Is Mixing an Art?

“Then there’s that beautiful moment at the height of creation

when someone will lean back in their chair and say

‘sounds like a record’, signifying that the whole

finally transcends the parts”.

Richard James Burgess

I

Introduction

The purpose of this paper is to examine the process of mixing a music record in order to determine whether it is an artistic activity or merely a technical procedure. I will begin by setting up a theoretical frame that will serve as a tool in my examination of this question. My theoretical frame is largely derived from Nelson Goodman’s autographic-allograhic distinction, Theodore Gracyk’s account of rock music as an autographic art, and Casey O’Callaghan’s sonic realism. These tools will be contrasted with several claims made by mixing engineers, and also with David Gibson’s method of expressing mixing procedures in a notational scheme as it appears in his book The Art of Mixing. I will conclude that mixing is an artistic activity because it changes the identity of records in an essential way, it is a creative activity in itself (independent of the performance and the production), and it implies, moreover, a degree of autonomy on the part of the mixing engineer. I will attempt to show that the contribution of the mixing engineer shapes the spatial and temporal content of the record’s sound, and also how this fact challenges O’Callaghan’s notion that recorded sounds entail an impoverished listening experience. Also, this contribution will raise some questions about artistic realism when applied to sound recording. However, I will claim that, as an art, mixing cannot attain full independence, because it is determined and constrained by the material produced by musicians and the artistic intention of both musicians and producers. If, despite its lack of independence, mixing can be understood as an art, I will finally attempt to show that it must be considered autographic.

II

Theoretical frame

a) Nelson Goodman

Since the invention of machines capable of automatically recording and reproducing sound, a considerable number of aesthetic theories have been presented, challenged and modified. Equally, the application of this new technology and the improvements it affords in the dissemination of music has changed the way in which we experience music and, at the same time, the way musicians create their work and are credited for it.

Before the widespread commercialization of records, the only way to experience music was by attending a live performance (or, indeed, by performing one oneself). Each performance of a musical work is unique and ephemeral as a whole and with regard also to each of its components, e.g. each of the sounds involved in, say, a performance of Beethoven’s 9th Symphony are as unique and ephemeral not only as the whole performance of the symphony but also as any other natural sound. From one performance to another several characteristics changed and others remained equal. Art theory and the acts both of musical creation and listening required discrimination between what was integral to the work of art and what was accidental, as concerned its identity and authorship.

Of course, what is accidental in the identification of a particular work can be integral to the enjoyment of a particular performance: two orchestra directors may have different interpretations of allegro ma non troppo un poco maestoso in the first movement of the 9th Symphony and the difference between them can have a dramatic impact on the experience and quality of the respective performance, but this is irrelevant when it comes to identifying that the piece is Beethoven’s symphony number nine. Moreover, although it is virtually impossible, in any case, to reproduce identically a prior performance of a musical piece, it is worth bearing in mind that there did not exist a means of comparing would-be identical performances before the advent of recording, because performances are confined to a determinate time and space, outside of which they are not available for examination. So, keeping in mind the ineluctable variability of performances, melody, harmony and relative tempo–the sound structure– are relevant elements of the identity of a piece of music, and scoring is the only reliable way to express them.

The existence and role of scores raise the question of notational systems and their characteristics. Nelson Goodman, in his Languages of Art, famously developed a theory of notation that led him to differentiate two kinds of art, one called “autographic” and the other “allographic”. Goodman addresses the differences between these two arts during his discussion of forgery:

“Let us speak of a work of art as autographic if and only if the distinction between original and forgery of it is significant; or better, if and only if the most exact duplication of it does not thereby count as genuine. If a work of art is autographic, we may also call that art autographic. Thus painting is autographic, music nonautographic, or allographic”.

One reason behind Goodman’s distinction between painting and music lies in the place that the final product of the artistic endeavour occupies in the experiencing a given work of art. A performance of the 9th Symphony, that is, the realization of Beethoven’s composition, is certainly more ephemeral than Las Meninas by Velázquez. Although Las Meninas will also disappear eventually (as do all material things) — thus painting does not completely escape ephemerality– we can experience a performance of any musical work only once. The painting-object remains over time allowing for many experiences of it, whereas every note vanishes within the duration of the symphony. This difference concerns “re-production” in the sense that for a musical work to be experienced at least twice it is necessary to have some sort of a tool that makes it possible to re-create the work without experiencing a different opus every time. It can be argued that this problem concerns paintings too because they can only occupy one place at a given time, and, hence, in order to experience them in two different places at the same time, we would need a way of reproducing them. However, in contrast to music, a painting has durable existence and remains constantly identical to itself. Therefore, a reproduction of a painting will always be another object, while upon the completion of a performance, the work itself vanishes. In addition, performances are also (at least, minimally) different from one another. However, we do not consider them different works, nor do we give credit, say, to the conductor of the NY Philharmonic Orchestra for the authorship of a piece composed by Beethoven. As Goodman says:

“Performances may vary in correctness and quality and even in “authenticity” of a more esoteric kind; but all correct performances are equally genuine instances of the work. In contrast, even the most exact copies of the Rembrandt painting are simply imitations or forgeries, not new instances, of the work”.

One reason for this difference is that what constitutes the identity of a musical work allows us to discriminate between what is accidental and what is integral in it. From the perspective of the autographic-allographic distinction, it is clear that the identity of a musical work does not depend on everything that makes up the concert experience. In contrast, in pictorial art everything contained within the frame is relevant to its identification. At the same time, for this distinction to be operative, it is necessary to have a way of communicating and receiving, without any loss or mistake, this “essence” from one performance to the next. Otherwise it would be impossible to know whether a performance corresponds to a particular work or not. For Goodman, the necessary factor is notation, and, in the case of music, scoring: “A score, whether or not ever used as a guide for a performance has as a primary function the authoritative identification of a work from performance to performance”. Scores define the integral parts of a work, rendering the differences between two or more performances of a given score secondary to the identification of the work, as occurs with the indications of tempo that I mentioned above:

“Thus the verbal language of tempos is not notational. The tempo words cannot be integral parts of a score insofar as the score serves the function of identifying a work from performance to performance. No departure from the indicated tempo disqualifies a performance as an instance –however wretched– of the work defined by the score. For these tempo specifications cannot be accounted integral parts of the defining score, but are rather auxiliary directions whose observance or nonobservance affects the quality of a performance but not the identity of the work”.

For a notational system to be complete and correct, according to Goodman, it is required that they possess certain properties, namely, “unambiguity and syntactic and semantic disjointedness and differentiation”. (fig.1)

(Fig.1)

Accordingto Goodman, a notational scheme is syntactically dens “if it provides for infinitely many characters so ordered that between each two there is a third,” and semantically dense if “the compliance-classes are so ordered that no insertion of others in normal position will destroy density”. However, although notation is a necessary condition for an art to be allographic, (“the allographic art has won its emancipation not by proclamation but by notation”.), it is not a sufficient condition. In a sense, everything could be transposed into a notational system (Goodman calls this possibility “trivial”), but this possibility does not mean that every art is susceptible to being allographic art. In order to introduce another necessary condition for an autographic art to become allographic, Goodman poses the question of whether or not painting could become an allographic art by instituting a notational system, and his answer is as follows:

“Where a pertinent classification is lacking or flouted, a notational language effects only an arbitrary, nominal definition of ‘work’, as if it were a word newly coined. With no prototype, or no recognition of one, there are no material grounds for choosing one systematization rather than any other. […] Thus the answer to the significant question about a notational system for painting is no. […] In sum, an established art becomes allographic only when the classification of objects or events into works is legitimately projected from an antecedent classification and is fully defined, independently of history of production, in terms of a notational system. Both authority and means are required; a suitable antecedent classification provides the one, a suitable notational system the other. Without the means, the authority is unexercised; without the authority, the means are footless”.

What is interesting about this second condition is that, although the notational system must be defined independently of the history of production, the requirement of an “antecedent classification” actually introduces history of production into the equation. Goodman is arguing, it seems, that the notational system must exist prior to the production of the work of art, in order for the work, or its “parts”, to fall into the classification. In this sense, the history of production and implementation of the notational system is indeed relevant.

In sum, allographic arts must be reproducible, that is, they must have a notational system that, on the one hand, distinguishes between its integral and accidental parts (notating only the former); and, on the other hand, it must be unambiguous, finite, semantically and syntactically differentiated, and it must be sufficient to identify every reproduction of the work as an instance of the same work of art. Music — as it was understood before audio recording — is a paradigmatic example of allographic art.

b) Gracyk

In his book, Rhythm and Noise: An Aesthetics of Rock, Theodore Gracyk argues that the emergence of Rock music changed the theoretical frame through which we understand musical art. Taking up Goodman’s notions of autographic and allographic art, he defends the claim that Rock music is an autographic art. Let us now turn to the reasons behind this claim.

Although the musical work, in the Western musical tradition, has generally been identified with its sound-structure (i.e., what appears in the scores), in Rock music, the sounds themselves, as contained in records, identify the work. Records and their history of production are both relevant elements of Rock. The process of recording in a studio fixes a definite set of properties of the work of which none is accidental: everything matters. Similar to etching, only those records copied from the original source, i.e. the final product of the studio process, are original instances of the work of art; whereas any other product with the same sound-structure, the same lyrics or arrangements, even performed by the same artists, is not.

“The Born to Run album is a musical work, but it is autographic, because notational determination is entirely irrelevant to the genuineness of its instantiations. […] If Springsteen or anyone else rerecords the songs on Born to Run, notational fidelity [to the song’s scores] may occur (we may genuinely have the same eight songs in the same order), and it may resemble Born to Run as closely as two performances might. But it won’t be Born to Run, the work that got Springsteen onto covers of Time and Newsweek in 1975. Notational accuracy is insufficient for access to the relevant piece of rock history”.

A problem arises: strictly speaking, from Goodman’s perspective, the song Born to Run is an allographic work, but what makes it what it is in the record includes many more aspects than the score reflects. And these other aspects that go beyond what Goodman held relevant for the identity of the musical work are rooted in the history of production of the “sound of the record”. The decisions made during the production of the record determine and fix what properties and meanings “adhere to those sound structures”. Garcyck adopts a particular interpretation of Goodman’s distinction in order to include the particular case of Rock music and records:

“Adopting Levinson’s alternative analysis of the distinction, that a work is allographic if a historically indicated structure is presented and autographic if notational determination plays no role in its genuineness, we again find that our fake Born to Run [a hypothesis of an exact reproduction of the content of Born to Run] offers genuine instances of allographic works (the songs) while forging an autographic work. So it is false that the only relevant musical works are allographic. Furthermore, musical works are not restricted to sound structures plus performance-means; that is, rock musicians compose works whose replication demands far more than structural accuracy plus various limitations on performance means”.

The change that Rock brought to musicology is rooted in the invention of the technology of sound recording. It is not the only reason, according to Garcyck, but without audio recording this change would have never happened. Originally, audio recording had the purpose of transparency, that is, a perfect transmission and reproduction of a sound as if the sound was present. In his book, Perfecting Sound Forever, Greg Milner explains how at the beginning of the commercialization of audio-reproducers (phonographs and so on) their value was mirroring. The more a phonograph is capable of deceiving the audience, making them believe that the sound coming out of the machine is real, the more it is considered good; as can be seen in the anedocte about the first public presentations of the Edison’s phonograph reported by Milner:

“Tonight’s concert was billed as a “tone test”.  […]The invitation had mentioned something about […] some sort of comparison between the live voice and its reproduced facsimile. Fuller (the head of the phonograph division at Edison’s company) told the audience that Edison’s machine could “hear” as sensitively as the human ear, and could therefore reproduce a sound that was indistinguishable from the original. […] Fuller brought Miller onstage and cued up a recoding of her singing “O Rest in the Lord” an aria from Mendelssohn’s Elijah. The record began, and Miller let it play for a while. She began singing along with it, and then stopped. There were audible gasps from the audience. It was uncanny how closely Miller’s recorded voice mirrored the sounds coming from her mouth onstage. The record continued playing, with Miller onstage dipping in and out of it like a DJ. The audience cheered every time she stopped moving her lips and let the record sing for her” […] “The audience craned forward to see when her lips stopped moving. It was the only ay they could tell when she wasn’t singing. (…) The stage lights went down. The audience was now staring into darkness as the music continued. Only hey ears could guide them, and their ears failed them. The lights came up, and there was Miller, her mouth frozen in  a smile. When had she stopped singing? The crows went wild.”.

However, very soon those who worked in the business of recording sound began to modify the way in which the music was performed and recorded in order to create new sounds, or new characteristics of old sounds, as can bee seen in another anecdote about Stokowski also reported my Milner:

“On october 6, 1929, The Philadephia Orchestra traveled to New York to play the first radio program broadcast over a network. […] He decided he wanted to do more than merely conduct these broadcasts. When he discovered that the show’s engineer controlled the sound level and mix as it went over the air, the maestro announced, “No one controls Stokowski’s sound but Stokowski”. He insisted that NBC rig up a portable mixing board that could be placced next to him while he conducted. If Stokowski thought  the orchestra sounded too loud or too soft, he’d reach over and adjust the levels. […] At a time when music at a session was recorded directly onto a disc-cutting machine, before postproduction sound mixing was a realistic possibility, Stokowski learned to “mix” musicians. […] As with radio, however, Stokowski’s insistence on complete control had its drawbacks. Not only did he require that the recording engineers consult him before moving any microphones, but he insisted that a panel that controlled the microphones be installed next to him at the podium so he could tweak the sound as he conducted. […] Besides the technical knowledge he was gaining, Stokowski was begining to conceptualize what it meant symbolically to record music electrically”.

If these changes had been just a matter of novelty, then we might merely have to address the emergence of novel instruments, like synthesizers, for example, or the introduction of the Broadwood piano, the first with a range of more than five octaves during the era of Romanticism, which led Beethoveen to use “new” notes in his late compositions. What Gracyck argues, however, is that studio work not only creates new sounds, but also changes the meaning of performances and scores by making it possible for some part of the musical art hitherto considered irrelevant for their identification, such as timbre, to become an integral part of the work.

This is why Gracyck also distinguishes between the live performance of a musical work and the performance that the artist records in the studio. Strictly speaking, a guitar riff recorded in a session is not a performance of a pre-given work. That particular guitar riff is part of a bigger work -like a brushstroke in painting- that is relevant as a whole, with all its properties (including –albeit not exclusively– its notable part). In this sense, not only those parts of the work that were considered accidental become integral parts of the recorded work, but, more important, once included in the final mix, accidental sounds — i.e., a musician coughing- that were not part of the work at all in both scores and live performances become deliberate, and must be considered an integral part of the musical work: a consequence of the artist’s intention, as is any other sound contained in the record. Furthermore, the fixed character of the record also allows for a distinction between finished and unfinished works:

“…it calls attention to the distinction between an artist’s musical activity and something more specific, namely the works that the artist sanctions as items for appreciation and critical evaluation. Musical activity includes everything from practicing scales to idle strumming of a guitar or trying out different chord sequences. Even if recorded, these snatches of musical activity are not performances in any usual sense; they are not put forth for appreciation. We distinguish finished from unfinished works.”

However, the most important difference between a live performance and a record is the capability of manipulation that musicians have in the studio, and the fact that such a capability is employed. Technology allows for a multiplicity of sound manipulation, even during live performances, but the kind of manipulation that is possible in recording is precisely the liberation of the process from the constraints of real time that apply to live performances. Records transcend the constraints of time and space, that is, the fact that a live performance must exist “here and now,” that every sound must be performed at a particular moment, with a particular duration, from a particular spot. On the contrary, among all the sounds included on a studio record, some are recorded on a certain day, and others on other days. However, once mixed and played they appear to belong to a single timeframe. Equally, some sound effects –from primitive techniques, such as the recording of vocals in a bathroom in order to obtain a particular echo, to that of a guitar recorded in an insulated room in order to obtain a flat sound, to digital technologies– also break the spatial constraints imposed on live performances. Once mixed and played, these sounds recorded in different places with different environmental characteristics produce a whole that is impossible to recreate in a single place. Records represent not only the fiction of a performance that never actually existed, but also the fiction of a performance that could, in fact, never exist.

In this sense, the medium is not “invisible”. It adds properties to the work. If the recording process were invisible, that is, in this case, inaudible, records would be secondary instances of the work, and the work would just be the studio performance; but they are not. If the medium were invisible, then, their specificity would lie squarely in the content they transmit. In Rock music, records qua media are “audible” and, moreover, an integral part of the work of art. They are not conceived and created as instances of another work, but as “primary texts”.

“Even if its subject and materials are ‘real things’ appropriated from ‘real life’, new properties are always added. In a truly invisible medium, never allowing properties over and above those present in the represented subject, the results would not be independent works or texts. It would be reduced to its content. Understanding artistic activity involves understanding the interpretative contribution of the medium to our grasp of the content”.

However, when we listen to a live performance of a pre-existing record, we can identify the songs as being the same as those on the record. Recorded songs can be reproduced by allographic means in a live performance, and they can be identified just like any other musical work, but they are only ever recorded in a way that renders the former distinction between “integral” parts and “accidental” parts irrelevant. Gracyck explains this fact by saying that Rock exemplifies two different sorts of musical works: autographic recording and allographic songs:

“I am not suggesting that rock repudiates prevailing concepts of the musical work. Rock’s use of recording as a medium does not require a wholesale rethinking of music. […] Instead, rock adapts existing norms so that when recording is the medium, recordings simultaneously exemplify two different sorts of musical works: the autographic recording and the allographic song”.

Thus, the record is not an instance of another work, the song, as were performances before Rock. The record does not transmit a given performance; but it represents a performance. This process of manipulation, this elaboration of a total sound is what configures the work’s identity. This totality is the reason behind Gracyck’s claim that Rock music is an autographic art. The work’s identity evolves during all the phases of the album-recording process, but once released it remains fixed and there is no way to create an instance of that work, only variations and representations of it are possible. The question of the identity of the work is central to Gracyck’s claim that Rock is autographic:

“The identity of the originating sounds differs from the identity of the musical work featured on the finished recording. […] The music’s identity was not fixed until the final stages in the record-making process. Thus the performances may be “of” that work rather than another only retrospectively. […] The music performed to generate the basic tracks had an ambiguous identity at the time of its recording. […] The identity of the music on the finished recordings cannot be reduced to the identity of the music when it was being performed; performances may cause all of the sounds heard upon playing the recordings without settling the identity of the musical work. Songs are present, but not in virtue of the causal relationship between recording and performance posited in the realist “recording” relationship”.

Among the things that constitute the final identity of a record there are at least two main sources: the sounds of the songs themselves and the “sound of the record”. As we have seen, songs can be understood as an allographic part of the whole record, since they can be notated, performed and recognized as the work itself. The sound of the record, however, is as unique as a painting is, and configures the autographic properties of rock, as Gracyck understands it.

III

Mixing and Casey O’Callaghan’s theory of sounds

Usually there are several individuals who contribute to the different “parts” of the record. I am not referring here to the luthiers, for example, who would be equivalent in the realm of painting to the company that makes the paint or the brushes, but rather, to be more precise, to the people that intervene in the production of the record, meaning its passage from an ambiguous identity to its final autographic product. There is no completely pre-determined pattern in the making of a record. Sometimes songs are composed prior to entering the studio, while other times songs are created in the studio, either as partial derivations of older songs or not. Sometimes, the same individuals do both the composing and the performing, different people perform these functions. Sometimes, well-funded bands can hire professional musicians to perform certain parts of the records; sometimes their own particular style in performing is essential to their sound. Sometimes the sound engineer is also the producer, and other times these two functions are separated.

Among all of these functions in the recording process, certain steps are necessary for the “sound of the record” to exist. These steps include, at the very least, composing, performing, producing and mixing (including mastering). The order in which the first three steps appear can vary depending on the recording process, but the final step is always mixing (this does not exclude the fact that mixing can take at early stages as well). I want to focus on this last step and attempt to provide an answer to two questions: first, is mixing an art? That is, should mixing be considered a creative contribution to the record, or rather a merely technical one? And second, is mixing an allographic part of the record -like songs- or is it autographic, like producing?

The work of a sound engineer in the process of recording an album includes many things. The sound engineer sets up the microphones and the jacks in a specific place; she decides how many channels to use to record a given instrument (i.e. typically, due to developments in technology, engineers place many microphones near the drums, one or even more for each drum or plate); she introduces effects in pitch, timbre, frequency, tempo, harmony, reverberation and so on. Some of these effects affect the quality or the identity of the sound itself, and others imply changes in the space where the sound is. For example, an echo added to the chorus adds a characteristic of the environment, rather than of the singer, while a change in harmonics to make a note sound tuned with the rest of the song is a modification of the sound itself. Finally, a sound engineer mixes all recorded material to create one sound, a whole from the parts. In this mixing moment, the engineer decides the grade of distinctness of each sound, the relationship and the hierarchy between them. Sound engineer Roey Izhaki, in his book Mixing audio. Concepts, practices and tools, defines mixing as follows:

“The following definition can be given: a process in which multitrack material –whether recorded, sampled or synthesized–is balanced, treated and combined into a multichannel format, most commonly two-channel stereo. But a less technical definition –one that does justice to music–is that a mix is a sonic presentation of emotions, creative ideas and performance” (bold in the original).

If a “presentation of emotions, creative ideas and performance” is one of the many necessary conditions for art, then, according to Izhaki and the majority of mixing textbooks written by sound engineers, mixing must be considered an art itself because it uses particular techniques to express emotions, ideas and performances. Moreover, if, as Duchamp said, claiming that one’s work is art is sufficient to consider any given work as a work of art, then mixing is doubtless an art if mixers claim it is. Against the technician’s account of sound engineers as mere assistants to the artist by virtue of their skills in making music sound clearer, Izhaki states: “intelligibility is the most elementary requirement of sonic quality, but it goes far beyond that”.

But, it seems, a sound engineer follows orders. There is a musician that wants a particular sound, a producer that decides how to create that sound on record, both give orders to the engineer and he executes them without adding or creating anything. The fact credit is almost never given to the engineer on an album’s cover suggests that, within the industry and among the audience, engineers are considered mere executors instead of creators of the record’s sound.

If this account were true, then as long as a sound engineer has the necessary technological skills to produce the sound desired by the musicians and the producers, it would not matter who the engineer was. This claim is difficult to contradict because the same song is almost never mixed by different engineers, and we would need two samples of the same song, taken from the same tapes, mixed by different individuals with the same technology in order to compare them. The two different mixings of Nirvana’s Nevermind, as reported by Izhaki, provided us with an example:

“One should listen to the four versions of Smells like teen Spirit mentioned below. The link between the sonic quality of a sound recording and its ability to excite us makes it fair to assume that the listed order would also make the appealing listening order –having the rehearsal demo as the least appealing listening, and the album version as the most appealing one. As per our recent discussion, it should be clear why most people find both the rehearsal demo and the live recording less satisfactory listening when compared to the studio versions. But comparing Vig’s and Wallace’s mixes gives great insight into what mixing is truly about, and what a huge difference a mix can make.

(fig2)

Both Vig and Wallace used the same raw tracks; yet, their mixes are distinctly different. Vig’s mix suffers from an unbalanced frequency spectrum that involves some masking and the absence of spark; a few mixing elements, like the snare reverb, are highly discernible. Wallace’s mix is polished and balanced; it exhibits high definition and perfect separation between instruments; the ambiance is present, but like many others mixing elements it is fairly transparent. Perhaps the most important difference between the two mixes is that Vig’s mix sounds more natural (more like a live performance), while Wallace’s mix sounds more artificial. It is not equipment, time spent or magic tricks that made these two mixes so dissimilar –it is simply the different sonic vision of Vig and Wallace. […] Straight after recording Nevermind, it was Vig that started mixing the album. Tight schedule and some artistic disagreements he had with Cobain left everyone feeling (including Vig) that it would be wise to bring fresh ears to mix the album. From the bottom of prospective engineers list, Cobain chose Wallace, much for his Slayer mixing credits. Despite the fact that Nirvana approved the mixes, following Nevermind’s extraordinary success, Cobain complained that the overall sound of Nevermind was too slick –perhaps suggesting that Wallace’s mixes were too listener-friendly for his artistic, somewhat punk-driven, taste”.

Several things are relevant in this anecdote. First, the difference between the first three and the last recording recalls Gracyck’s concept of fixed identity. The rehearsal version is the unfinished work, thus its identity is still ambiguous. The live performance version plays the role of documentation of a historical fact, namely, a performance, and the mixer does not manipulate it: its goal is fidelity and a kind of transparency. The difference between the third and the fourth version lies in Cobain’s intention. Thus, strictly speaking, only Wallace’s mix counts as the original work. The fact that Cobain decided to change Vig for Wallace shows that, actually, the artistic intention belongs to him, because he makes the relevant decision to fix the final identity of the album. However, both his “artistic disagreements” with Vig’s mixing, and his complaints about Wallace’s final mixing also show that the engineer maintains a certain degree of autonomy that allows him to change the final identity of the work beyond the intention of the musician.

The autonomy of the mixer raises a central question: what exactly is changed in the mixing? If we take a look at the history of mixing, beyond the obvious fact that it is determined by the technological development in recording and mixing, it is easy to see that the change that made mixing relevant is related to the possibility of breaking with the notion that the mixer’s job is to record the sound exactly as produced by the musician. Bobby Owsinsky explains the change in these terms:

“In the early days of recording in the 1950s, there really wasn’t mixing per ser because the recording medium was mono and a big date used only four microphones. Of course, over the years, recording developed from capturing an unaltered musical event to one that was artificially created through overdubs, thanks to the innovation of Selsync (the ability to play back off of the record head so that everything stayed in sync), introduced in 1955. The availability of more and more tracks from a tape machine begot larger and larger consoles, which begot computer automation and recall just to manage the larger consoles fed by more tracks. With all that came not only an inevitable change in the philosophy of mixing, but also a change in the way that a mixer listened and thought”.

These technological possibilities and the fact that they were used in a particular way and with particular goals led to the same debate that every change in artistic technologies and techniques produces: the question of fidelity, which is a form of the philosophical problem of realism. The fact that these debates are reproduced every time, and that the terms of the debate always express a tension between standard practices and new ones supports Goodman’s claim that realism is a matter of habit. Virgil Moorefield reports this same tension in the case of mixing through the reservations voiced by two of the most influential producers of the XX century when mixing opened the door to creating some new kinds of fictions in sound recording:

“Like [John] Hammond, [Mitch] Miller was ambivalent at best about multitrack recording, faulting the modern processes of overdubbing and punching in as techniques which rob music of its spontaneity and vitality. For both producers, and probably for their generations as a whole, modern recording processes such as editing, splicing, overdubbing, and remixing were a form of dishonesty. In this worldview, a valid musician is a virtuoso, and the ability to perform in real time is paramount”.

However, there was a real change with the introduction of these new practices in the manner that recorded sound was perceived. I would like to claim that the work of the mixing engineer changes the ontological character of sounds. In order to make such a claim I need to rely on a realistic conception of sound. Sounds must be something, a kind of material with determinate properties, which, once changed, transform their identity. I will follow Casey O’Callaghan’s theory of sound to outline such a conception.

Against “visuocentrism”, a traditional conception of the objects of perception that gives preeminence to vision, and considers sounds as mere properties of objects -like colors- that are ultimately a product of our brain as it decodes those properties, O’Callaghan offers a very different account that suggests sounds are things in themselves:

“Suppose sounds are not merely mental artifacts of sensation. Suppose that a sound can seem distinct from oneself, and that subjects sometimes really do hear sounds. Sounds, that is, frequently are the targets of perspective. Realism bout sounds —sonic realism– is the view that the world contains sounds whose existence is not entirely dependent upon the auditory experiences of subject. Realism about sounds as I shall develop it maintains, furthermore, that the world of sound is not a distant realm occupied by alien entities or attributes. Sounds are in the world. Sounds are entities that, in the first instance, we auditorily perceive”.

If sounds exist as such, then what kind of things are they?

“Sounds, I propose, are events. My proposal aims to capture not only the sense in which sounds seem located in one’s environment, but also the sense in which sounds are creatures of time. Much as visible colors are bound up with the extended spaces of volumes and surfaces, sounds are inextricably connected with time and have durations”.

Thus sounds are particulars that happen in a place and during a time. Their space and time are their circumstances, and their identity as events is formed by their properties. We can perceive –that is, hear– at least some of them. And these audible properties let us recognize their identity and distinguish them from their sources:

“…sounds appear auditorily as distinct particulars that bear similarity and difference relations to each other based on their complexes of audible qualities –the properties of pitch, timbre, and loudness -to which their identities are tied. Sounds, I want to suggest, have identity, individuation, and persistence conditions that require us to distinguish them from properties of the sources that we should understand to make or produce sounds”.

The fact that we can perceive these particulars, the fact that these particulars happen in a given place and time and the fact that we identify them through their properties imply that sounds provide information about the space where they exist as well as their temporal duration:

“Audition furnishes 360º awareness of a three-dimensional spatial field and directs the orientation of visual attention toward the sources of perceived sounds. Awareness of sounds provides information about the sorts of objects that populate the environment, about what those objects are doing and about how they interact”.

Records are combinations of sounds. Thus, according to O’Callaghan view, they are events, creatures of time that have a location and provide information about the environment and the passing of time. However, recorded sounds are different from natural sounds. The most important difference, I believe, is the ambiguity of their source. In some sense, the source of the sound of the album Vaivén by Jorge Drexler that I am listening to while writing this paper is the speaker of my computer. As O’Callaghan says, the sounds I hear from my computer furnish 360º awareness of a three-dimensional spatial field. However, in another sense, I hear the guitar and the bass and the voice of Jorge Drexler and I know that those are the sources of these sounds. Moreover, I hear the voice as being closer to me than the guitar. And the bass, I hear it as somehow underlying the voice. In my head, especially when wearing headphones, I can see where these instruments are placed in an imaginary space: Jorge Drexler is in front of me, to his left the guitar player, to his right, a little further behind him, there is the bass player and behind all of them there is the drummer with his jazz drums. Furthermore, as if they were very distant, I hear a sort of choir, but if I draw my attention to it I recognize Drexler’s voice singing in multiple layers. Also, the guitar has some reverb, as if it were played inside a small bathroom, but the bass sounds dry, as if it were played in a room full of pillows.

Not only can I tell the difference between the speakers and the instruments as sources of the sound, I can also reach two simultaneous forms of awareness in time and space. Changes in timbre, pitch and volume change my image of the band, but not the awareness of my computer’s speakers as a source. Moreover, just as I know perfectly well that it is impossible for Drexler to sing two different melodies at the same time, it is also impossible that the sounds occuring simultaneously on his album – one, for example, sounding like an effect produced in a bathroom, and the other the effect of being produced in a padded cell – have their sources in the same space.

As we have seen above, the unnatural properties of records are responsible for their autographic character. The question remains, however, how much of this fiction is the mixing engineer responsible for?

IV

The Art of Mixing

In his book, The Art of Mixing, mixing engineer David Gibson enumerates 11 aspects of a recorded piece of music: concept or theme, melody, rhythm, harmony, lyrics, density, instrumentation, song structure, performance, quality of the equipment and the recording, and the mix. Since his book is a manual for mixing engineers, he focuses on teaching how to mix. His first claim is that what mixing does is place the sounds recorded in the studio in an imaginary space:

“[One] way we perceive sound is by imagining sounds between the speakers. The apparent placement of sounds between the speakers is called “imaging” because it is a figment of our imagination. So you see, we’re not talking about reality here. When we imagine a sound, like a vocal, to be between the speakers, there is, in actuality, no sound there. The same sound is coming out of both speakers, traveling throughout the room, and we just imagine the sound to be between the speakers. The same thing happens when you listen through headphones:

(fig3)

[…] A wide range of dynamics are created by different placements of sounds between the speakers, and these dynamics are utilized to create all the various styles of mixes that fit all types of music and songs.

Gibson has invented a way of representing this space between speakers that (in a stereo record) we imagine. Drawing a three-dimensional space between two speakers, the mixing engineer places each sound in a relative position to the limits of the space and to the other sounds. Panning is represented as a place left to right; volume as front to back; and pitch as up and down (fig4)

Fig4

(fig4)

Each instrument recorded separately is represented as a globe, sometimes as a perfect sphere and other times as different round geometrical shapes. Their size expresses absolute volume and its shape, a frequency range. The distance between spheres expresses their distinctness or their overlapping. Many characteristics of the mix are represented by the different relationships between their shape and the space they occupy on this imaginary stage between speakers. An example of a section of a recorded song would be fig5.

What this method shows is that the mixing engineer actually creates the fiction of a space and time that goes beyond the intention of the artist, which has a language of its own, and which is an integral part of the final identity of the record.

Moreover, acknowledging this implies a challenge to O’Callaghan’s conception of recorded sounds. According to O’Callaghan, recorded sounds provide an impoverished and illusory experience compared with the original or a non-recorded sound: “According to my ‘illusory perspectives’ account, hearing the recording in your living room is a way of perceiving the sound of Hendrix performing ‘The Star-Spangled Banner’. Your experience of that sound, however, is impoverished and illusory in a number of respects, including its spatial and temporal perspectival contents”.

Apart from the fact that O’Callaghan does not distinguish between recording a live performance and a recording an album in the studio, there is an important difference between “illusory” and “impoverished”. What the mixing engineer does with the tapes he works on is an illusion, but what he does in creating this illusion is manipulate the audible properties that configure the identity of the sound in order to produce an effect on the listener, so that the latter may imagine a space and a time equivalent to the awareness of any other type of non-recorded sound. In order to see if this illusion implies an impoverished experience or an experience as rich as the one that non-recorded sounds produce we should ask whether or not this illusion completely deceives the listener. Here again, we are facing the classical question of realism. To start with, the spatial and temporal content of the record, as we have seen, is a fiction. It is a representation of a performance that never existed as such. This case introduces a new perspective in the dispute about mind-dependent and mind-independent accounts of artistic realism.

O’Callaghan claims that recorded or transmitted sounds do not need a decoding to be understood, and in this sense, recorded sounds would be transparent. Therefore, the experience of listening to a record would not be an impoverished experience. Although it would only be an “illusory” experience, it would nevertheless have full spatial and temporal content. On the other hand, as I mentioned earlier, one can perfectly tell the difference between the speakers and the instruments as sources. But, is it due to the other senses, namely vision, that one can tell the difference? Is the fact that I am do not see Jorge Drexler but rather only my computer, the give-away that I am listening to a record? Or is there something in the sound of a record that tells me that I am not listening to a live performance? Are audiences fooled by playback?

Fig5

fig5: Visual O. The Alarm Clock Section in “Time” on Dark Side of the Moon by Pink Floyd

Fig6

Fig 6:Visual H. Acoustic Jazz Mix. Gibson: “Note the incredible cleanliness and clarity of the overall mix. The bass is panned to the right and doesn’t have much high end. The guitar is right out front with the piano and the hi-hat. The kick is quite loud, which is not typical”.

If we answered this last question in the affirmative, then the illusion created by the mixing engineer would be complete. But, following Goodman, if one could be fooled by a playback performance, would this deception be caused by rock bands’ habit of using records as primary texts and live performances as representations of those records; or would it be the use of amplified electric instruments in live performances, as Gracyk points out, that sound like the sounds on the record?

Be that as it may, it is clear that the contribution of a mixing engineer in the production of a record is essential to the identity of the record; and it is also a creative and autonomous activity. A record is a collective product, and there is a director –sometimes the musician, sometimes the producer, sometimes the marketing director of a recording company, etc– but the role played by the mixing engineer is an artistic one.

Finally, we need to determine if the art of mixing is an autographic or an allographic part of the broader autographic rock music recording art. As we have seen, in order for us to consider art allographic the difference between original and forgery must be significant, the history of production must be irrelevant, and the art must have a notational system as an authoritative precedent. Even if Gibson’s method could be considered an authoritative precedent, it is clear that his notational scheme is not a notational system, because it is not semantically nor syntactically differentiated and finite, nor is it unambiguous (some positions in the stage may signify both panning and pitch, depending on the way that perspective is understood in the three-dimensional stage); and it is also fatally redundant (volume, for example, is represented in two different ways, that may contradict each other). Between each pair of successive marks and between each pair of the sounds denoted by these marks there can always be a third, more similar to each member of the pair than each member is to the other. The question of forgery is rather difficult because, since every mixing engineer works with different material and with different goals, the best one can do is imitate a style (Fig6). Is it possible for a mixing engineer to take old tapes of an old record and copy the exact mixing that shaped that old record? Would Gibson’s system be precise enough to copy a mix? No. And finally, the history of production of the mix is as relevant as the history of production of the rest of the record. Therefore, if it is assumed that mixing is an art, then we must say, for now, that it is an autographic art.