off in the original recording are not replaceable later.  The sound studio has better tools and is a better environment in which to judge how much bass needs to be removed.

A prudent approach is to filter bass only when excessive rumble or wind noise requires it or when trying to compensate for a microphone that is overly sensitive to low frequencies.  Then filter only the minimum amount to improve the sound.  If filtering is done, keep it consistent in the scene.