Read an Excerpt
Signal Traffic
Critical Studies of Media Infrastructures
By Lisa Parks, Nicole Starosielski UNIVERSITY OF ILLINOIS PRESS
Copyright © 2015 Board of Trustees of the University of Illinois
All rights reserved.
ISBN: 978-0-252-09741-6
CHAPTER 1
Compression
A Loose History
JONATHAN STERNE
The use of the word compression to describe a communication technology process comes rather late in its history. According to the Oxford English Dictionary, the term compression is at least six hundred years old. Its use to describe the "condensation of thought and language" dates to the eighteenth century. The term was first applied to machinery—steam engines—in the mid-nineteenth century. Compression as a description of representation thus predates its use to describe a technical operation by about one hundred years.
Today, compression in communication engineering refers to one of two things: data compression or dynamic range compression. People encounter data compression every day in the form of zipped files, mp3s, jpegs, online videos, and mobile-phone voice algorithms. All of these technologies save precious bandwidth by eliminating categories of data that engineers have decided are redundant and therefore unnecessary to store or transmit. Dynamic range compression refers to reducing the distance between the loudest and quietest parts of an audio signal. It is useful because a signal with less variance can have a higher overall average volume.
Most writers outside the engineering world, and especially most humanities scholars, still understand compression as something that happens after the fact, as supplemental to communication and its purposes, to perception, to interaction, and to the experiences attending them. In the wake of poststructuralism, few humanities writers would argue for verisimilitude as a guiding norm for representation—whether technological or otherwise. Yet too often we still write our media theories and histories as if the primary aesthetic criteria for technical media are verisimilar. In such work media representations are judged not only in terms of their realism but also in terms of their self-sufficiency; perceptual and definitional abundance; and immersive characteristics. This is the story of communication as being about the anxiety over the loss of meaning through a succession of technical forms. The assumption here is that progress in technology comes through its ability to produce verisimilitude.
For example, we can hear this set of assumptions as the warrant behind an implicit criticism of the sound of contemporary music in a New York Times story about vinyl records:
The last decade has brought an explosion in dazzling technological advances—including enhancements in surround sound, high definition television and 3-D—that have transformed the fan's experience. There are improvements in the quality of media everywhere—except in music. In many ways, the quality of what people hear—how well the playback reflects the original sound—has taken a step back. To many expert ears, compressed music files produce a crackly, tinnier and thinner sound than music on CDs and certainly on vinyl. And to compete with other songs, tracks are engineered to be much louder as well.
But it is not only journalists who argue this way, and we should be grateful, since academics are more likely to render their assumptions explicitly. Writing about optical devices like the telescope and microscope in the nineteenth century, Anne Friedberg argued that their "entertainment function ... relied not only on the verisimilitude of the images seen and the recording capabilities of mediated vision, but also on the illusion of verisimilitude, the very virtuality of the experience produced." The comment is interesting both because Friedberg explains her own logic and because if presented with the proposition in the abstract—verisimilitude is the basis of virtuality and entertainment—as an author influenced by poststructuralist thought, she would not, I suspect, accept the proposition.
In writing about technical media, verisimilitude is often tied to signal definition. Definition is the amount of signal that can fit in a given transmission or be stored in a file. It is the available bandwidth or storage capacity of a medium in terms of how much of its content can be presented to an end user at any given moment. It measures the density of materials available to perception; available is the key term here, because signal definition guarantees neither robustness of perception nor intensity of experience for listeners or viewers. The number of pixels on your screen is a measure of definition; high-definition television is measured in pixel density. The number of bits-per-second transmitted as your digital audio file plays back is also a measure of definition. It is a traditional line of marketing rhetoric to assert that increased signal definition leads to increased realism in the representations of a given medium or format, and that the increased realism will lead to greater intensity of experience and deeper meanings for audiences. But this is not actually the case. As Michel Chion writes,
Current practice dictates that a sound recording should have more treble than would be heard in the real situation (for example when it's the voice of a person at some distance with back turned). No one complains of nonfidelity from too much definition! This proves that it's definition that counts for sound, and its hyperreal effect, which has little to do with the experience of direct audition.
Analogous arguments can be made for images and video: definition is not verisimilitude, definition is not realism, and realism is not reality, but these terms are still often confused.
There is an aesthetic tradition that goes in the opposite direction. If we want to understand compression as a cultural phenomenon, as something other than a perversion or diminishment of more primary, higher-definition sensory experience, it would be wise to begin with how end users experience it. Aesthetics matter here for several reasons: so much of the writing that orbits around verisimilitude makes aesthetic arguments, and so if I want to pose a viable alternative, I need to at least gesture in that direction. We also tend to think of the storage and transmission dimensions of media as anaesthetic phenomena, mere engineering matters—more telecommunications than communication. But in fact they are absolutely central to culture, experience, and action at a distance. They help shape the texture of mediatic experience.
Consider this account of telegraphic conversation from a mid-nineteenth-century piece of short fiction, "Kate: An Electro-Mechanical Romance":
Mary replied instantly, and at once the two girl friends were in close conversation with one hundred miles of land and water between them. The conversation was by sound in a series of long and short notes—nervous and staccato for the bright one in the little station; smooth, legato and placid for the city girl....
[T]he two friends, one in her deserted and lonely station in the far country, and the other in the fifth story of a city block, held close converse ... for an hour or more, and then they bid each other good night, and the wires were at rest for a time.
Here, the basis of intensity and intimacy is precisely a lack of definition. Whole modes of being are condensed into the rhythms of telegraph signals, which in turn index the subtle and quick movements of operators' hands. We could attribute this description to a standard nineteenth-century literary conceit were it not so common elsewhere. In When Old Technologies Were New Carolyn Marvin tells stories of telegraphic and telephonic weddings and deceptions, feats of long-distance intimacy and intensity, whether in shared passion or cruelty.
Frantz Fanon's treatment of radio in A Dying Colonialism follows a similar pattern. In his chapter, "This Is the Voice of Algeria," he writes, "The whole nation would snatch fragments of sentences in the course of a broadcast and attach to them decisive meaning. Imperfectly heard, obscured by an incessant jamming, forced to change wave lengths two or three times in the course of a broadcast, the Voice of Fighting Algeria could hardly ever be heard from beginning to end. It was a choppy, broken voice." Lucas Hilderbrand coins the phrase "bootleg aesthetics" to describe the grainy images of analog video that gave rise to fair-use law and presaged many file-sharing practices in the United States. He, too, finds these videos all the more affectively powerful because of their low definition. Blurred images and distorted sounds could serve as material traces of a video's illicit circulation, adding a potential thrill or at least a call to identify with countercultures of circulation.
Limited definition can produce particularly intense modes of experience. Intense experience shaped by limited definition. This is an old point from Marshall McLuhan. In his classic essay "Media Hot and Cold," McLuhan discusses definition as an affective problem, rather than a reality problem. "A hot medium is one that extends a single sense in 'high definition.' High definition is the state of being well filled with data." Cool media are low definition "because so little is given and so much has to be filled in." With the television image, he writes, the eye must "act as hand in filling in and completing the image." Derived as it was from the everyday experience of watching black-and-white images flicker on cathode ray tubes and hearing sounds emanate from tiny monaural speakers with cheap transistor amplifiers, McLuhan's description of television as cool might well have felt ontological to the end user of 1964. Today, the range of television experiences available to the average person—from mobile-phone screens to HD—reveals television's coolness as a specifically infrastructural, industrial, and cultural condition. Coolness was an aesthetic that tuned perception to the limits of transmission infrastructure, and tuned transmission to the then-understood limits of perception. Using McLuhan's terminology, to say all media follow an historical trajectory toward high definition, to write media history in terms of a general history of verisimilitude, is to say that all are on a historical path toward hotness. But this is clearly not the case, either in his time or ours. Whether in its audio or data varieties, compression accommodates signals to infrastructures. But it also transforms infrastructures by enabling them to carry different kinds of signals.
More generally, we can define compression this way: compression is the process that renders a mode of representation adequate to its infrastructures. But compression also renders the infrastructures adequate to representation. There are thus at least two long-term tendencies in media history, and at least two grand narratives of media history we need to consider (and almost certainly more). The dominant paradigm is the general history of verisimilitude, where progress in technological history is therefore imagined as progress in terms of greater and greater definition. Too often, humanist critics of media forms echo advertising copy, as sound and images are unmoored from previously fixed stations governed by immediate experience, or reduced in definition or sophistication by virtue of their technological transmission (but with the promise of better definition—and more realism—in the next generation of technology). Humanists still write like travelers who want to anticipate every possible social or climatological contingency at their destination. The clothes preexist the trip—they take up less space in the suitcase if we roll them up and squeeze out the extra air to allow for a few extra garments to make the journey with us. In this line of thinking, compression squeezes the extra air out of sound recordings, phone calls, or videos that would otherwise take up more space as they reach out toward a horizon of fulsomeness. This chapter sketches out a general history of compression as another path through media history, and gives a rough outline of its contours. A general history of compression considers communication as based in a relational reality (often, though not necessarily, networked) and presupposes large-scale, collective activity in its positivity. Following Gilbert Simondon, there is a particularly useful insight from a history of compression for thinking about communication in general. Starting from compression, communication has a "network reality." This is to say that it is not a binary relationship between sender and receiver mediated by a medium but rather an ensemble of relations that only produce the moments of transmission and reception after the fact. For Simondon this is a kind of circular causality, or at least a relational causality, where a relation must exist to produce the things on which it has effects. As he argues in On the Mode of Existence of Technical Objects:
Elements that materially are to constitute the technical object, and that are independent one of the other, lacking an associated milieu that precedes the constitution of the technical object, must be organised in relation to one another by means of circular causality which will exist once the object is constituted. What is involved here, then, is a conditioning of the present by the future, or by what up to now does not exist.
Simondon's prime example of this process at work is the Guimbal dam in the Philippines, where the water moved by the turbine cools the turbine, enabling it to work at the properly regulated temperature to move water. For media scholars, this means that media are not like suitcases; and images, sounds, and moving pictures are not like clothes. They have no existence apart from their containers and from their movements—or the possibility thereof. Compression makes infrastructures more valuable, capable of carrying or holding materials they otherwise would or could not, even as compression also transforms those materials to make them available to the infrastructure. Having taken a short detour through theory, let us now take a second detour through the history of technology, to consider how compression gets named as a problem in the twentieth century.
* * *
In the field of communication technology, compression was first applied to audio. The Oxford English Dictionary gives 1938 as the earliest known use of the term compression as it is applied to audio. The December 1937 issue of the magazine Communications offers further insight into the term's use at the time in the United States. "Compressors" reduced the distance between the loudest and quietest parts of audio signals. Engineers started to call them "limiters" because of how they worked. Today this technique is called "dynamic range compression," and particular types of compression are called "limiting." The Communications article discusses the Western Electric 110-A amplifier and the products of two competitors, all of which had come onto the market in the previous year. In less than a year, "over half the radio stations in the country" had purchased one. To understand why radio stations rushed to buy these devices, we have to understand a bit about loudness and radio.
Compression solved a problem created by the Federal Communication Commission's method of regulating radio stations after 1927. Each station was allotted a certain channel and a certain maximum broadcasting power. Exceeding this allotted power, even for a moment, was called "overmodulation"—it was against regulation and it was considered rude. As explained by John P. Taylor, author of the Communications article, overmodulation introduced various kinds of signal distortion into the sound, but it also interfered with adjacent channels. It was the equivalent of shouting down your neighbor.
At the same time, the upper limit placed on broadcast power introduced certain aesthetic problems for radio broadcasters. Roughly speaking, broadcast power in wattage was a measure of how loud a station could be relative to other stations on the dial. Stations wanted to be as loud as possible for as much of the time as possible. This was part of a logic of capitalist competition, where it was expected that the consumer flipping through the dial would gravitate toward the louder signal. Although watts are not a very good measure of loudness, peak wattage represented peak loudness, and all other levels were relative. The range of possible loudnesses was called dynamic range. So if a station's average levels were quite a distance below its peak levels, most of the time it would not be using all of its allotted wattage, and that station would be effectively too quiet compared with its competitors, placing it at a commercial disadvantage.
(Continues...)
Excerpted from Signal Traffic by Lisa Parks, Nicole Starosielski. Copyright © 2015 Board of Trustees of the University of Illinois. Excerpted by permission of UNIVERSITY OF ILLINOIS PRESS.
All rights reserved. No part of this excerpt may be reproduced or reprinted without permission in writing from the publisher.
Excerpts are provided by Dial-A-Book Inc. solely for the personal use of visitors to this web site.
