Dirty Pixels


Dirty Pixels by Stella Brennan

A particle of representation, a pixel is the self-same grain of an image, given meaning by its structured relationship with other pixels. The pixel is a poignant, sometimes even palpable symptom of digital culture. Visual atomism, gridding and breaking images into monochrome, indivisible chunks, has a history stretching from classical mosaic to grainy photo id. Dirty Pixels explores the anatomy of the pixel, the slice and dice of the grid that underlies much of our visual culture. The artists in Dirty Pixels build worlds, fields of information from tiny repeating units.

The pixel is analogous to any other repeating element used to build a whole — the brush-strokes in a Seurat, grains of silver in photographic emulsion, stitches in a tapestry, any array of exchangeable components subsumed to the logic of a greater entity. In spite of existing only in informational space, pixelvision colours our worldview. Graphic fashion has seen the return of the pixel — bigger, bolder, more ironic. Chunky GIFs and jaggy fonts recall Lego-strewn bedrooms and all-night Atari binges. As consumer-electronic grainlessness approaches with the highly rendered scenes of Playstation2, and the smoothly perfect synthespians of Final Fantasy, those blocky styles have retro cachet (not to mention quick download times).

Dirty Pixels is about ideas and attitudes that feed into and out of digital culture, about corrupting a Cartesian dream. From Piet Mondrian’s austerely low-res boogie-woogie to heavily compressed jack-in-the-box porn, the grid gets grubby in the real world.

Dirt challenges the notional exchangeability of pixels, dirt has specificity, a pattern, it is locating. A dirty pixel is no longer an uninflected container of information; it has its own, corrupted character. How does a pixel get dirty? How does it become so corrupted? By coming down from the discreet binary world into the land of objects, by making the passage from the notional realm of numbers through the real-time crunch of dodgy hard-drives, crackly phone lines, flickering monitors and fugitive ink on paper.

The dirty pixel is rooted in synaesthetic thoughts about how pixels look, smell, taste. Lego is a key, remembering low-res childhood struggles to make buildings, or vehicles, or anything else for that -matter, out of the small pile of available pieces. Imagine hard, indivisible pixels, like those little shiny plastic blocks. Sharp-edged, painful underfoot, roll them around in your mouth, suck on the hard plastic like insoluble candy, clacking against your teeth.

Lego Worlds
Michel Gondry’s meticulous stop-motion Lego animation for Detroit band The White Stripes plays with this childhood analogy. In Fell in love with a girl Gondry filmed the band, pixellating the footage and then building each blocky image frame-by-frame in primary plastic brightness. The resulting music video is a combination of low-fi and high-tech, echoing the arch simplicity of the one guitar, one voice, one drum of the artpunk duo.

The dirty pixel challenges the purported transparency of digital media, its supposed texturelessness. Not all pixels are equal. The resolution and quality of its component pixels make up the character of a medium — the coarse cathode blue scan lines of a television screen, the pale silky grey precision of an LCD monitor. Pop artists knew this, lovingly painting the dot screen of the printed page, Roy Lichtenstein rendering every inky comic splotch.

Sara Hughes’ work, Software for Ada is a swarm of grey spots, a haze on the white wall, like printed dots on a white page, or holes in a punch card. The work is a meticulous combination of the mechanically produced and the handcrafted: each vinyl spot sliced out by computer-controlled blade, then painted a specific shade — dove grey, battleship grey, bone black. Like pointillist wallpaper the work frames the doorway, folding around the gallery surface, flickering and shifting its dimension. Digital photos of the work dissolve into moiré patterns as the grid of the camera display overlays the dot screen of the work.

Unbuttoning the coat, he thrust his hands into the trouser pockets, the better to display the waistcoat, which was woven in a dizzy mosaic of tiny black-and-white squares. Ada Chequers, the tailors called them, the Lady having created the pattern by programming a Jacquard loom to weave pure algebra.1
— William Gibson and Bruce Sterling, The Difference Engine.

Software for Ada references Ada Lovelace, recently resurrected mathematician, daughter of Lord Byron, and collaborator of nineteeth-century inventor Charles Babbage. Babbage was the inventor of the Difference and Analytical Engines, machines that foreshadowed the modern computer. Ada herself has been hailed as the first computer programmer. The revival of interest in these eminent Victorians marks the reentanglement of the computer, historically stereotyped as a masculine technology (notwithstanding all those women labouring in the cleanrooms of microchip factories) with the history of weaving. The punched card programmes of old room-sized mainframes are again intertwined with the punched pasteboard patterns of steam-powered Jacquard looms, cards which enabled them to weave complex scenes of foliage, flowers, fruit.

The hygienic labour of the shrink-wrapped clean-room worker is not so distant from domestic handiwork. As the history of the computer is reconnected to the history of the loom, feminist historians have traced occluded female labour from Victorian textile worker, to the crisp-haired and bright-suited ladies tending corpulent mainframes, to today’s electronics assembly piece-workers, labouring long hours with dangerous materials for little pay. The pixel has dirty hands.

Dot Matrix
The embroidery Tuesday, 3 July 2001, 10:38 am is a stitch-for-pixel representation of my computer screen, bearing the datestamp of the desktop snapshot it replicates. The time-consuming and often communal labour required to replicate the now commonplace metaphors of the Graphical User Interface forms part of the work. Taking over a year to complete, the production of the embroidery was an experience of the embodiment of information. Analogous to a rendering process, every stitch of thread has run through human fingers. Minor deviations from the pattern mean no two file icons are identical. Alterations in tension and thread colour are other indications of the hand-made. Making the work was a journey into all those Photoshop commonplaces we take for granted — you’ll never feel the same about anti-aliasing once you’ve done it by hand. Often I had the sense I had shrunk, and was crawling across the desktop as if across some enormous terrain.

Tuesday, 3 July 2001, 10:38 am is a self-portrait, a mapping of a personal space. The work represents a world unto itself — the mnemonic device of the hard drive — but simultaneously points away, to the technologies it is a component part of. While the embroidery represents a specific desktop image of a particular computer at a precise moment, it also refers out to all other similar desktops constructed according to the metaphors of the Graphical User Interface.2

Obsessive, repetitive labour characterizes the works in Dirty Pixels. Strange programmes are rigorously implemented. Working on standard 1mm graph paper, Martin Thompson builds up paired drawings, squares of meticulously rendered pattern. The drawings’ creased and dirty edges betray the time and labour contained in their construction. Thompson draws freehand, colouring directly, using tape and scalpel to graft new sections and erase errors. The artist builds fields of intricately arranged blocks, shapes with intimations of stars, Pac-men and snowflake-patterned knitwear. Order and symmetry are apparent, but at times the images almost teeter into white noise. Thompson’s preferred Day-Glo inks create shimmering figure-ground ambiguities. The tiny squares swarm, resembling the glowing phosphenes of hallucinogenic or closed-eye vision, the white noise of the eye.

The dirty pixel draws on notions of digital decay, particularly on Lev Manovich’s argument that lossy compression, the squeezing out of information in order to fit large amounts of data down narrow channels, or onto say, a DVD, represents the true aesthetic of digital media.

…rather than being an aberration, a flaw in the otherwise pure and perfect world of the digital, where not even a single bit of information is ever lost, lossy compression is the very foundation of computer culture, at least for now. Therefore, while in theory computer technology entails the flawless replication of data, its actual use in contemporary society is characterized by loss of data, degradation, and noise. 3
— Lev Manovich, The Language of New Media.

Taking Manovich’s lossy aesthetics to their logical extreme, Tim Ryan’s work Crash Media riffs on the fancy algebraic compressions we rely on to digest video and sound into bandwidth-friendly formats. Ryan crunches and stretches the stream of digital images till the colours smear and pixels dissolve into blocky swarms. There is a sympathy between the panel-crunching car crashes he samples and the grinding codecs he abuses. Transformed by this process into ambiguous boiling, tumbling smears and blocks of colour, the footage occasionally reveals a recognisable glimpse of wheel or fender through the algorhythmic haze.

The Grid
The pixel gains meaning through its co-ordinates, its insertion in the Cartesian grid. The order and evenness of the grid, its democratic distribution, rules the pixellated image. Uncompressed, it is a uniform, continuous space.

In her 1978 essay, “Grids”, Rosalind Krauss draws a distinction between two types of grid. Firstly she describes the perspectival grid of Leonardo or Duhrer, where the perspective lattice structures the transformation of perceived space onto the picture plane, and secondly the self-contained use of the grid in modern art.

Perspective was the demonstration of the way reality and its representation could be mapped onto one another, the way the painted image and its real-world referent did in fact relate to one another… everything about the grid opposes that relationship… if it maps anything, it maps the surface of the painting itself. It is a transfer in which nothing changes place.4
— Rosalind Krauss, “Grids” The Orginality of the Avant-Garde

Krauss asserts that perspective contains an indissoluble tie to the representation of an external world, while the modernist grid refers back only to itself. But technologies of perspective, and the contemporaneous development of accurate cartography have worked to impose the abstract lines of the grid onto the form of the land. Conversely, modelling software is designed to generate perspectivally accurate renderings of imaginary wire-frame landscapes.

Joyce Campbell’s sprawling photographic grid LA Bloom confounds this distinction between the grid as the structuring logic of a landscape and the grid as a world unto itself. LA Bloom is a series of photograms infected with the images of fungi and bacteria gathered from around Los Angeles. From her field samples, the artist inoculates nutrient agar plates the same size as the final photogram. Days or weeks later, after the colonies of microbes have grown into their characteristic splotches and nets, the plates are direct printed onto cibachrome paper. Created in a one-to-one relation without recourse to camera or lens, they resemble Borges’ tale, related by Baudrillard, of a one-to-one scale map of a kingdom lying obscured beneath it. 5

LA Bloom traces the grid of Los Angeles. The photograms are arrayed on the wall according to the location of their parent samples in the plan of the city. These microbial samples stand in for landscape, the invisible city of spores and scrapings. The photograms trace the colonies’ slow spread across their agar terrain, revealing in grainlessly intimate detail delicate tendrils, greasy spots, the odd stray hair. In total, the work creates an image of the sprawling city, formed from the bodies of its lowliest inhabitants.

How does a pixellated image differ from these chemically produced photographs, where grains of silver in emulsion serve the same recording purpose as the raster grid? Converting images to binary information gives the digital image its main distinction, its mutability. In its coded essence the pixel can contain many different values, represent many kinds of image, whereas once a film is shot or paper printed, it is capable only of accreting more images. The photographic emulsion cannot shake itself off, wipe itself clean and begin again. LA Bloom complicates the relation between the cold, hard Cartesian pixel and goopy self-replicating organic worlds. Cells, subsumed to the logic of the colony, are, like the pixel, or like the grain of the cibachrome paper that images them, similar parts of a greater whole.

LA Bloom exhibits the indeterminate scale we experience in a world where libraries are folded into the surface of an etched wafer of silicon, where resolution, the level of detail in an image, is unhinged from its physical dimension. Diminutive in relation to the city grid that forms its structuring logic, Campbell’s grainless photograms encapsulate vast fields of data. In terms of stored visual information they are enormous. This ambiguity enables each agar plate to form its own microcosm. It is only the dirt and hairs trapped in the photograms that pop the work into scale.

Dirt has specificity, it is evidential, showing passage through the world, and acting as a marker of authenticity. The two channel video work Zen DV uses the built-in Dust and Scratches filter from the video editing software Final Cut Pro and a preprogrammed Record Noise plug-in. The filters are applied to video signal generated by the software itself — bluescreen and bars and tone — images and sound that mark an absence waiting for information (bluescreen) or an equipment test (bars and tone). The filters emulate on digital video the degraded emulsion of film stock and the hiss and scratch of old vinyl.

Zen DV plays on the strangeness of simulacral dust and scratches and pays homage to Nam June Paik’s 1965 work Zen for Film. Paik’s film plays out in light John Cage’s aleatory means of constructing artworks. Zen for Film is a clear loop of film with no sound-track. The array of injuries to the film’s surface creates the work. ZenDV is subject to different kinds of loss than Paik’s slowly degrading loop — image compression, smears and scratches on the surface of the disc — but unlike the film which is eroded by each performance, the scratched and dirty image that the digital video presents will be the same the first and the thousandth time it is played.

ZenDV also bears a strong relation to Paik’s work Zen TV. Paik manipulated the scanning beam of a television so that it formed a single line. The electron gun, rather than playing democratically over the screen’s surface then accreted every scan line on top of every other. The visual information was compressed into a tiny strip of indecipherable light running along the center of the screen. Modifying a mass medium to extract new meanings from it, Zen TV is, in retrospect, a parable of compression — all the flickering data of the televisual image flattened into a narrow band of light.

The vagaries of chance embraced by Cage and Paik in generating artworks are rigorously excluded by the mathematical permutations programmed by John Simon’s 1997 work EveryIcon. The work displays a 32 by 32 pixel grid, the standard dimension of computer icons. The grid is bitmapped black or white, on or off, and within this tiny universe every possible combination is trialled by the software. The grid flickers as the computer’s processor runs through thousands of possibilities per second. But even in this constrained environment the time required to exhibit all possible combinations will exceed the life of the universe. Chance is excluded, but only notionally.

Once I went to Texas, to Donald Judd’s Chinati Foundation. We weren’t allowed to take pictures of the artworks there, so I spent my time photographing the animals and plants slowly invading the blank minimal structures — the tendril of grass forcing its way between two concrete slabs, the spiders in the corners. For me the photographs became images of the exaggerated entropic state of these programmatic artworks. Like the dirty pixel, the cobwebbed Judd snaps into specificity. Its equivalence and replicability are compromised. No longer a subset of an infinitely extending system, it is located by that blade of grass and that spider, placed in a field under the hot afternoon sun with the tourists and the rattlesnakes.

Like the primary structures of Judd’s concrete cubes, you could think of a monochrome as an image of a single pixel. Think of Malevich’s white square, tilted, self-contained, grubby with the passage of time. Perfectly self-reflexive, yet marked by its yellowing pigment as a historic artifact. Similarly self-contained, the pixel shines. Hanging in dark Cartesian space, pinioned on a fixed axis, a radiant mathematical possibility. But once a pixel comes down into the world, once it stops being an abstract sequence of numbers, once a pixel decays, gets dirty, fingered, reconfigured, it exits the virtual. Once this particle of representation becomes a familiar object of use it becomes ours. As a pulse of pure electricity it is clean, discrete and distant, but as we draw it through the smeary depths of the screen it becomes our own.


1. Gibson, William and Bruce Sterling. “The Difference Engine” quoted in Zeroes and Ones Sadie Plant London: Fourth Estate 1997 p23

2. This pointing is literally depicted in the Microsoft Internet Explorer icon: a small, pink, generic hand points to a globe. The cartoonish paw traces lines of power, its indicating finger touching down somewhere around New York.

3. Manovich, Lev. The Language of New Media Cambridge, Massachusetts: MIT Press, 2001 p55

4. Krauss, Rosalind. “Grids” The Originality of the Avant-Garde and other Modernist Myths Cambridge, Massachusetts: MIT Press, 1985 p10

5. Baudrillard, Jean. “Simulations” translated by Paul Foss, Paul Patton and Philip Beitchman. Semiotext(e) New York 1983 p1