In the summer of 1877, in a workshop in Menlo Park, New Jersey, Thomas Edison spoke a nursery rhyme into a brass mouthpiece while turning a hand crank, and a needle traced the vibrations of his voice into a sheet of tinfoil wrapped around a cylinder. When the crank turned again, the rhyme came back, thin and wavering but recognisable, out of the same horn. Nothing in that moment sounded like an industry. It sounded like a parlour trick. Yet the trick contained the whole argument of this article, because how recorded music changed music begins with that reversal: a performance, which for all of human history had vanished at the instant it was made, had become an object that could be kept, carried, sold and heard again by people who were never in the room.

Recording did not simply preserve music. It changed what music was made of, because the moment a performance became a repeatable object, the incentives governing length, arrangement, improvisation, virtuosity and fame all inverted. A song that once had to fill an evening now had to fit a disc side. A solo that once evaporated now had to survive rehearing. A singer who once had to reach the back row of a theatre now had to reach one listener at a time, at close range, through a machine. Mark Katz, in Capturing Sound, published in 2004, gave this cluster of effects a name, the phonograph effect, meaning the ways a recording technology reshapes the musical practices it was built only to capture. Katz treated the effect as reciprocal rather than one way: musicians changed the machines by demanding more from them, and the machines changed the musicians by rewarding some habits and punishing others. That reciprocity is the thread through every section that follows, from the three minute disc to the multitrack studio, and from the village field recording to the global trade that carried local sounds into distant homes.

One plain fact shows the scale of the inversion, and it deserves stating without decoration. Before recording, a listener who wanted to hear a particular performance again had exactly one method, which was to persuade the performer to perform it again, in person, in the same place or another place the listener could reach. After recording, the same listener could hear the same performance a hundred times without the performer being present, awake, alive or even aware that the hearing was taking place. Every other change described in this article, the shortened song, the frozen solo, the quiet croon, the edited master, the exported tradition, is a consequence of that single mechanical fact working its way through a century of musical life.

How music travelled before recording

Before a performance could be carried, music travelled by three routes, and each route taxed the music in a different way. The first route was the travelling musician, who carried repertory in memory and in the hands, and who changed it at every stop because memory edits as it carries. A ballad that crossed a county acquired new place names. A dance tune that crossed an ocean acquired a new tempo suited to new floors and new shoes. Variation was not a defect of this system. Variation was the system, the means by which a tune stayed useful to communities whose lives differed from the lives of the people who had made it.

The second route was notation, the written score, which allowed music to travel without its maker and to outlast its maker by centuries. Notation carried pitch and rhythm with reasonable fidelity, and it built the Western classical tradition on the premise that the written work was the durable object and the performance a temporary realisation of it. Yet notation carried only what it could encode. Timbre, attack, vibrato, the grain of a particular voice, the exact push and drag of a rhythm section, all of that stayed behind with the performer. A score is a recipe rather than a meal, and two honest performances of one score could differ as widely as two kitchens working from one page. The gap between the written work and the sounded event was where interpretation lived, and listeners before recording accepted that gap as a law of nature rather than a problem awaiting a machine.

The third route was teaching, the master and pupil chain, in which technique and repertory passed hand to hand inside guilds, families, churches and courts. Teaching preserved subtleties that notation missed, because a teacher could correct a wrist, a breath or a bow arm in the moment of error. Teaching also rationed access. A tradition held by a small lineage stayed small, however fine it was, and it died when the lineage broke, through war, migration, poverty or simple failure to find a pupil with the patience to receive it. Jonathan Sterne, in The Audible Past, published in 2003, situates sound reproduction inside this older landscape of listening practices rather than outside it, arguing that the phonograph answered habits of attention and desires for permanence that listeners already held. The machine did not invent the wish to keep a sound. It industrialised a wish that teachers, copyists and mourners had carried for centuries, and it did so at a price and a scale that turned a wish into a market.

What none of the three routes could do was move the sound itself. A famous singer’s voice stopped at the walls of the room where she sang, and reports of it travelled as metaphor, as comparison, as reputation, never as evidence. A listener in another country who read that a tenor had a golden tone had no way to test the claim. Fame before recording was therefore local and testimonial, built from the accounts of witnesses, and it faded with the witnesses. The sections ahead trace what happened when the sound itself became portable, first as a fragile cylinder, then as a disc that could be pressed by the thousand, and at last as tape and files of data that could be copied without the original performer being consulted at all.

The first recording technologies

The phonograph of 1877 was not the first device to write sound down. In 1857, in Paris, Edouard Leon Scott de Martinville patented the phonautograph, which traced airborne sound as a wavy line on soot blackened paper turned by a crank. The phonautograph could draw a voice but could not play it back, because its purpose was study rather than reproduction, and for two decades its traces sat in archives as silent pictures of sound. Edison’s contribution was the return journey: a groove cut into a moving surface could guide a needle back through the same vibrations and set a diaphragm moving in sympathy, so the stored sound became audible sound again. Charles Cros in France described a similar method on paper in the same year, 1877, without building a working machine, and the near simultaneity matters because it shows the idea was ripe. Several workshops stood one step from it, and the step was taken where capital, patents and publicity stood ready to exploit it.

The tinfoil machine was a novelty that faded within a few years, and recording survived its first death through office dictation rather than music. Alexander Graham Bell and his associates replaced tinfoil with wax, producing the graphophone, and Edison answered with an improved phonograph using solid wax cylinders that could be shaved and reused. Through the 1880s and 1890s the cylinder trade sold spoken word, comic monologues, brass bands and popular songs to parlours and to coin operated arcades, where a listener paid for one hearing through rubber tubes held to the ears. Each cylinder at first had to be made by performing into the machine, or by pantograph copying that weakened the sound with every generation, so a popular title existed in small numbers and varied from copy to copy.

Emile Berliner broke that bottleneck. His gramophone, patented in 1887 and marketed through the following decade, recorded sound as a sideways wiggle in a groove on a flat disc, and the flat master could stamp out copies by the hundred in a press. The disc was the decisive industrial form because it separated the performance from the manufacture: one good master, cut once, became ten thousand identical objects shipped in paper sleeves. Greg Milner, in Perfecting Sound Forever, published in 2009, follows this separation through the whole history of the industry and treats it as the founding bargain of recorded music, fidelity traded against reach, the single event multiplied until the multiplication itself became the business. By 1900 the elements of the modern trade stood assembled: a machine in the home, a catalogue of titles, a star whose name sold the object, and a groove whose length quietly dictated how long a piece of music was allowed to be.

What early machines could and could not capture

The acoustic recording process, which governed the industry until the middle of the 1920s, captured sound by funneling it through a horn onto a diaphragm linked to a cutting stylus, with no electrical amplification anywhere in the chain. The energy that cut the groove was the energy of the performance itself, so loud instruments cut deep and quiet ones barely cut at all. The frequency span was narrow, roughly the middle band of human hearing, which meant that a double bass lost its foundation, a bass drum lost its weight, and the high shimmer of cymbals and violin harmonics vanished entirely. What survived was what sat in the middle: voices, cornets, clarinets, pianos thinned to their percussive core, and the bright attack of banjos and xylophones, which is one reason early dance records sound so plucked and so eager.

Performers adapted their bodies and their ensembles to the horn. Singers stood close and turned away on high notes that might blast the diaphragm into distortion. String sections were cut down or replaced, in some studios, by Stroh violins, instruments fitted with metal horns of their own that aimed their sound at the recording horn like trumpets. Drummers abandoned full kits for woodblocks and muted pads. A whole craft of recording balance grew up in which the engineer arranged people in space, nearer or farther, higher or lower, and the arrangement was the mixing, done with feet and furniture rather than with faders. Timothy Day, in A Century of Recorded Music, published in 2000, stresses how completely these constraints shaped the recorded repertory of the acoustic era: what the horn could not hear, the catalogue largely did not sell, and later listeners who met early orchestral music only through acoustic discs inherited a repertory filtered by a funnel.

The limits also taught listeners a new tolerance. A cylinder or disc arrived with surface noise as a permanent companion, a hiss and crackle that the ear learned to listen through the way an eye learns to read in poor light. Fidelity in the acoustic era was therefore a negotiation rather than a measurement, and the word itself, as Sterne shows in his account of the period’s listening culture, carried moral weight: a faithful machine, a faithful copy, a faithful servant of the original sound. That moral vocabulary never left the industry. It returned with electrical recording, with high fidelity marketing in the 1950s, and with every later format war, each generation accusing the previous one of betrayal and promising, at last, the sound itself. The promise was always partly true and always partly a sales line, and learning to hear the difference between the two became one of the basic skills of the recorded century.

The technologies can be compared directly, because each one traded a freedom against a constraint and the music settled around the trade.

Technology	What it made possible	What it constrained	The musical change that followed
Wax cylinder	A single performance kept and replayed in the home	Two to four minutes a side and copies made one at a time	Parlour repertory of short songs, monologues and band selections
Flat disc at seventy eight revolutions	Mass pressed identical copies from one master	Roughly three minutes on a ten inch side	The three minute popular song and the compressed jazz solo
Electrical recording with the microphone	Full frequency range and quiet, close singing	Balance decided at a control desk away from the players	Crooning, restored orchestral weight and the professional engineer
Magnetic tape	Cutting, splicing and assembling a master from takes	The master no longer proved one continuous performance	Edited classical performance and songs built rather than captured
Multitrack tape	Parts recorded at different times and balanced afterward	Simultaneity ceased to be required for ensemble sound	The studio album as a constructed work and the producer as co author
Long playing record	About twenty two minutes a side	The side break and the album length as organising facts	Sequenced albums and extended jazz and classical programmes
Compact disc and digital editing	About seventy four minutes, silent surfaces and note level correction	The copy lost nothing, weakening the scarcity of the object	Longer albums, routine correction and the file sharing crisis

The three-minute limit and why it existed

Why were early songs about three minutes long?

Early discs turned at about seventy eight revolutions each minute, and a ten inch side held roughly three minutes before the groove ran out. Singers and bands fitted verses, choruses and solos inside that span because a longer performance simply could not be pressed onto one side of the disc sold.

The arithmetic behind that span was physical and unforgiving. A disc turning at about seventy eight revolutions each minute carries its groove in a spiral from the rim toward the label, and the spiral has a fixed amount of room. Cut the groove wide, with loud passages swinging broadly from side to side, and the spiral must space its turns apart, which shortens the playing time. Cut it narrow and quiet, and more turns fit, at the price of a faint record drowned by surface noise. The ten inch disc, the industry’s standard popular format for half a century, settled at roughly three minutes a side as the working compromise between loudness and length. The twelve inch disc held four to five minutes and was reserved for classical excerpts and prestige titles, which cost more and sold in smaller numbers, so the popular song lived its life inside the smaller circle.

Cylinders imposed the same discipline in another shape. The standard cylinder ran two minutes, and the later four minute cylinder doubled that by halving the groove spacing, which doubled the fragility as well. Performers who recorded for both formats learned to prepare two arrangements of the same material, a habit that reveals how completely the object governed the art. Nobody in 1905 sat down to write a three minute song because three minutes was felt to be the natural length of a musical thought. Writers and arrangers worked backward from the groove. The verse that repeated once rather than twice, the introduction cut to four bars, the instrumental chorus dropped entirely, each of these decisions was made in the shadow of a spiral that ended where it ended, and the cumulative effect of millions of such decisions was a musical culture that experienced three minutes as a norm long after the physical reason for it had weakened.

What the limit did to musical form

A limit repeated often enough stops being felt as a limit and becomes a form. The three minute side trained songwriters to front load a song, to state the hook early, to keep introductions short and to treat the ending as a button rather than a winding down, because a listener deciding whether to buy, to replay or to drop another coin could not be asked to wait. Verse and chorus architecture tightened accordingly. The thirty two bar chorus of American popular song, four eight bar phrases in a pattern that states, restates, contrasts and returns, fitted the side almost exactly at a danceable tempo, and the fit was not a coincidence. The form and the format grew up together, each confirming the other’s sense of how long attention lasts and how a song earns its close.

For jazz and blues the side acted as a frame for the solo. An ensemble recording of three minutes could hold a statement of the theme, two or three solo choruses and a return, so the soloist learned to build an argument in thirty seconds, entering at full character, developing one idea rather than surveying several, and exiting cleanly. That compression produced a distinctive rhetoric, the recorded solo as epigram, and musicians studied those epigrams by replaying them until the phrasing entered their own hands. The frame could also falsify. A band whose live performances stretched a tune across fifteen minutes of collective exploration appeared on disc as a tidy miniature, and generations of listeners mistook the miniature for the music’s natural scale, a misunderstanding that later live recordings and longer formats only partly corrected.

Classical music met the limit from the other side, as amputation. A symphony movement that ran twelve minutes had to be cut, hurried or spread across several discs sold as a set, and early classical recording is full of all three expedients. Conductors took faster tempos to fit sides, orchestras recorded single movements as calling cards, and the four minute side break taught a generation of listeners to hear great works as strings of excerpts. The album, in the original sense of a bound set of discs holding one long work, took its name from the photograph album it resembled, and the word survived the object by decades. When the long playing record arrived in 1948 the classical world treated it as liberation, while the popular world, shaped by two generations of the short side, largely kept the three minute habit as a matter of craft, radio programming and audience expectation rather than necessity.

Radio deserves a footnote in this section because it offered repetition of a different kind. A broadcast could be heard by millions at once and then vanished unless somebody recorded it, so radio taught audiences breadth while the disc taught them depth, and musicians worked both systems at once, using broadcasts to advertise the objects that shops could sell. The division of labour between the two media, one ephemeral and vast, one durable and private, structured musical fame for half a century, and performers calibrated their art to it, saving their most repeatable achievements for the groove and their most expansive for the air and the stage. When home tape recorders spread after the Second World War, listeners began capturing broadcasts for themselves, and the line between the two systems blurred in the parlour long before the law or the industry admitted that listeners had become archivists of their own listening.

Repeatability and the improviser

The deepest change recording brought to improvised music was not capture but return. A performance that can be replayed can be studied, and a solo that can be studied can be memorised, copied, taught and surpassed by people who never met its maker. Katz describes this as one of the clearest phonograph effects: repetition turned improvisation, the art defined by its disappearance, into a text. Students of the trumpet in the 1920s and 1930s lifted the needle and dropped it again on the same chorus until the chorus was theirs. Conservatories that had once dismissed jazz as unteachable acquired, in the disc, a syllabus that wrote itself, one celebrated solo at a time.

Repeatability also changed the improviser’s own relation to the work. A player who knows that tonight’s solo may be pressed, sold and compared against last year’s solo plays inside a hall of mirrors. Some players responded by treating each recording as a draft to be improved, varying their famous solos from take to take so that no single version could harden into a cage. Others responded by repeating themselves deliberately, giving buyers the solo they had paid to hear again. Neither response existed before the groove. The pre recording improviser answered to the room. The recording improviser answered to the room, to the catalogue and to earlier versions of the self, all at once, and the strain of that triple audience is audible across the century wherever a celebrated player records a signature tune for the fifth time.

There is a quieter effect as well, which is what repetition did to listening itself. A passage heard once is an event. The same passage heard twenty times becomes a landscape, with landmarks the listener anticipates and small details, a breath, a cracked note, a drummer’s accent, promoted from accident to meaning. Recordings trained listeners to prize those details, and performers, knowing the details would be caught, began placing them on purpose. The improvised aside became a composed effect wearing the clothes of spontaneity, and telling the genuine accident from the rehearsed one became one of the pleasures, and one of the honest difficulties, of listening to recorded improvisation at all.

How jazz solos changed once they were fixed

The early recorded jazz solo was short because the side was short, and it was self contained because the side had to be. When Louis Armstrong recorded with his Hot Five and Hot Seven groups in the middle and late 1920s, in sessions made for the OKeh label in Chicago, the solo became the organising event of the record rather than an interlude between vocals or ensemble passages. Armstrong built solos with an architect’s sense of pacing, opening statements that climbed in register and intensity toward a climax placed where a three minute frame could hold it, and those records travelled to musicians who had no other way to hear him. Players in distant cities learned the solos note for note from the discs, which meant that a style worked out in one studio became a common language within a few years, at a speed no touring circuit could have matched. The full career behind those sessions is traced in the series biography of Louis Armstrong and his life and career, and the wider frame sits in the pillar survey of the history of jazz and blues music.

Fixing the solo also created the problem of the definitive version. Once a solo existed on disc, later performances by the same player were heard against it, and a musician could be accused of declining for failing to reproduce, at forty, what the groove had preserved from twenty five. The accusation misunderstands what improvisation is, yet the groove gives it force, because the evidence sits in the listener’s cabinet rather than in a reviewer’s fading memory. Some celebrated solos acquired such authority that younger players reproduced them entire in tribute performances, an act that would have been impossible and meaningless before recording, since nobody can reproduce entire what nobody can hear twice.

The fixed solo changed pedagogy down to its smallest habits. Transcription, once the skill of writing down a tune heard in a room, became the discipline of slowing a record, or enduring its speed, and writing down a master’s chorus exactly. Published solo transcriptions followed, then method books built on them, and the improvised line entered the classroom wearing the uniform of the composed etude. Jazz education, for better and for worse, was born at the needle drop. The gain was a shared technical inheritance of enormous richness. The cost was a subtle pressure toward the already recorded, toward solos that behave like the solos on the canonical discs, and against the kind of playing whose value lies in never being repeatable at all.

The arrival of the microphone

What did the microphone change about singing?

The microphone let a singer work quietly and closely instead of projecting across a hall toward a recording horn. Breath, consonants and small shifts of pitch became audible, so intimacy replaced volume as a saleable skill, and engineers began placing voices forward in a balance that acoustic recording had seldom allowed before.

Electrical recording reached commercial studios in 1925, when Western Electric’s system, developed in the Bell telephone laboratories, replaced the horn with a microphone, an amplifier and an electrically driven cutter. The change sounds like a mere improvement in fidelity, and it was, but its musical consequences ran deeper than clarity. The frequency span widened at both ends, so bass instruments regained their weight and voices regained their air. More radically, the energy that cut the groove now came from a wall socket rather than from the performer’s lungs and arms, which freed the performer from the obligation to be loud. A singer could stand back, sing softly, and still fill the record. The studio stopped being a funnel that musicians shouted into and became a space whose acoustics an engineer could shape.

The microphone also changed what counted as a voice. Under the horn, the voices that recorded well were the voices built for theatres, trained tenors and sopranos and vaudeville belt, instruments of projection. Under the microphone, a small voice with fine control of words could outsell a large one, because the record played in a parlour at parlour volume, where projection reads as shouting and closeness reads as confidence. Susan Schmidt Horning, in her research on recording engineers and studio practice published across the first decade of the twenty first century, documents how engineers became mediators of this new closeness, choosing microphones, distances and rooms, and in doing so acquiring a share of authorship in sounds that credits assigned to singers alone. Singing technology as a subject in its own right, from the ribbon microphone to later processing, belongs to the companion survey of how technology changed singing, and the present article keeps to what the microphone did to the act of singing itself.

The crooner’s art also depended on a new kind of accompaniment. Because the voice no longer needed to carry over an orchestra by force, arrangers could write softly around it, strings and muted brass and rhythm sections playing at conversational level, and the record as a whole acquired a dynamic scale matched to the room it would be heard in. Engineers learned to place the voice slightly forward of the band, a perspective no concert seat provides, so the listener occupied a position inside the performance rather than in front of it. That impossible seat, closer than the front row and steadier than any head in a hall, became the default viewpoint of recorded popular music, and singers ever since have sung for a listener who is, acoustically, nearer than a lover and more attentive than any audience, a fiction the microphone maintains and the listener gratefully accepts.

Crooning as a technologically enabled technique

Crooning is the name the industry gave to microphone singing, and the name began as an insult. The crooner sang quietly, closely and conversationally, shaping words the way a person speaks to one other person rather than the way an actor declaims to a gallery. Rudy Vallee, who began recording in the 1920s, at first used a megaphone to reach the back of dance halls, a relic of the acoustic world, and then discovered through radio and records that the small voice was the product. Bing Crosby completed the transition in the early 1930s, singing with a relaxed baritone ease that depended entirely on amplification: without the microphone, his manner would have reached the third row and stopped. With it, he could sound as though he were singing in the listener’s own room, which, through the loudspeaker, he was.

The technique drew moral criticism that now reads as a measure of how much had shifted. Crooning was attacked in the early 1930s by clergy, editorial writers and at least one prominent cardinal as effeminate, decadent and a threat to vigorous manhood, because it separated singing from visible physical power. The attacks conceded the essential point. A technology had made a new kind of masculinity audible and saleable, one based on nearness rather than force, and the market ratified it in sales that the belters could not match. Women singers gained the same freedom in the same years, trading projection for inflection, and the microphone era’s defining vocal art, in jazz, in popular song and later in soul and rock, descends from that exchange of volume for detail.

Crooning also rewired the relation between singer and song. The theatre singer interpreted a song for a room that heard it once. The crooner interpreted it for a record that would be heard in private, repeatedly, often by one person alone, and the interpretation adjusted toward the confessional: slower tempos, rubato that followed speech, and lyrics addressed in the second person to a listener who could plausibly feel singled out. Katz counts this manufactured intimacy among the central phonograph effects, and it is worth separating from mere softness. Plenty of quiet singing predates the microphone. What the microphone added was distribution, the same intimate performance delivered identically into a million rooms, intimacy manufactured at industrial scale without losing its air of privacy.

Electrical recording

The electrical era, from 1925 into the years after the Second World War, widened everything the acoustic era had narrowed. Orchestras could record with their real string sections restored, because microphones heard what horns had missed, and the symphonic repertory entered the catalogue in something closer to its concert balance. Dance bands gained a floor of bass and a ceiling of brass that made records swing harder. The blues and country musicians recorded in the great location sessions of the late 1920s, when the Victor and Columbia labels sent portable electrical rigs to Memphis, Atlanta, Dallas and Bristol, were captured with a fullness that acoustic field trips could not have given them, and the 1927 Bristol sessions in Tennessee, which recorded the Carter Family and Jimmie Rodgers in the same week, stand as the clearest case of a technology arriving in a place at the moment a repertory was ready for it.

Electrical recording also professionalised the studio. The recording manager of the acoustic era, who arranged bodies around a horn, gave way to the engineer at a control desk, adjusting gain, riding levels during a take and judging a performance through loudspeakers in another room. Performance and judgment separated by a wall and a pane of glass, which is the architectural fact of the modern studio: musicians on one side producing sound, listeners on the other side evaluating its recorded shadow. Schmidt Horning’s work on the engineering profession traces how these men, and they were almost entirely men in the electrical era, built a craft culture with its own standards of good sound, standards that musicians had to satisfy to get a record made at all.

The wider span had a cost that took years to name. Acoustic records, for all their thinness, presented an ensemble in one acoustic space, blended by the air of one room. Electrical recording permitted separation, and separation invited control, and control invited the thought that a performance could be assembled rather than given. That thought needed tape and multitracking to reach its full form, but its root is here, in the moment the microphone let each instrument be captured on its own terms and balanced by a hand that was not playing any of them.

The producer’s rise also redistributed blame and credit in ways the industry argued about for the rest of the century. A failed record could be charged to the producer’s choices of song, key, tempo and take, and a successful one raised the question of whose success it was, a question with money inside it once producers began claiming royalty points alongside fees. Singers who wrote nothing and played nothing on their own records pressed the question hardest from the other side, since the voice on the disc was indisputably theirs while nearly every other decision had been taken around it. The mature industry answered with contracts, credits and custom rather than with principle, which is why producer credits vary so wildly in fullness across labels and decades, and why the sessionographies compiled by collectors, reconstructing who decided what in which studio on which afternoon, read like attempts to restore an authorship that the commercial record was designed, in part, to blur.

The record producer appears

Once the studio acquired technology worth directing, a new figure appeared to direct it: the producer. In the acoustic and early electrical eras the role hid under other titles, the recording manager, the artist and repertory man, the engineer, the label owner who chose the songs. Fred Gaisberg, who recorded for the Gramophone Company across Europe and Asia from the late 1890s onward, scouted Enrico Caruso in Milan in 1902 and paid a fee for one afternoon’s session that the company considered reckless, ten sides at a hundred pounds, and the resulting records sold in numbers that justified the recklessness and taught the industry that choosing the singer was itself a creative act. Gaisberg combined talent scouting, session planning, accompaniment at the piano and commercial judgment in one person, which is the producer’s job description before the job had a name.

The producer’s power grew with the technology. When a record was a single take of a fixed ensemble, the producer chose the artist and the song and then witnessed the result. When tape allowed editing and multitracking allowed assembly, the producer chose takes, balances, orders and effects, and the finished record bore the producer’s decisions in every bar. Figures such as George Martin at Abbey Road in the 1960s, working with the Beatles from 1962 onward, and Sam Phillips at Sun Studio in Memphis in the 1950s, who recognised and framed the sound of Elvis Presley, stand at opposite ends of the producer’s range: Martin the trained musician arranging and engineering possibility, Phillips the facilitator creating conditions and recognising the take. Both cases show production as authorship shared with the performer, a sharing that contracts and credits have argued about ever since. The dedicated history of the building and the desk belongs to the survey of the rise of the recording studio, while the present article treats the producer as evidence for its thesis: recording created musical roles that live performance had never needed, because the object, not the evening, had become the work.

Producers also became the industry’s memory. A singer might make one record a year; a producer made dozens, across artists and styles, and carried techniques from session to session, the echo that suited one voice becoming the signature tried on the next. Cross pollination that once required musicians to share a stage or a town now happened inside one person’s working week. That quiet brokerage explains a good deal of how recorded styles hybridised so quickly in the middle of the twentieth century, faster than live circuits alone can account for, because the studio concentrated encounters that geography would have kept apart.

Magnetic tape and editing

Magnetic tape reached Allied hands as a German wartime technology, the Magnetophon, refined in Germany during the 1930s and used for broadcast recording during the Second World War. After the war, American engineer Jack Mullin brought machines and tape back to the United States, demonstrated them in 1946, and the Ampex company built commercial recorders from 1948 onward, with Bing Crosby among the early backers because tape let him record his radio programmes in advance and edit out the mistakes. The musical significance of tape was not fidelity, though tape sounded fuller than disc cutting. The significance was the razor blade. Tape could be cut with scissors and spliced with adhesive, which meant that a performance could be assembled from pieces recorded at different moments, and no listener could hear the join.

Editing rewrote the meaning of a recorded performance at a stroke. Under disc recording, a side was evidence: flawed, perhaps, but continuous, the trace of one stretch of real time. Under tape, a master was a construction, the best two bars of one take joined to the best eight of another, and the splice marks existed only on the physical tape, invisible in the sound. Classical recording felt the change first and deepest, because classical performance had staked its honour on continuity. Pianist Glenn Gould, who recorded the Goldberg Variations in 1955 and again in 1981 and who withdrew from concert performance in 1964 to work only in studios, defended the splice openly, arguing that the listener’s ideal performance mattered more than the fiction of one unbroken attempt. Critics who prized the concert as the real event heard the same splice as a kind of dishonesty. Both sides were describing the same technical fact, that tape had separated the recorded work from any single act of playing it.

Singers and bands absorbed the lesson at different speeds. The pop session of the 1950s still aimed at a complete take, with editing reserved for rescue. By the 1960s, editing was composition: songs were built, introductions were flown in from other takes, and solos were compiled. Milner treats the splice as the hinge of recording history, the point where the industry’s founding promise, a faithful copy of an event, quietly inverted into its opposite, an event that never occurred until the tape made it. Whether that inversion counts as liberation or as loss depends on what a listener believes a record is for, and the argument had, by 2010, been running for sixty years without exhausting itself.

Stereo gave multitracking a second gift, a stage inside the record. Two channel stereo, commercialised on disc from 1957 and 1958 onward, let engineers place instruments across an imaginary width, and the mixing desk became a kind of architecture in which a band could be arranged more legibly than any room arranged it. Producers of the 1960s experimented with separation as meaning, voices answering across the stereo field, drums built as structures rather than heard as kits, and listeners with headphones, a constituency that grew through the 1970s, received a private version of the mix that loudspeaker listeners only partly shared. The point for this article’s argument is that placement joined performance as a compositional decision. Where a sound sits in the recorded space, how near it feels, how much air surrounds it, these are musical facts in the age of the desk, and they belong to the record alone, since no concert can guarantee them to every seat.

Multitracking

What is multitracking?

Multitracking records separate parts of a performance onto separate tracks of one tape, so each part can be recorded at a different time, balanced afterward and corrected without redoing the whole performance. A singer might add harmony to a finished backing track days later in another studio across town entirely.

Les Paul built the practical path to multitracking out of a guitarist’s impatience. Working with disc machines in the late 1940s and then with tape, Paul layered guitar parts on top of one another, recording with Mary Ford’s voice stacked in harmonies that no single throat produced, and in 1954 the Ampex company built him an eight track recorder using selective synchronisation, which let him record on one track while listening to the others in time. The commercial industry followed within a decade: two track machines for stereo, then four track, eight track, sixteen and twenty four track machines through the 1960s and 1970s, each expansion weakening the requirement that musicians be present together, or present at all in any single moment, for a record to sound like an ensemble.

The compositional consequences were enormous. A rhythm section could lay a foundation in January and a string arrangement could be written against that foundation in March, fitted to the recorded performance with a precision no live rehearsal allows. A singer could harmonise with the self of an hour earlier. A mistake in one track cost the retaking of one part rather than the whole. Producers and arrangers gained the ability to decide balances after the performance ended, which moved a share of musical judgment from the floor of the studio to the desk, and from the day of recording to the weeks of mixing that followed it. Albums grew longer to make, more expensive, and more deliberately shaped, and the band as a group of people playing in a room became, on many records, a legal and social fact layered over a manufacturing process that no longer required the room.

Multitracking is sometimes described as the point where recording stopped being honest, a framing the stance of this article declines to adopt whole. The framing belongs to a wider argument about authenticity in recorded popular music, and it is one side of that argument: the other side answers that ensemble playing in one room is one musical value among several, that layered construction produced works no room could have produced, and that a record assembled from parts is a performance of a different kind, distributed across time rather than faked. What multitracking ended was not authenticity but simultaneity as a requirement, and the next section takes up what that ending made possible.

The end of the single performance

Once parts could be recorded separately and wholes assembled from takes, the single performance lost its status as the unit of music. A record no longer needed a corresponding event. There might be no afternoon, anywhere, on which the released song was played from beginning to end by the people named on its sleeve. The Beatles’ later studio work of 1966 and 1967 made the point famous, because the group withdrew from touring in 1966 and produced records whose sounds, tape loops, reversed passages, orchestral collages and layered vocals, could not be reproduced on a stage by four musicians at all. The record had become the primary work and the concert, where one happened, a translation of it, reversing the relation that had held since Edison: the performance the copy, the object the original.

This reversal changed how musicians thought about their own output. The live player of the acoustic era left a reputation; the studio artist of the multitrack era left a catalogue, a body of fixed works against which every later act would be measured. Careers acquired the shape of literatures, with early and late periods, masterworks and minor works, because the works persisted in a form that invited exactly that literary treatment. Critics could compare a singer’s phrasing across twenty years in an afternoon, and did, and singers answered by curating their catalogues, re recording old songs, revising arrangements and contesting the standing of their own earlier objects.

It also changed the meaning of a mistake. In a single performance culture, the wrong note is part of the event and dies with it. In an assembled culture, the wrong note that survives to release is a decision, somebody heard it and let it stand, and listeners learned to read such survivals as signatures of intent, roughness kept as proof of life inside a polished construction. Producers of the 1970s and 1980s spoke of leaving the human element in, a phrase that concedes the whole argument: humanity, in the studio, had become an ingredient to be dosed rather than the medium in which music happened by default.

The album’s authority can be measured by the behaviour it trained. Listeners learned to play albums whole, in order, as a condition of respect for the work, and to judge artists who sequenced carelessly the way readers judge a book with its chapters shuffled. Record shops organised browsing by the object, sleeves face out like book covers, and the sleeve carried the work’s public face into homes and bedrooms where it functioned as furniture, poster and credential at once. Milner notes how completely the industry’s economics came to rest on the album unit in the decades after 1948: advances, royalties, chart positions and career narratives all counted in albums, so the format’s artistic prestige and its commercial convenience reinforced each other until neither musicians nor listeners could easily imagine the song as the primary unit again. The digital file would eventually break that settlement, which is one reason the break felt, to the album generation, like a change in the nature of music rather than a change in its packaging.

The album as a designed object

The long playing record, introduced by Columbia in 1948, spun at thirty three and a third revolutions and held roughly twenty two minutes of music on each side, which multiplied the side length of the 78 era by about seven. The seven inch forty five, introduced by RCA in 1949, took over the single song market, and between the two formats the industry acquired a grammar of lengths: the single for the song, the album for the statement. At first the album collected singles the way its disc era ancestor had collected sides. Through the 1950s, jazz musicians and classical labels exploited the span for extended works, and singers such as Frank Sinatra sequenced albums as mood programmes, songs chosen and ordered to sustain one emotional weather across both sides.

In the 1960s the rock album claimed the span as an artistic unit. Sequencing became composition at one remove: the opening track as threshold, the side break as an interval with its own dramaturgy, the closing track as the note the listener would carry out of the room. Sleeve design, gatefolds, printed lyrics on inner sleeves in defiance of no rule but custom, and studio credits all enlarged the object into something handled, read and shelved as well as heard. The compact disc, introduced in 1982 with a playing time of about seventy four minutes, stretched the unit again and abolished the side break, and albums swelled toward the new ceiling the way songs had once shrunk toward the old one, a symmetry that supports this article’s central claim better than any manifesto: musicians compose toward the capacity of the object that will carry them, and call the result natural for exactly as long as the object lasts.

The designed album also created the designed career artefact, the record as the unit by which a musician would be remembered and ranked. Lists of great albums, reissue programmes, anniversary editions and boxed sets all presuppose the object. So does the peculiar grief of the album era listener when a favourite work is remixed or resequenced, the sense that a fixed thing has been tampered with, which is the grief of a reader handed a revised edition of a memorised book. That grief is a recording invention. No audience of 1850 could feel it, because no performance of 1850 stayed still long enough to be revised.

There is a discipline hidden inside the studio as instrument idea that deserves a sentence of its own. An instrument must be playable, which means its results must be repeatable enough to practise and surprising enough to reward practice, and the great studios met both tests. Engineers learned their rooms and desks the way organists learn a particular organ, its registrations, its stubbornnesses, its weather, and producers booked particular engineers with particular rooms as deliberately as they booked singers. When digital workstations standardised the tools in the 1990s, something of that local knowledge dissolved, since the same software sat in every studio on earth, and partisans of the older order mourned the rooms the way an earlier generation had mourned the regional orchestras that recording itself had levelled, a recursion the history of this technology produces at every stage, each standardisation lamented in the vocabulary of the standardisation before it.

The studio as an instrument

By the 1960s the leading studios had ceased to be neutral rooms with machines in them. Engineers and producers treated echo chambers, tape delay, compression, equalisation and the mixing desk itself as musical resources, voices in the arrangement that happened to be operated rather than played. Phil Spector’s productions of the early 1960s stacked orchestras and rhythm sections inside echo until the ensemble fused into a single surging timbre. Jamaican producers and engineers, above all King Tubby in the 1970s, took the mixing desk further, stripping finished songs down to drum and bass and rebuilding them in performance at the controls, which produced dub, a genre made out of the studio itself. In Germany in the 1970s, Conny Plank engineered records for groups such as Kraftwerk and Can in which electronic texture was the substance rather than the decoration. Each of these cases satisfies the test of the studio as instrument in the strong sense: remove the studio work and the music does not become plainer, it ceases to exist.

Treating the studio as an instrument changed who counted as a musician. The engineer at the desk, choosing a microphone and riding a fader in time with the music, performs, and the credits and royalty structures of the industry spent decades catching up with that fact, unevenly and often unfairly. It also changed what skill meant for the credited artists. Timing a vocal entry against a tape echo, singing into compression so the machine’s squeeze becomes part of the phrasing, playing a guitar part designed to be reversed or slowed: these are instrumental techniques whose instrument sits partly outside the player’s hands. Listeners trained by records learned to hear them as musicianship, which they are, while concert trained listeners sometimes heard them as trickery, which mistakes the location of the craft for its absence.

The studio instrument had a geography. Certain rooms acquired sounds that records carried around the world: the echo of particular chambers in Los Angeles and London, the dry tightness of specific Memphis and Muscle Shoals rooms, the spaciousness that Jamaican studios achieved with limited equipment and unlimited ingenuity. Musicians travelled to rooms the way earlier generations travelled to teachers, to be recorded inside a sound. A room cannot tour and cannot be interviewed, yet its contribution is on every side it ever shaped, which makes the studio the clearest example of recording’s habit of turning conditions into collaborators.

Language rode in the grooves as well, with consequences that outlasted the sessions. A recorded singer fixed a pronunciation, a dialect and a poetic diction as the audible standard of a language for listeners who might never hear it spoken at length anywhere else, and language revival movements of the later twentieth century mined early discs for exactly that evidence. Songs also crossed communal boundaries that live performance had policed. Material sung by one caste, clan, gender or age grade, in settings closed to outsiders, became purchasable by anyone with the price of a disc, which some communities experienced as theft of the sacred and others as overdue recognition of artistry the closed setting had hidden. The record did not ask which reading was correct. It simply made the boundary permeable in one direction, outward, permanently, and left each community to renegotiate its rules around a leak that could not be plugged.

What recording did to traditional musics

When recording reached musics that lived by oral transmission, it offered preservation and imposed a freeze in the same gesture. A tradition that had survived by changing a little at every performance now acquired reference versions, pressed in shellac or vinyl, against which change registered as drift or betrayal. The reference version was rarely chosen by the tradition itself. It was chosen by whoever held the machine: the label scout deciding which singer represented a region, the visiting collector deciding which ceremony, which verse, which take. One afternoon’s performance, by one practitioner, on one day, in one mood, became the sound of a culture for listeners on other continents, and often, through reissue and radio, for the culture’s own grandchildren.

The effects ran in both directions at once, and honest writing about them has to hold both. Records carried endangered repertories across distances that migration had scattered, reconnected diaspora communities with the sounds of places their grandparents had left, and gave young musicians access to masters they could never have studied in person. Indian classical music, to take one large case, was reshaped by the gramophone from 1902 onward, when the Gramophone Company recorded singers in Calcutta: short disc sides compressed performances that had unfolded across an hour in concert, hereditary musicians reached audiences beyond their patrons, and women singers of the courtesan traditions, including Gauhar Jan, who recorded in 1902 and announced her name at the end of sides because the company required identification, became public recording artists at a moment when respectability politics pressed them toward silence. The disc cut the raga down and carried the singer out, and both halves of that sentence stayed true for the century that followed.

Recording also standardised. Where ten villages held ten variants of a tune, the recorded variant had a distribution network and the others had a locality, so the recorded variant became the tune and the others became deviations, in the ears of outsiders and, over time, in the ears of insiders taught by records. Instruments changed to suit the horn and then the microphone, repertories tilted toward what recorded well and what sold in three minute units, and ceremonial musics with durations set by ritual were excerpted into concert lengths. None of this required malice. It required only a market, a machine with preferences, and time, which is why the same pattern appears on every continent the industry reached, wearing local clothes.

Field recording and its ethics

Field recording began as scholarship and commerce at once, and the mixture has never separated. Jesse Walter Fewkes recorded Passamaquoddy singers and storytellers in Maine in 1890 on wax cylinders, among the earliest uses of the phonograph outside the laboratory, in the belief that the cultures he studied were vanishing and that the cylinder could hold what the future would otherwise lose. Ethnographers followed across every continent, building archives of cylinders and discs in Berlin, Vienna, Paris and Washington. In parallel, commercial labels sent scouts through the American South, the Caribbean, West and East Africa, South Asia and the Pacific, recording local musicians for local markets, because a community would buy records of its own music and the machinery to make them had become portable enough to travel.

The ethical difficulty sits in the encounter itself. A collector arrives with a machine that the recorded community does not own, records performances governed by protocols the collector may not understand, and leaves with objects whose value will be realised elsewhere, in archive prestige, in sales, in scholarly careers. The recorded singers of the cylinder era rarely controlled copies, could seldom hear how their recordings were used, and in some traditions had sung material that was restricted, seasonal or sacred, material whose unrestricted replay in another country violated the terms on which it was given. Alan Lomax, the most energetic field recordist of the middle twentieth century, whose expeditions for the Library of Congress and later labels carried him across the United States, the Caribbean, Britain, Ireland, Spain and Italy, argued across his career that recording dignified musics the concert world ignored and that royalties and credits should flow back to source communities. His critics answered that energy is not consent, that dignity is not payment, and that the archive’s doors opened more easily to scholars than to the families of the recorded. Both positions describe real features of the same expeditions, which is why the argument survives its participants.

The stance this article takes is structural. What was legal, what was customary and what was fair were three different standards in the field recording era, and collapsing them serves nobody. Much early collecting was legal under the law of the place and time. Much of it was customary within the young recording industry, which treated a small session fee as the whole transaction everywhere it went. Whether it was fair depends on questions the participants rarely asked on the record: who could refuse, who understood the machine’s reach, who shared in the object’s later value, and whose rules governed sacred or restricted material once it sat on a shelf abroad. Later archival practice, repatriation of copies to source communities, and shared rights agreements all grew out of those unasked questions, and the growth itself concedes that the questions were real.

Consent, credit and royalties for recorded communities

Who owns a field recording?

Ownership depends on the contract, the law of the place of recording and who fixed the sounds on tape or disc. A collector or label often held legal title while performers and communities held moral claims to credit, control and payment that contracts of the period frequently failed to recognise fully.

The practical history divides into three patterns. In the first, musicians were paid a flat session fee and signed, or never signed, papers that assigned all rights to the label, a pattern standard in the commercial location sessions of the 1920s and 1930s in the United States and in the early recording industries of India, Egypt and West Africa. In the second, scholarly collectors recorded without commercial release in view, deposited cylinders in institutions, and understood themselves as custodians, a pattern that avoided the session fee’s finality but placed control in an archive rather than with the recorded. In the third, which belongs mainly to the later twentieth century, collectors and labels negotiated credits, royalties and community rights in advance, sometimes under the guidance of ethnomusicological codes of ethics that the discipline adopted as its field practice came under scrutiny.

Credit is the cheapest of the three goods and was the most often withheld. Early discs frequently named a region, a tribe or a genre where they should have named a singer, and some collectors presented themselves as the authors of material they had only carried. Where names survive on labels and in ledgers, later reissues could restore them, and the restoration programmes of archival labels from the 1960s onward treated naming as a form of repayment. Royalties are harder, because a royalty requires an accounting chain, sales records, identifiable rights holders and a mechanism for paying people who may be dispersed, deceased or collectively rather than individually entitled under their own law. Defenders of the old collectors point to exactly this machinery problem, and to the fact that many recordings sold in tiny numbers: the fortune extracted from any single village session was usually small, and the larger fortune, where one existed, belonged to the catalogue as a whole.

The critique answers that small sums do not settle the principle, that consent given without an understanding of replay across oceans and generations was consent to a different act than the one performed, and that control matters independently of money, above all for restricted material. For the defence to be right in full, it would have to be true that recorded communities understood the machine, could refuse without cost, and valued the preservation offered above the control surrendered. For the critique to be right in full, it would have to be true that collectors understood the reach of what they took and that the communities gained nothing they wanted from the exchange. The documentary record supports each of those conditions in some cases and refutes each in others, which is why honest catalogues now record the circumstances of a recording alongside the recording, treating the encounter itself as part of the evidence. Listeners assembling their own notes on such histories sometimes keep session details in an offline encrypted notes app obtained from the VaultBook download page, a filing habit that at least keeps the names beside the sounds they belong to.

The record industry’s global expansion

The gramophone industry was multinational before the word was fashionable. By 1900 the Gramophone Company in London and Victor in the United States drew on linked patents and divided territories, and by 1902 and 1903 recording expeditions were at work in Calcutta, where the company’s agent Fred Gaisberg recorded local singers for an Indian market that the company intended to build by selling machines alongside the records only those machines could play. The pattern repeated across the following three decades: engineers travelled with portable cutting equipment to Cairo, Bombay, Shanghai, Buenos Aires, Lagos, Jakarta and a hundred smaller markets, recorded local repertory with local musicians, shipped the masters to pressing plants, and returned finished discs to shops in the cities where the music had been made. Historians of the early industry sometimes call the catalogue that resulted the first world music market, and the description is fair in the narrow sense that local music, industrially produced, was being sold back to its own community and outward to curious buyers elsewhere at the same time.

The expansion followed imperial and commercial routes, cables, shipping lanes and colonial administrations, and that parentage marked the catalogue. What got recorded in a colony reflected what a visiting engineer, working through local fixers and theatre managers, could assemble in a fortnight, which favoured urban professionals over rural ceremonial musicians, short forms over long ones, and the languages of ports and capitals over the languages of the interior. Yet the same machinery handed local musicians an instrument they quickly turned to their own purposes. Egyptian singers built record careers that radio later amplified. Indian film music grew on a foundation of disc sales. Argentine tango, Cuban dance music, Hawaiian guitar styles and West African highlife all used the record as an export engine, sending styles abroad in grooves and receiving foreign styles back by the same channel, so that by the 1930s a musician in Havana, Hamburg or Hanoi could hear, and answer, music made on other continents within a year of its making. The machinery of that two way trade, the labels, scouts and licensing deals, is set out in the survey of how the record label system worked, and its broadcast cousin is traced in the account of how radio changed popular music.

Price decided a good deal of what travelled. A disc was a luxury purchase in most of the markets the industry entered, and the buyers who could afford machines and records were urban, salaried and exposed to foreign goods, so the export catalogue of any local music leaned toward the tastes of that class and taught rural musicians which of their styles the town would pay for. Migrant workers formed a second market with opposite tastes, buying records of home as medicine for homesickness, and labels served both, the prestige style for the parlour and the village style for the barracks and the boarding house. Between those two demands a local music could split into a polished export line and a preserved home line, each regarding the other with suspicion, and the split, once cut into shellac, acquired the authority of a fact about the tradition rather than a fact about its market.

Local musics as export commodities

A local music becomes an export commodity when its sound can be detached from its occasion. A wedding band sells an evening; a record of the wedding band sells a token of the wedding to people who were not invited, in places the band will never visit. That detachment is the record’s gift and its distortion in one act. Styles travelled in the grooves long before musicians could afford to tour: Hawaiian steel guitar records circled the world in the 1910s and 1920s and taught guitarists on several continents a technique their teachers had never seen played. Tango records and sheet music reached Paris from Buenos Aires and returned to Argentina wearing Parisian prestige. Blues and jazz records reached Europe in the hands of soldiers and merchants after the First World War and founded schools of playing in cities their makers never entered.

Export also edited the product. A style sold abroad shed what foreigners could not follow, long narrative ballads, insider humour, ritual pacing, and kept what travelled without translation: rhythm, timbre, melodic hooks, virtuosity. Musicians learned the lesson quickly and began composing for the groove’s passport, writing the version of their music that a distant listener could receive. Critics of the trade called the result dilution, and sometimes it was. Defenders answered that every style in the export catalogue was already a hybrid, that tango, highlife, calypso and country music were all born from mixtures that trade had made, and that the record merely accelerated a blending that ports and migrations had practised for centuries. The strongest version of each case concedes something to the other: the critics are right that the market selected for exportable features and against local depth, and the defenders are right that local musicians were agents in the selection, not merely its raw material, and often profited in money and standing from the very records that simplified them for foreign ears.

One more mechanism deserves naming, because it closed a loop. Export created return influence. A style shipped abroad, transformed by foreign players, and shipped back could confront its makers as a fashion to imitate, a standard to resist, or a mirror in which the home version suddenly looked old fashioned. Jamaican musicians listening to American rhythm and blues records on distant stations, West African bands answering Cuban discs, and British groups of the 1960s returning American blues to American audiences are all instances of the same circuit. Recording did not merely spread local music. It set local musics in conversation at a distance, with the groove as interpreter, and the conversations changed every participant.

The archive itself became an actor in the story. Institutions that held cylinders and discs, the Berlin Phonogramm Archiv founded at the start of the twentieth century among the earliest, set standards for documentation, cataloguing and access that decided whose recordings would be findable by later generations and whose would sleep unlabelled in boxes. A recording without a catalogue entry is nearly as lost as a recording never made, so the archivist’s handwriting on a ledger joined the engineer’s hand on the machine as one of the silent co authors of the world’s audible past. Repatriation programmes of the later twentieth century, returning copies and documentation to source communities, acknowledged that the archive’s custody had always been a relationship rather than a fact of nature, and that relationships can be renegotiated when the weaker party finally gets a hearing.

Preservation and what was frozen in place

The preservation claim for recording is strong and partly deserved. Cylinders and discs hold performances by musicians who left no notation and no pupils, singers whose art would otherwise survive as a name in a memoir and a reputation nobody could test. Archives of early recordings let later generations hear the tempo choices of nineteenth century trained pianists, the ornamentation of singers who had learned their art before the microphone, and regional styles that radio and roads later levelled. For oral traditions disrupted by war, forced migration and language loss, a recording can be the only surviving teacher, and revival movements across the world have rebuilt repertories from discs, learning songs from the grooves their grandparents’ generation pressed.

The freeze is the same fact read from the other side. A tradition preserved on record acquires a fixed point that can inhibit the change by which traditions stay alive. Students learn the recorded master’s version as the correct version, including its accidents: the day the tempo ran fast, the verse the singer substituted under session pressure, the tuning adopted because the studio piano sat where it sat. Day’s account of classical recording documents the parallel process in concert music, where recorded performances by a few celebrated artists became the standard against which students worldwide were measured, narrowing the range of acceptable tempo, tone and rubato even as the recordings enriched everyone’s access to great playing. Preservation, in other words, preserves one state of a moving thing and then exerts authority over the movement that continues around it.

The freeze also operates on repertory. What was recorded in the brief commercial window of the 1920s and 1930s, when location recording swept through region after region, became the documented tradition, and what the scouts missed, because a musician was away, ill, unwilling or unfashionable, sank toward invisibility. Later scholars inherit the catalogue as if it were a census, and the correction of that error, seeking the unrecorded alongside the recorded, is slow work against the groove’s authority. A responsible listener treats an early recording the way a historian treats a photograph: as evidence of surpassing value about one moment, and as a trap if mistaken for the whole scene.

Repertoire narrowing in classical music

Classical recording narrowed the working repertory through a simple commercial ratchet. A label deciding what to record chose works that buyers already recognised, because recognition sold the expensive object, and each successful recording deepened the recognition that justified the next one. The symphonies of Beethoven, Brahms and Tchaikovsky accumulated dozens of competing versions across the twentieth century while symphonies by their contemporaries, performed in their lifetimes and printed in editions, waited decades for a first recording or never received one. Concert programming followed the catalogue’s lead, since audiences buy tickets for what they know, and they know what they have heard, and they have heard what was recorded. The circle is nobody’s plot. It is an incentive structure, and it tightened as recording budgets rose and the cost of a failed adventurous release fell on fewer, larger companies.

The narrowing had a stylistic twin. Recording rewarded precision, because errors replay, and rewarded the kind of brilliance that survives close listening: clean articulation, steady intonation, controlled climaxes. Orchestras around the world grew more uniform in sound across the century, a convergence that historians of performance attribute partly to musicians studying the same celebrated records and partly to recording itself selecting for blend and accuracy over local idiosyncrasy. The distinctive wind timbres of French orchestras, the string portamento of pre war Vienna, the rhythmic freedoms documented on early discs, all thinned as the recorded standard became international. Something was gained, a floor of technical command below which leading orchestras rarely fall. Something was lost, a plurality of answers to the question of how this music goes, and the early records preserve the lost answers precisely because the narrowing came later.

Countercurrents existed and matter. The long playing record lowered the cost of exploratory repertory, independent labels built identities on neglected composers, and the early music movement used recording as its loudspeaker, carrying performances on period instruments from specialist circles to a worldwide audience that no concert circuit could have assembled. The net verdict is therefore double, and the evidence supports the double verdict rather than either half alone: recording concentrated fame and money on a narrowing canon at the centre of classical life, and simultaneously gave the margins of that life, the early, the regional and the forgotten, a distribution channel they had never possessed. Which force dominates depends on the decade and the label one examines, and any account that reports only one of them is selling something.

The canonical performance problem

A canonical performance is a recording so widely taken as the reference for a work that other performances are described by their distance from it. The problem it names is not that reference recordings exist. Listeners need landmarks, and a great recording earns its authority honestly, through playing that survives rehearing. The problem begins when the landmark stops being one performance among many and becomes the work itself in the public ear, so that a conductor who chooses a different tempo is heard as deviating from the music rather than from a predecessor, and a student who plays the passage another way is corrected toward the record.

The mechanism is partly technological and partly institutional. A recording can be assigned, broadcast, excerpted and quoted in a way a concert cannot, so the canonical version is usually the version that was available, in print, in the shops, on the radio, when a generation formed its taste. Record company mergers and catalogue deletions have quietly made and unmade canons by keeping one version in circulation and letting rivals lapse, decisions taken for commercial reasons whose musical consequences lasted for decades. Gould’s two recordings of the Goldberg Variations, made twenty six years apart with strikingly different tempos and character, offer the cleanest demonstration that the canonical object is a performance and not the work, because the same musician produced two incompatible landmarks for the same score, and listeners divided into camps over which Gould was the real one, as though a work in two authorised versions were a contradiction rather than a reminder of what a score leaves open.

Jazz and popular music hold the same problem in a different key, where the canonical object is often the hit take, and the artist who made it must perform against the record’s memory nightly. The healthiest musical cultures in the recorded century developed a counter habit: treating the canonical recording as a text to argue with, covering it, answering it, sampling its authority in order to borrow or break it. Argument keeps a canon alive as a provocation. Reverence alone embalms it, and the difference between the two treatments is one of the most reliable measures of whether a tradition is still moving.

Bootlegs and the unauthorised record

The bootleg is the recorded century’s shadow catalogue: performances captured or copied without the rights holder’s permission, circulated because the authorised catalogue would not supply them. Its history begins early. Lionel Mapleson, librarian of the Metropolitan Opera in New York, recorded excerpts of live performances from the house’s proscenium and catwalks between 1901 and 1903 on cylinders never meant for sale, and those fragile cylinders preserve voices of the era’s operatic stage that the commercial studios, which preferred their own conditions, captured thinly or not at all. In the 1950s, collectors pressed historic broadcast material onto small label discs. In 1969 the rock bootleg acquired its modern form when an album of unreleased Bob Dylan recordings, packaged in a plain sleeve, appeared in American shops and sold in numbers that proved a market existed for what the artist and label had chosen to withhold.

Bootlegs matter to this article’s argument because they expose what the official object leaves out. The authorised record presents the performance the artist approved; the bootleg economy exists on the difference between that approval and the audience’s curiosity, the live night when the song went somewhere else, the outtake where the singer tried the braver reading, the concert that no studio session reproduced. Artists have divided over the traffic. Some prosecuted it as theft of both property and artistic control, which in law and in fairness it can be. Others tolerated audience recording as the price of a devoted following, and a few, the Grateful Dead most famously, formalised tolerance by designating taping areas at concerts, converting an illicit practice into a sanctioned archive that spread their live reputation further than their studio catalogue did.

The historian’s position is compromised in an instructive way. A large share of what is known about live performance practice in the twentieth century survives only through unauthorised capture: broadcasts taped off the air, house tapes, audience tapes, the Mapleson cylinders themselves. Condemning the method while using the evidence is a discomfort the literature carries openly. What can be said without contradiction is that the bootleg is a phonograph effect in its own right, created by the gap between the repeatability recording made possible and the scarcity the industry maintained, and that every change in copying technology, from the cassette to the recordable compact disc to file sharing after 1999, widened that gap and tested who controlled the object once it left the studio.

The solitary listener also became a performer of a modest kind. Choosing a record, cleaning it, lowering the needle at the right band, turning the disc at the side break, these small ceremonies gave listening a craft of its own, and the compact disc and the file each retired parts of the ceremony while inventing others, the playlist and the shuffle standing where the side break and the shelf once stood. Attention studies of the gramophone era worried that machine listening produced a distracted consumer; the diaries and letters of collectors from the same era show the opposite figure as well, the listener as scholar of the self’s own collection, keeping catalogues, annotating sleeves and corresponding with strangers about pressings and takes. Both figures are real, and both are phonograph effects: the machine made listening private enough to become either a hobby of great intensity or a habit of great lightness, and most listeners, across a lifetime, have been both figures in turn.

The listener alone with a record

Recording privatised listening. For most of human history, music was a social act performed in company, in worship, in labour, in dance, in ceremony, and hearing it meant being somewhere with other people. The phonograph moved music into the parlour, the gramophone into the home’s best room, the transistor era’s players into the bedroom, and the personal stereo of 1979, the Sony Walkman and its successors, into the listener’s head in public space, where music became a private weather carried through shared streets. Each step loosened the bond between hearing music and gathering for it, and each step was celebrated as access and mourned as isolation, often by the same writers in the same decade.

The solitary listener acquired powers no audience member had held. Repetition, already discussed for the student of solos, belongs equally to the private devotee, who could live inside one record for a season, memorising it the way earlier generations memorised poems, until the work became part of the furniture of a self. Selection passed to the listener as well: the home collection is a canon of one household’s choosing, assembled against the programmes of concert societies and radio stations, and the arguments conducted across that collection, what to keep, what to lend, what to play for a guest, formed a kind of domestic criticism with real consequences for what the industry recorded next, because labels read sales the way politicians read weather.

Solitude also changed the emotional contract. A record can be played at grief, at insomnia, at the hour when no ensemble could be summoned, and a voice can accompany a listener through years of private life with an intimacy no living acquaintance matches, the singer ageing not at all while the listener ages entirely. Writers on listening have worried over this asymmetry since the 1920s, when critics first charged the gramophone with replacing music making in the home, the piano abandoned for the machine that plays itself. The charge contains a truth, amateur performance did decline as professional reproduction rose, and a blindness, since the listeners were not passive in the way the charge required: they were studying, comparing, collecting and, in numbers the critics missed, picking up instruments to imitate what the records taught them.

Background listening and what it changed

The same machines that enabled devoted solitary listening also enabled its opposite, music attended to barely at all. The phonograph in the cafe, the radio in the workshop, the wired music services that Muzak built from the 1930s onward, piping programmed sequences into factories, offices and lifts: recorded music became the first music in history that could be supplied continuously, at controlled volume, without musicians present, and therefore the first music designed, in some of its branches, not to be listened to. Muzak’s programmers arranged material in fifteen minute blocks of rising intensity, a system the company called stimulus progression, matched to the working day’s fatigue curve, which is composition inverted, music engineered around the listener’s inattention.

Background supply changed foreground music by competition and by habit. Restaurants, shops and public spaces acquired soundtracks as a matter of course, and silence in public became rare enough to feel pointed. Listeners habituated to a constant floor of music brought different expectations to deliberate listening: less tolerance for silence inside a work, for slow openings, for the dynamics of a concert hall where quiet passages are genuinely quiet. Composers and producers working for the foreground market adapted, consciously or not, raising average levels, front loading interest and treating the opening seconds as an audition against the listener’s wandering attention, a pressure that connects the cafe speaker of the 1930s to the mastering choices examined later in this article.

A balanced account has to give the background its due. Programmed music in factories was studied, promoted and sometimes genuinely welcomed by the workers it accompanied, and the democratisation of access that the critics lamented was real: a listener of modest means in 1935 could hear, across a week of incidental exposure, more fine performance than a wealthy listener of 1835 heard in a season of concerts. The question recording poses here is not whether ubiquitous music is good or bad in bulk. It is what happens to a culture’s attention when its most powerful musical technology serves two opposite purposes at once, the record as an object of devotion and the record as wallpaper, and produces both from the same groove.

The economics of attention shifted with the carrier change as well. The compact disc’s seventy four minutes arrived without a side break to organise listening, and its remote controls and programmable players invited skipping, so the album’s sequence lost the physical enforcement the groove had given it. Producers answered by sequencing for a listener with a skip button, placing strong material early and treating the closing tracks as the album’s most expendable real estate, an arrangement the vinyl era, which saved a climax for the side’s end, would have found perverse. None of these adaptations appears in any manifesto. They are the phonograph effect operating at the level of the track listing, the carrier quietly rewriting the rhetoric of the works it carries, exactly as the three minute side had rewritten the rhetoric of the song three generations earlier.

Analogue to digital

The compact disc, launched commercially in 1982 by Philips and Sony, replaced the groove with a spiral of microscopic pits read by a laser, and sound with numbers: the audio waveform sampled forty four thousand one hundred times each second and stored as sixteen bit values. The change of carrier changed the copy. An analogue copy loses a little of the original at every generation, hiss rising and detail softening, which is why the industry could sell fidelity as a ladder. A digital copy, made correctly, loses nothing, which collapsed the ladder and, with it, a pricing structure the industry had enjoyed for a century. When the MP3 format, standardised in the early 1990s, shrank audio files by discarding information the ear is least likely to miss, and the Napster service of 1999 showed how quickly such files could move between strangers, the recorded object lost the scarcity on which its economics rested. That story belongs to the economics section of this article; the musical consequences belong here.

Digital recording changed studio practice as thoroughly as tape had. Editing, which tape had made possible with a blade, became effortless and reversible on screen in the digital audio workstations that spread through studios in the 1990s, with Pro Tools, launched in 1991, the best known of them. Performances could be corrected in pitch and nudged in time at the level of the individual note. The splice that Gould defended as a route to the ideal performance became invisible, undoable and therefore routine, and the argument about it changed character: the question was no longer whether a record may be assembled, but whether any audible trace of the unassembled performance need survive at all. Some producers answered by fetishising the single take and the analogue desk, others by treating correction as one more instrument, and the disagreement is best reported as unsettled, because the evidence of the records themselves cuts both ways, polish and sterility arriving, as they always had, from the same tools in different hands.

Analogue and digital listening divided audiences in the same unsettled pattern. Vinyl retained and then regained a devoted market through the 1990s and 2000s on arguments about warmth, continuity of waveform and the ritual of the side, while measurements showed the compact disc superior in noise, distortion and dynamic range, and double blind listening tests repeatedly shrank the differences listeners reported under sighted conditions. Katz’s treatment of format loyalties in Capturing Sound gives the most useful frame: listeners defend not only a sound but a whole practice of handling, collecting and attending that grew around a carrier, so a format argument is partly an argument about how music ought to be approached, prosecuted in the vocabulary of frequencies. Neither side is dismissed here. The groove and the file each made some kinds of listening easy and others nearly impossible, and the music composed for each bears the marks.

Loudness and mastering

Mastering, the final preparation of a recording for release, acquired its modern importance when records began competing in environments where level is advantage: the jukebox, the radio playlist, the shop floor, the car. A louder record seizes attention beside a quieter one, and from the 1950s onward mastering engineers learned to raise average level with compression and limiting, squeezing the distance between a recording’s quietest and loudest moments so that more of it sat near the maximum. Within limits, the practice served listeners in noisy surroundings. Pushed across decades of competitive releases, it produced the escalation that listeners and engineers of the 1990s and 2000s named the loudness war, in which compact discs arrived with waveforms flattened toward a constant maximum and quiet passages engineered out of music whose expression depended on them.

Hearing an overcompressed master requires no equipment beyond attention, and the skill generalises. Crescendos that fail to grow, choruses no louder than verses, a singer whose whisper and shout arrive at the same weight, drums whose impact feels dense rather than sudden, and a fatigue after ten minutes that the same music in an earlier master does not produce: these are the signatures. The comparison is easiest where a work exists in an early master and a later remastered edition, played at matched volume, because loudness itself flatters, and the louder version wins any casual test regardless of its dynamics. Matching the level before judging is the single discipline that separates hearing a master from being sold one.

The loudness escalation is a phonograph effect wearing a fader. No composer of concert music wrote those flattened dynamics; they accreted from thousands of competitive decisions inside a distribution system that rewarded level, each defensible, the sum corrosive. Its history also carries the counterweight: broadcasters adopted level normalisation in various markets, audiophile labels marketed dynamic range as a virtue, and musicians with sufficient standing refused masters they disliked. What the episode demonstrates, for this article’s purposes, is that the recorded object is never finished by the performance, and rarely by the mix. The last hands on the music belong to a stage most listeners never think about, and the shape of their hearing, what loud and quiet can mean in a song, was decided there. Readers preparing comparison notes or session slides on such questions can open the files in a browser-based presentation viewer through ReportMedic, which is the humblest possible illustration of the section’s point: how a work is presented is part of what the work becomes.

What the recorded object cannot capture

For all its reach, the record is a reduction, and naming the losses keeps the gains honest. A recording captures sound at the expense of the event: the sight of the players, the effort in a singer’s body, the room’s answering resonance felt in the chest, the risk shared between performers and audience in a single unrepeatable stretch of time. It captures one performance, or an assembly wearing a performance’s clothes, at the expense of the work’s other possible lives, and it fixes interpretation at the expense of the variation by which music in oral and concert traditions stays responsive to place, occasion and audience. Directions in a score that call for improvisation, for ornament, for the performer to complete the work, meet a medium that can hold only one completion, and holds it with an authority the composer never granted it.

Social music suffers a particular loss in the groove. Dance music recorded is dance music with the dancers abstracted away, their tempo pressures and floor answering energy removed, so that the record of a dance band is a portrait of a conversation with one speaker missing. Ceremonial music loses its function and keeps its sound, becoming a specimen of the thing it was rather than the thing itself. Even the concert, the setting recording serves best, loses the economics of attention that make live performance an act of communal concentration: nobody coughs on a record, nobody’s neighbour whispers, and nobody has travelled through rain to be there, and those frictions, however annoying, are part of how a live audience manufactures the silence in which quiet music can exist.

The honest conclusion is not that recording falsifies music, a claim the preceding sections have given no support, but that it translates music into a new medium with its own native excellences and its own characteristic blindnesses, the way notation translated it before. Music survived notation by becoming partly a literature. It survived recording by becoming partly an archive of objects. In both translations the survivors who flourished were the practices that learned the new medium’s grammar without surrendering the old one’s virtues, and the losses fell, as losses usually do, on the practices with the least power to negotiate the terms.

Live performance after recording

How did recording affect live performance?

Recording gave audiences a fixed version to compare against the concert platform, so listeners arrived knowing tempos, solos and balances by heart. Performers answered by matching the record, extending it deliberately, or treating the stage as the place where risk and variation survived after the studio version had settled permanently.

The concert after recording is a different institution from the concert before it, because the audience arrives pre instructed. Listeners who once met a symphony first in the hall now meet it first on record, repeatedly, in a favourite performance, and hear the live event as a reading of a known text. Performers feel the pressure in both directions. A celebrated recording artist who departs from the record risks disappointing the listeners who paid to hear the object made flesh; an artist who reproduces the record exactly risks the charge of redundancy, since the object already exists and plays at home without ticket or travel. The durable solutions divide into three broad kinds. Some performers match the record and sell fidelity. Some treat the stage as the record’s alternative, stretching tempos, reordering sets and improvising against the fixed version’s grain. Some, above all in jazz, rock and traditional music, rebuild the relation entirely, letting live and recorded work diverge until each is judged on separate terms.

Recording also raised the technical floor of live performance in ways so complete they are hard to see. Amplification, born in the studio and the broadcast chain, remade the concert: singers who had built microphone technique on records carried it to stages fitted with the same tools, and the intimate voice that could not have filled a hall in 1910 filled stadiums by 1970, not by growing louder but by being carried. Musicianship itself absorbed the record’s standards, since players who study canonical recordings arrive with intonation, precision and repertory that earlier generations assembled slowly through local teaching. Audiences, in turn, grew less forgiving of the ragged edge that live events once carried as a matter of course, and the wrong note in a concert acquired a new sting, measured against the corrected objects in the listener’s pocket and shelf.

The deepest change may be the simplest. Before recording, the live event was music’s only form and its memory was the audience’s. After recording, the live event became one form among two, valued increasingly for exactly what the object cannot supply: presence, risk, community, the unrepeatable. Concerts did not die of the record, as prophets of the 1920s predicted they might. They were repriced by it, their scarcity inverted, the fixed performance becoming the cheap and common thing and the shared hour becoming the luxury. That inversion, from a world where music was an event and objects were rare to a world where objects are numberless and events are precious, is the whole history of this article stated as an economy.

Armstrong, Holiday and Presley as recording cases

Three careers isolate the mechanisms this article has described, each in a different register. Louis Armstrong’s Hot Five and Hot Seven records of 1925 to 1928 turned the solo into the record’s centre and the record into the solo’s school, and they did it inside the three minute frame at the moment electrical recording was arriving to carry their full weight. Armstrong understood the studio as a place where a performance could be built to be studied, and the solos hold their architecture under the rehearing that he knew, being a devoted listener to records himself, they would receive. Billie Holiday’s career demonstrates the microphone’s side of the story. Her voice, small by theatre standards and unbeatable at close range, was a recording instrument: the phrasing behind the beat, the words weighed one at a time, the grain of the sound itself carrying feeling that projection would have blown apart. Holiday’s records with small groups in the middle and late 1930s made intimacy into virtuosity, and the singer who learned from them learned a technique that no conservatory taught because no concert hall required it. Her full story is told in the series biography of Billie Holiday and her life and career.

Elvis Presley completes the triangle with fame. Presley’s records for Sun Studio in 1954 and 1955, produced by Sam Phillips, who recognised a sound he had been hunting for years, fused country and rhythm and blues sources that the record trade had kept in separate catalogues, and the fusion travelled by disc into regions where its sources had never played live. When Presley moved to RCA in 1956, the machinery of national distribution, television and film amplified a fame that records had founded, and the sequence matters: the object came first, the phenomenon second. His biography in this series, Elvis Presley and his life and career, follows the life; the analytical point here is that Presley’s career is unintelligible without the groove’s ability to carry a regional hybrid across a segregated market faster than any touring band could travel. The same mechanism, at larger scale and lower cost, drives the later chapters traced in the pillar on the history of rock and pop music.

Set side by side, the three cases show recording acting as amplifier, as instrument and as vehicle in turn: Armstrong’s art amplified into a common language, Holiday’s art made possible as an instrument of closeness, Presley’s art carried as a vehicle across boundaries that live performance alone crossed slowly or not at all. A technology that can be all three in three consecutive careers is not a neutral pipe for music. It is a participant in the music, and treating it as one is the precondition for hearing the century’s records as they deserve to be heard.

The economics of the recorded object

The recorded object inverted music’s economics at the root. Live performance sells time once: the seat, the evening, the labour of the players on that night, and when the night ends the value ends with it. A master recording sells indefinitely, each copy costing a fraction of the first, so the economics of the object reward the fixed and the repeatable and penalise the ephemeral that music had always been. Copyright law scrambled to fit the new object into older categories, dividing rights among composers, performers and the owners of masters, and the divisions, settled at different dates in different countries, distributed the object’s earnings unevenly for the whole century. Britain recognised a copyright in sound recordings as such in the Copyright Act of 1911, among the earliest such recognitions anywhere, while performers in the United States worked for decades under a system in which the composition earned broadcast royalties through societies founded for songwriters, and the recording itself, when played on American radio, earned the performer nothing, an asymmetry musicians campaigned against across generations.

The object’s economics also decided who could have a career. A touring musician’s income stopped at the edge of endurance and geography; a recording artist’s catalogue kept earning in the artist’s absence, in illness, in retirement, and after the artist’s death, which created the modern catalogue business, the care of estates and reissue programmes that the survey of how posthumous albums work examines in detail. The same permanence created a new vulnerability. When copying became cheap, first through home taping in the 1970s, which the industry campaigned against as theft of its object, and then through file sharing after 1999, which collapsed the price of the copy toward zero, the inverted economics inverted again, and musicians of the early twenty first century found income migrating back toward the live event, the one thing that cannot be copied. The wheel had turned full circle with a difference: the live event now sold at prices set by its scarcity against a recorded catalogue that promotion treated as the advertisement for the concert rather than the product itself.

Fame built on records also acquired a strange patience. A live celebrity whose public stopped attending was finished; a recording artist whose catalogue kept selling could wait out fashion, and several of the century’s careers include decades in which the records worked while the person rested, taught, or simply lived unobserved, until taste rotated back and the catalogue stood ready, indistinguishable from new product to a young listener meeting it without prejudice. Day observes that the record collection made each listener a curator of the famous dead, programming evenings across centuries in a way no concert society could imitate, and that curatorship is the final form of recorded fame: not the crowd in the hall, which disperses, but the shelf in the home, which stays, and from which the voices go on being chosen, nightly, by people who owe the singers nothing and keep choosing them anyway.

What recording did to fame

Fame before recording was testimony. A singer was as great as the witnesses said, in the cities the singer had visited, for as long as the witnesses’ memories held. Recording replaced testimony with evidence that anyone could audition, and fame changed species accordingly: it became portable, comparable and cumulative. Caruso’s records made him, in the first decade of the twentieth century, the first musician whose voice was a household possession on several continents, including in homes he never entered in countries he never toured, and his record sales taught every later star’s management the foundational lesson of the industry, that the object scales and the body does not. A fame built on objects also survives its owner in working order. The dead singer’s voice stays in the catalogue, in the shop, in the radio’s rotation, competing with the living on equal acoustic terms, a condition no earlier musical culture had to accommodate and one this series, devoted to musicians no longer living, treats as its founding circumstance.

Comparability cut both ways. Where every claimant can be heard, judgment sharpens: the exaggerated reputation withers when the record is played, and the neglected one rises when a reissue puts the evidence back in circulation. Yet the same machinery manufactures fame as readily as it tests it, because distribution is a choice and promotion a budget, and the voices a generation never heard are as absent from its judgments as if they had never sung. Recording democratised the audition and oligopolised the stage door at the same time, and the tension between those two facts runs through every argument about canons, charts and greatness that the industry has produced since.

The last turn of the mechanism is the one the listener performs. In a fame economy of objects, the audience’s repeated play becomes a vote that never stops being counted, and the catalogue of 2010, the year of this article’s writing in the series calendar, stood shaped by a century of such votes, some cast in shops, some in parlours, some by collectors with machines in Maine in 1890 and some by teenagers with cassettes a hundred years later. Recording did not simply preserve music, and it did not simply distort it either. It made music into something that could be owned, studied, frozen, exported, argued with and inherited, and it made musicians answerable to their own past performances in a way no earlier generation had been. The groove ended, the tape spliced, the file copied, and through every carrier the same reversal held: the performance became the object, and the object remade the performance, and music has been negotiating the terms of that exchange ever since the needle first found its way back along the tinfoil and a rhyme spoken once in New Jersey refused, politely and permanently, to disappear.

A word about method belongs here, because the sources for this history are themselves recordings and the literature built upon them. Day writes as a historian of listening, Katz as a scholar of technology and practice, Sterne as a historian of sound culture, Milner as a chronicler of the studio, and Schmidt Horning as the historian of the engineers who ran the rooms, and the five accounts correct one another at useful points, the studio literature reminding the technology literature that machines were operated by people with habits, and the listening literature reminding both that a record is finished only when somebody hears it. Where their accounts conflict on dates or credit, this article has followed the documentary record of patents, catalogues and session ledgers, which is the closest thing the subject has to neutral ground.

Frequently Asked Questions

Q: How did recording change music?

Recording changed music by turning performance into a repeatable object, and every other change follows from that one. Once a performance could be replayed, it could be studied, memorised, sold at a distance and compared against rival versions, so improvisers acquired fixed texts, singers traded projection for microphone intimacy, and songwriters composed toward the capacity of a disc side. Tape editing and multitracking then separated the recorded work from any single act of playing it, making the studio a compositional space and the producer a co author. Mark Katz named the whole cluster the phonograph effect in Capturing Sound in 2004: the technology reshaped the practices it was built to capture, and musicians reshaped the technology in return, generation after generation.

Q: What was the first music recording technology?

The first technology that could record sound and play it back was Edison’s phonograph of 1877, which traced a voice into tinfoil wrapped around a hand cranked cylinder and then retraced the groove to reproduce the sound. An earlier French device, the phonautograph of 1857, had drawn sound as lines on soot blackened paper but could not reproduce it, so it recorded without replay. Edison’s machine was improved through the 1880s with wax cylinders by his laboratory and by Alexander Graham Bell’s associates, and Emile Berliner’s gramophone of 1887 added the flat disc, whose master could be pressed into many identical copies. Music recording as an industry begins where capture, playback and mass copying meet, and that meeting happened across those three decades rather than on a single date.

Q: Why did records have a time limit?

Records had a time limit because a groove is a physical spiral with a fixed amount of room. A ten inch disc turning at about seventy eight revolutions each minute could carry roughly three minutes of music at a loudness that rose above surface noise, and a twelve inch disc could carry four to five minutes at a higher price. Louder passages forced the groove wider and its turns further apart, which shortened the side, while quieter cutting bought time at the price of a faint, hiss drowned record. Cylinders imposed the same discipline, two minutes in the standard form and four in the later finer grooved form. The limit was therefore arithmetic before it was aesthetics: the spiral ended where it ended, and performers arranged their material to finish before the groove did.

Q: Did the three-minute limit shorten songs permanently?

The limit shortened the commercial song for as long as it ruled, and its habits outlasted the machinery that imposed them. Popular songs had often run longer in the parlour and music hall era, with multiple verses sung in sequence, and the disc side compressed that practice into a verse and chorus architecture that stated its hook early and ended decisively. When the long playing record arrived in 1948 and removed the physical constraint, album artists stretched songs freely, yet the three minute form persisted in singles, radio programming and songwriting craft, because two generations of listeners had learned it as the natural span of a song. The honest verdict is that the limit ended as a law and survived as a custom, strong in commercial songwriting and weak wherever album, concert or ceremonial forms set the terms instead.

Q: What is crooning?

Crooning is the style of close, quiet, conversational singing that the microphone made possible and the recording industry sold from the middle of the 1920s onward. Instead of projecting across a theatre, the crooner sang at the level of speech addressed to one person, letting the microphone and amplifier supply the carrying power, so breath, consonants and small inflections became expressive tools. Rudy Vallee and Bing Crosby were its defining early practitioners, Crosby above all, whose relaxed baritone ease of the early 1930s depended entirely on electrical recording and broadcast chains. The style drew moral attacks in its first years as soft and decadent, which conceded its novelty: a technology had separated singing from visible physical power and replaced public declamation with manufactured intimacy, delivered identically into millions of private rooms.

Q: How did tape editing change what a performance is?

Tape editing changed what a performance is by making the recorded performance separable from any single stretch of real time. Magnetic tape, commercialised by Ampex from 1948 onward, could be cut with a razor and spliced, so a master could join the strongest bars of several takes into a whole that nobody had ever played in one pass. Under disc recording, a side was flawed evidence of one continuous event. Under tape, the master became a construction whose joins were inaudible. Glenn Gould defended the splice as a route to the ideal performance a listener deserved, while critics who prized the concert heard construction as dishonesty. The disagreement persists because both describe the same fact: after tape, a recording stopped being proof that an event occurred and became a work in its own right.

Q: What is the difference between a live and a studio recording?

A live recording documents an event shared with an audience in one place and one stretch of time, while a studio recording is made for listeners who are absent, under conditions built for the microphone rather than the room. The live document keeps the event’s risks and accidents, its tempo pressures and its crowd, and its value lies partly in that testimony. The studio version trades testimony for control: balances set by engineers, takes chosen and edited, parts layered across days, and a perspective, close and steady, that no seat in a hall provides. Neither is the real music in a simple sense. They are two kinds of musical work answering to two contracts, one promising that something happened and the other promising that something was made as well as its makers could make it.

Q: Did recording help or harm improvised music?

Recording did both to improvised music, through the same mechanism. It helped by giving improvisation a memory: solos could be replayed, transcribed and studied by musicians continents away, so a style worked out in one studio became a shared language within years, and jazz education eventually built itself on transcribed recorded solos. It harmed by fixing what was meant to vanish, creating definitive versions against which the same player’s later work was judged, rewarding repeatable brilliance over risk, and compressing the long explorations of live sets into three minute miniatures that listeners mistook for the music’s natural scale. Louis Armstrong’s Hot Five and Hot Seven records of 1925 to 1928 show the help at its strongest, while the decline accusations thrown at ageing improvisers who no longer matched their young recorded selves show the harm in its most personal form.

Q: What is a field recording?

A field recording is a recording made outside a commercial studio, in the place where the music ordinarily lives, usually with portable equipment and usually by a visitor: an ethnographer, a collector, a label scout or a broadcaster. Jesse Walter Fewkes made some of the earliest, recording Passamaquoddy singers in Maine on wax cylinders in 1890, and commercial scouts of the 1920s carried portable disc cutters through the American South and far beyond. The category covers scholarly documentation and commercial location sessions alike, and the two were often the same trip. Field recordings preserve repertories that studios never reached, and they carry the ethical complications of the encounter itself, since the machine, the copies and the later value usually left with the visitor rather than staying with the recorded community.

Q: Were recorded communities paid royalties?

Usually not in the early decades, and the pattern varied by place and by who held the machine. Commercial location sessions of the 1920s and 1930s typically paid musicians a flat fee for the session and assigned all rights to the label, sometimes with a signature and sometimes without paperwork at all, so later sales earned the recorded singers and their communities nothing. Scholarly collectors often paid nothing and treated archive custody as the benefit, depositing cylinders in institutions far from the source. Royalties, credits and negotiated community rights belong mainly to later twentieth century practice, pressed forward by ethnomusicological ethics codes, reissue programmes and repatriation projects. Where names and ledgers survived, archival labels could restore credit and sometimes payment. The gap between what was legal, what was customary and what was fair in the early sessions is the unresolved centre of the subject.

Q: How did records spread local music worldwide?

Records spread local music by detaching sound from occasion and shipping it along the trade routes of an industry that was multinational from its first decade. From 1902 onward, engineers travelled with portable equipment to Calcutta, Cairo, Buenos Aires, Shanghai and hundreds of smaller markets, recorded local musicians, pressed the masters abroad and sold the discs back locally and outward. A style could therefore circle the world without its makers touring: Hawaiian steel guitar records taught players on several continents in the 1910s and 1920s, tango travelled from Buenos Aires to Paris and back wearing foreign prestige, and American blues and jazz reached Europe after the First World War in soldiers’ and merchants’ luggage. Export also fed back, as foreign transformations of a style returned home as fashions to imitate or resist, setting distant traditions in conversation through the groove.

Q: Did recording narrow the classical repertoire?

Recording narrowed the central classical repertoire while widening its margins, and the evidence supports both halves. At the centre, labels recorded what recognition could sell, each successful version of a Beethoven or Brahms symphony deepening the recognition that justified the next, while capable contemporaries waited decades for a first recording, and concert programming followed the catalogue because audiences buy what they have heard. Performance style narrowed with the repertoire, as musicians worldwide studied the same celebrated records and converged on international standards of precision and blend that early discs show were once far more regional. At the margins, the long playing record, independent labels and the early music movement used cheap distribution to carry neglected composers and period instrument performance to audiences no concert circuit could have assembled. The narrowing is real, and so is the countercurrent.

Q: What is a canonical performance?

A canonical performance is a recording treated as the reference version of a work, the one other performances are measured against and students are corrected toward. It earns that standing through a mixture of quality, availability and timing: the version in print, in the shops and on the radio when a generation formed its taste tends to become the landmark, and record company decisions about which versions stay in circulation quietly make and unmake canons for commercial reasons. The problem it names appears when the landmark is mistaken for the work itself, so a different tempo reads as an error rather than an interpretation. Glenn Gould’s two recordings of the Goldberg Variations, made in 1955 and 1981 with strikingly different character, are the standard antidote, since one musician produced two incompatible landmarks for a single score.

Q: Should I prefer the original release or a remaster?

Prefer whichever edition survives an honest comparison at matched volume, because loudness flatters and remasters are often louder before they are better. A good remaster returns to the original tapes with cleaner transfer equipment, repairs damage and can reveal detail that worn stampers and early digital transfers buried. A poor one raises level, squeezes dynamics and brightens the sound into a hardness the original engineers never approved. The disciplines are simple: compare the quiet passages, where overcompression shows first, check whether crescendos still grow, and distrust any edition whose chief virtue is that it sounds bigger in the shop. Original releases carry the balances the artists and producers signed off in their own era, which is evidence in their favour, but tape decay and primitive early compact disc transfers mean the original is a starting point for judgment rather than a verdict.

Q: How do I tell an overcompressed master by ear?

Listen for the absence of contrast rather than for any single defect. In an overcompressed master, quiet and loud passages arrive at nearly the same weight, a whisper and a shout from the singer feel equally close, choruses fail to lift above verses, and drums sound dense and flat instead of sudden. Crescendos that should grow across eight bars seem to lean against a ceiling, and a background tiredness sets in after ten minutes that the same music in a more dynamic master does not produce. The test works best as a comparison: play an earlier master and a later remaster of the same recording at matched volume, because unmatched volume rigs the result, and follow the quietest passage in each. If the quiet passage in the remaster sounds nearly as loud as the climax, the dynamics have been spent on level.

Q: How do I date an old recording from its sound?

Date a recording by its carrier clues and its acoustic signature together. An acoustic era recording, made before electrical methods reached studios in 1925, sounds thin at both ends, with little bass, no room air, a prominent surface hiss and ensembles arranged unnaturally around a horn, so string bass may be replaced by tuba and drums by woodblocks. Electrical recordings after 1925 gain bass weight and vocal closeness. Mono tape era records of the 1950s sound fuller and quieter, stereo arrives commercially on disc in 1957 and 1958, and multitrack productions after the middle 1960s show separation and layered parts no single room produced. Surface format helps as well: the coarse groove and three minute side point to the 78 era, the long playing record begins in 1948, and silent surfaces with wide dynamic swings belong to the compact disc era from 1982 onward.

Q: Is analogue genuinely better than digital?

Neither format wins the argument outright, and the honest evidence cuts against simple verdicts in both directions. Measured on noise, distortion and dynamic range, the compact disc surpasses the vinyl record comfortably, and listening tests conducted blind repeatedly shrink differences that listeners report confidently when they can see which carrier is playing. Vinyl partisans answer, with some justice, that the argument was never only about measurements: the ritual of the side, the large sleeve, the enforced sequence and a certain forgiving quality in well cut vinyl shape an attention that the file and the disc discourage. Mark Katz frames format loyalty as loyalty to a whole practice of listening built around a carrier. Judge by use, therefore. For archival stability, convenience and measurable fidelity, digital wins. For a listening ritual that some ears and households genuinely prefer, analogue keeps its case.

Q: Which historic recordings are worth hearing first?

Begin where the medium and the music explain each other. Enrico Caruso’s Milan sides of 1902 show the early industry discovering its first global voice, and Louis Armstrong’s Hot Five and Hot Seven records of 1925 to 1928 show the solo becoming a studyable text inside the three minute frame. Billie Holiday’s small group records of the middle and late 1930s demonstrate what the microphone did to singing, and Glenn Gould’s 1955 Goldberg Variations make the case for the studio as a place of thought. Add the 1927 Bristol sessions recordings of the Carter Family and Jimmie Rodgers for the field and location tradition, and any complete symphony from the early electrical era for the sound of an orchestra meeting the microphone. Heard in that order, the records teach the history this article describes while standing, each of them, as music worth the hour on its own account.

Q: How should I listen to a very old recording?

Listen through the surface rather than at it, the way the eye reads in poor light, and give the ear a few minutes to adjust before judging anything. Play the recording at a modest volume, since cranking an acoustic era disc only enlarges its noise, and follow one strand at a time on repeated hearings: the singer’s phrasing first, the accompaniment second, the room and the ensemble balance third. Context repays effort, so learn what the machines of the year could and could not capture before blaming a thin bass line on the players, when the horn simply could not hear the bass at all. Above all, resist remastered prettiness as the standard. A clean modern transfer helps, but the goal is to meet a performance across a century of noise, and the hiss is part of the distance being crossed, not merely an obstacle in front of the music.

Q: Is a studio-built record still a performance?

Yes, under a widened definition of performance that the technology itself forced. A studio built record distributes the performance across time and people: parts played on different days, takes joined by editors, balances set at a desk, and the producer’s choices running through every bar. Nothing in that process is fake in itself, since every contribution was really played, sung or decided by somebody, and audiences have always accepted constructed artefacts, from the edited studio film to the novel revised across drafts, as works rather than as frauds. What changes is the kind of trust involved. A concert performance asks trust in an event. A studio record asks trust in a construction, and it betrays that trust only when it claims the first kind while delivering the second without the listener’s knowledge. Judged as construction, the great studio records are performances of the highest order, performed at the speed of weeks.