Acoustic Erasure: Class, Capital, and the Material Bias of the Sound Archive

The Politics of the Unheard

The historical sound archive is an archive of capital. A recording survives through a physical chain, and every link in that chain carried a price. The firm paid for the session, chose the pressing compound, stamped enough copies to outlast their owners, and sold them into homes where playback left the grooves readable. Every condition favored music that commanded an established market. Sound reproduction was an industrial commodity, manufactured and marketed under the commercial arrangements that governed telephony, radio, and consumer goods.^[1]

Legal access narrows the inheritance further. Rights holders have reissued on compact disc only 14 percent of pre-1965 recordings, and no more than 10 percent of recordings made before the Second World War.^[2] Twenty percent of country recordings issued between 1890 and 1964 were commercially available in the United States, against 10 percent of blues recordings and 1 percent of recordings marketed as ethnic.^[3] Playback hardware for obsolete formats became institutional equipment, and copyright left reissue to rights holders without incentive to release back catalogs.^[4] For the unreissued majority, the original disc is the only witness, and the condition of that disc limits what any later listener can hear.

Silence in the sound record is a manufactured condition. Power enters the production of history at the creation of sources, at the assembly of archives, at the retrieval of facts into narrative, and at the moment a society decides what the past means.^[5] Sound recording adds a physical encounter between needle and groove, repeated thousands of times in parlors, kitchens, and boardinghouses. Chemistry and price set the cumulative effect of that contact. The industry sorted vernacular music into racially segregated catalogs, dividing race records from hillbilly releases in numbering, advertising, and distribution.^[6] Independent labels recording Black, rural, and immigrant performers pressed cheap compounds in short runs.^[7] Their discs reached listeners who owned the most worn machines. The absence of that music from the reissue catalog completes a loss that began at the press.

The Industrial Chemistry of the Groove

What left the press - 78's - was mostly stone. Manufacturers bound powdered slate, limestone, or clay with imported lac resin, added cotton flock for tensile strength, and colored the mixture with carbon black.^[8] Shellac made up a minority of the mass, varying with the price of imported lac, the grade of filler, and the formulation of each plant. Mixers rolled the heated compound into sheets and cut them into portions the trade called biscuits. A press operator softened each biscuit on a steam table, laid it between two nickel stampers, closed a hydraulic press, and chilled the platens until the disc set.^[9] The groove that resulted was a negative of the stamper cast in resin and mineral grit, and the fineness of that grit set the floor of audible noise before any needle touched the disc.

Compound formulation separated premium labels from budget pressings. A premium compound used finely ground, evenly dispersed mineral, a cleaner grade of lac, and little reground scrap. Columbia pressed laminated discs that bonded a thin layer of fine surface compound over a coarse core, keeping the needle on the expensive layer.^[10] Reground scrap from rejected pressings, edge trimmings, and returned inventory went back into the mixers at every plant. A higher proportion of regrind carried fibers, label paper, and uneven grit into the playing surface. Each impurity registered as crackle and hiss, weakening the groove wall against a passing needle.

While major firms separated vernacular music through marketing and distribution, the sharpest material divide fell between firms.^[11] Paramount was a sideline of the Wisconsin Chair Company, which manufactured phonograph cabinets and pressed records to sell with them.^[12] From the outset Paramount pressed popular releases on coarse shellac, and with electrical recording both sound and compound fell below major-label standards.^[13] Paramount's pressing plant also manufactured discs for Black Swan, the first major Black-owned label, and Paramount acquired Black Swan's catalog when that company failed. Through the late 1920s Paramount became a principal source of blues and gospel recordings in the United States.^[14] It later moved recording into a studio inside the Grafton plant that pressed its records. That room was acoustically poor enough that engineers hung burlap on the walls.^[15] The compound, the studio, and the pressing runs answered to the economics of a furniture business, and many Paramount issues now survive in a handful of copies, most of them worn.^[16]

Abrasive filler was not an accidental defect of budget pressings, but an industry design specification. Shellac grit was harder than the steel of a gramophone needle, so the record shaped the needle during opening revolutions until the point matched the groove contour.^[17] Manufacturers printed instructions on needle tins and machine manuals directing owners to fit a fresh needle for every side. A 1909 patent for a jewel-tipped stylus documented the trade's structural compromise. The steel needle seated within opening turns, but when listeners avoided replacing it after each side, the blunted steel planed the record.^[18] The format moved the mechanical cost of fitting needle to groove onto a cheap consumable part. That cost passed to the disc surface whenever the consumable went unreplaced.

The Domestic Chisel and Needle Economics

In the home, the damage caused by an unreplaced needle depended on the reproducer driving it across the groove. Playback hardware fell into economic classes tracking household income. At the top stood cabinet machines sold as parlor furniture. In 1925 the Victor Talking Machine Company introduced the Orthophonic line, whose reproducers and folded horns engineers designed alongside electrical recording. At the bottom stood portable phonographs, secondhand cabinets, and mail-order models bought on installment. Acoustic reproducers of every class bore down on the groove with forces measured in ounces rather than grams, and early electrical pickups remained heavy through the 1930s.^[19] Force alone did not separate expensive machines from cheap machines. Condition did. A reproducer whose rubber gaskets had hardened with age transmitted needle motion stiffly, dragging the point against the groove wall instead of following it. A misaligned tone arm pressed the needle harder against one wall than the other. A worn turntable bearing let speed waver and the needle chatter. These faults accumulated in machines that passed through several owners, and machines that passed through several owners accumulated in poorer households.

The needle completed the economy of wear. Steel needles sold in packets of a hundred for small sums, yet changing needles with every side multiplied that expense across an evening of listening. Fiber and thorn needles spared the groove but required special cutters and frequent resharpening, while tungsten styli cost more per unit; both sold groove protection at a surcharge.^[20] Manufacturers assumed a listener who bought needles freely and changed them after every play. Thrift, distance from a dealer, and neglected maintenance worked against that assumption. The resulting wear concentrated on the records a household played most, which in a household that owned few records meant the same discs night after night.

A conical steel needle rides on the flanks of a V-shaped groove. Abrasion cuts two flat facets into the steel at contact, leaving sharp edges where the flat meets the original cone.^[21] Fitted to the next disc, whose groove differs in width and angle, the worn needle no longer seats. Its edges scrape the groove walls, and its broadened contact face can no longer trace tight modulations. These tighter curves carry the highest frequencies and compress toward the center of the disc, where groove speed drops. A tip too broad to follow them rides across their crests and planes them flat.

The material lost in this process carries articulation. Sibilants and plosives in a singer's diction, the pick attack of a guitar, the strike of a cymbal, and the upper partials distinguishing one voice from another sit in the upper register of the groove. Each pass of a chisel-worn needle shaved compound from the walls, leaving fractured shellac and embedded grit. Each later needle, fresh or worn, rode through that damage. The rising noise floor of a heavily played budget disc is the physical record of its use.

Field Recording and Instantaneous Media

Yet commercial pressings, however abrasive their compounds, distributed the risk of loss across thousands of copies stamped from metal matrices. Field recordings had no pressing plant, no production run, and no commercial market to multiply them. They existed only as unmultiplied originals cut into soft wax, bare aluminum, or lacquer. From 1890, government ethnologists and academic fieldworkers carried cylinder phonographs onto reservations, into rural communities, and later into prisons and labor camps, recording songs, narratives, and word lists on soft wax.^[22] The Library of Congress holds roughly fifty thousand cylinders, the largest collection in the world, many dating to 1890.^[23] Wax cracks with temperature shifts and age, softens in heat, and supports mold.^[24] Recordings traveled by wagon, train, and automobile through summers in the Southwest and Deep South, waiting in trunks for weeks or months before reaching a laboratory. In 1979 the Federal Cylinder Project copied and inventoried the field cylinders held in federal collections, and from 1985 returned copies to the communities whose music the cylinders held.^[25] By the time of that survey, many cylinders had cracked, molded, or worn past recovery.

Fieldwork wore its recordings at the moment of their creation. Fieldworkers played cylinders back to performers on site. Playback confirmed that the recording had taken, drew out corrections and commentary, and persuaded singers to continue once they had heard their voices returned by the horn.^[26] Each playback drove a stylus through soft, freshly cut wax on equipment built for portability rather than gentleness. The practice that secured consent and accuracy in the field was also the first act of wear on the recording.

In the summer of 1933, John Lomax and his teenage son Alan drove across the South for the Library of Congress with a three-hundred-pound recording machine installed in the trunk of a Ford sedan. They recorded in state penitentiaries in Texas, Louisiana, Mississippi, and Tennessee, including the work songs, hollers, blues, and spirituals of incarcerated Black men and women at Sugar Land, Angola, and Parchman.^[27] The machine embossed a shallow groove into blank aluminum discs with a steel stylus, running without a speed governor across a narrow bandwidth. Nondestructive playback required a fiber or thorn needle; playing the disc back with a steel needle in a heavy pickup gouged the soft metal.^[28]

Lacquer discs, which by the mid-1930s replaced bare aluminum in the field, cut a deeper, quieter groove into a coating of nitrocellulose spread over a rigid base.^[29] The coating held more signal and resisted wear even less. When aluminum diverted to war production in the 1940s, manufacturers of instantaneous blanks substituted glass for the base. Glass-based lacquers shatter when dropped or jolted, and delaminate more readily than the aluminum discs they replaced.^[30] Lacquer deteriorates from within. A plasticizer migrates out of the coating as a white crystalline residue of palmitic acid, the coating shrinks against the base, and the lacquer cracks, peels, and flakes away until the signal is lost.^[31] Once delamination occurs, no treatment reverses it.

Fieldwork delivered its recordings to institutions marked from their creation by the conditions of collection. The cylinder bore the wear of its first playback. The aluminum disc bore the shallow groove of an underpowered cutter and the gouges of steel needles. The lacquer carried chemical instability from the day of manufacture. Recordings outside commercial markets, made among populations with no claim on institutional preservation, entered the archive on the most fragile media the twentieth century produced.

Optical Recovery and the Testimony of Noise

When mechanical contact risks destroying fragile media, optical recovery reads a groove without touching it. In 2003, experiments demonstrated that audio stored in a groove could be reconstructed by imaging groove topography with precision metrology and processing the images digitally.^[32] Two-dimensional imaging reads the lateral modulation of a disc groove from above. Three-dimensional scanning with a confocal probe, measuring surface height to a tenth of a micron, maps the hill-and-dale modulation of cylinders and vertically cut discs.^[33] Optical systems read media no mechanical stylus can touch, including lacquers in advanced delamination and glass discs shattered into fragments.^[34]

Optical recovery is allocated by funding. Scanning proceeds slowly, a disc at a time, under equipment and specialist labor few institutions can support. At one university collection of poetry recordings on deteriorating lacquer, the cost ruled out scanning the full holding, and curators hoped scholarly interest in particular discs would attract funding.^[35] Selection by institutional budget reproduces, at the stage of rescue, the commercial logic of market and attention that governed pressing and playback.

The method's advantage follows from its geometry. A stylus touches only the portion of the groove wall on which it rides. An optical system can sample any region of the groove, including portions no needle ever contacted, while image analysis compensates for surface scratches, cracks, and contamination.^[36] It cannot restore material a worn needle has cut away. Where a chisel-faced needle planed modulation crests flat, the scanner images a flattened crest. Groove damage and acoustic signal arrive together in the same measurement.

Because metrology cannot separate performance from the physical degradation of its carrier, archival practice separates preservation from presentation. International standards require an unaltered preservation transfer capturing the signal as faithfully as the carrier allows, reserving noise reduction and subjective processing for secondary access copies.^[37]

The listening public encounters versions released for sale and streaming. Reissue producers working from noisy shellac declick, apply spectral repair, and subtract broadband noise. Statistical restoration methods model surface noise and subtract the estimate from the signal. Because noise and the upper partials of voices and instruments occupy the same frequency bands, every gain in quiet costs signal and introduces processing artifacts of its own.^[38] The more aggressive the setting, the more of a singer's consonants and a guitar's attack goes with the hiss. For recordings surviving in a few worn copies held by private collectors, the restored reissue becomes the public form of the work. The access copy presents the performance as though compound, needle, and field trip had left no trace.

The worn disc, the unprocessed transfer, and the optical scan retain that trace. The grain of hiss under a Paramount side measures the grade of its compound and the size of its pressing run. The scraped sibilants of a heavily played ballad measure the price of steel needles against a household budget. The thin, gouged groove of a 1933 aluminum disc measures the limits of a portable cutter and the hardness of the needles that played it back, first for the singers who made it and then for the listeners who followed them. Surface noise, chiseled groove walls, and the residue of friction are the material record of the economic conditions under which early sound was made, sold, and heard, inscribed in the same groove that carries the song. To strip that residue away in the name of fidelity is to mistake the survival of capital for the survival of the past. The silence that results preserves the music only on condition that it forgets the world that paid for it.