In 1972 Roland Barthes built a theory of the singing voice around a flaw he could not quite name: the thing that made one baritone thrilling and another, more technically flawless, dead on arrival. He called it the grain — “the materiality of the body speaking its mother tongue”. Twenty-five years later, an engineer who had spent his career reading seismic echoes for the oil industry built, almost as a dinner-table joke, the precise machine for erasing it. Put Barthes and Auto-Tune in the same room and his short, difficult essay turns into something useful: a test for which voices, in an age of perfect pitch, still let a body be heard.

The body in the voice

Barthes locates the grain in the flesh. It is, he writes, “something which is directly the singer’s body, brought by one and the same movement to your ear from the depths of the body’s cavities, the muscles, the membranes, the cartilage”. It is not, or not only, timbre; it lives in the friction between the music and the language being sung, and he files it under signifiance rather than ordinary meaning — borrowing from Julia Kristeva the pair of terms pheno-song (everything serving communication, expression, correctness) and geno-song (the level where the body works on the language, beyond any message). He even attacks the technical cult of breath control as “the myth of breath”: the lungs, he says, swell but do not signify; it is in the throat, the place of friction, that something happens. To make the distinction concrete he sets two baritones against each other. The supremely accomplished Dietrich Fischer-Dieskau he hears as a voice without grain — “here it is the soul that accompanies the song, not the body” — and, more bluntly, “I seem to hear only the lungs, never the tongue, the glottis, the teeth, the sinuses, the nose”. The older, rougher Charles Panzéra is his man of the grain, whose “entire art was in the letters, not in the bellows”.

The machine built to erase it

Now the machine. Auto-Tune was created by Andy Hildebrand, a signal-processing engineer who had analysed seismic data for the oil industry before turning the mathematics to music; it shipped in 1997. The origin story is almost too apt: at a trade-show lunch a distributor’s wife challenged him — “why don’t you make me a box that would have me sing in tune?” — and, as Hildebrand recalls, “everyone just stared at their lunch plates, they didn’t say a word.” He thought it a lousy idea, then realised it was straightforward to build. The hard, robotic “discretize” setting that produced the famous effect on Cher’s “Believe” (1998) was nearly cut from the product; its producers then hid the technique as a trade secret, claiming in print that the sound came from a vocoder pedal. By 2009 Rick Rubin could say flatly, “Right now, if you listen to pop, everything is in perfect pitch, perfect time and perfect tune. That’s how ubiquitous Auto-Tune is.” Melodyne pushed the project further still, letting an engineer re-pitch a single note inside a chord. The pheno-song — the realm of correctness — had been fully automated. The body could be sanded out of the signal entirely, and usually was.

The body, kept on purpose

Which is exactly what makes Barthes newly useful, because it turns the grain from a default into a decision. Some voices keep the body audible on purpose. Tom Waits’s gravel — a voice one critic described as sounding “soaked in a vat of bourbon, left hanging in the smokehouse for a few months, and then taken outside and run over with a car” — is almost pure geno-song, where the damage is the meaning. Billie Eilish’s intimacy is built by putting the microphone so close that the proximity effect lifts the low end and the ear reads weight, breath and presence — the larynx itself made audible. The scholar Anna Muchitsch reads ANOHNI’s wide, irregular vibrato straight through Barthes: it “calls forth a body in motion, the oscillating tone being a sonic signature of muscular contractions in the larynx”. The grain, in each case, is not the absence of craft. It is craft pointed at the body rather than away from it.

Pitch-correction did not kill the grain. It quietly moved the responsibility for it onto the listener.

Machine against body is the wrong fight

The tidy opposition — machine erases the body, brave artists preserve it — is false, and the honest essay has to say so. The very singers who keep the body audible are often the heaviest processors: Bon Iver runs his voice through a bespoke rig and Auto-Tune, yet the result sounds more, not less, embodied. As his engineer Chris Messina put it, “It’s merely an effect… The thing that makes the sound of it new is the way that Justin plays it.” Eilish’s closeness is an engineered artifact, not raw nature. So the grain is not the refusal of technology but a use of technology that lets a body remain legible in the signal. And Barthes himself complicates any populist hymn to the “authentic” voice — he was a mandarin who openly preferred a pre-mass-culture art. The real question his essay leaves us is not machine versus body. It is which uses of the machine keep a person audible inside the sound, and which dissolve them.

Barthes, tellingly, located the grain not in the recording but in his own ear: “Am I alone in hearing it? Am I hearing voices in the voice?” The essay was never a rulebook for singers; it was a wager on the listener — that a body in a voice can be heard only by a body willing to listen for it. Pitch-correction has not settled that wager so much as raised the stakes. In a signal where everything can be made smooth, choosing to hear the cracked note, the caught breath, the trembling vibrato is no longer passive reception but an act of recognition: one body acknowledging another across the wire. The grain matters now precisely because it has become optional — and an optional thing is one we can decide, every time we listen, to keep.

Sources

  • Roland Barthes, “The Grain of the Voice” (trans. Richard Howard, in The Responsibility of Forms) — the grain, pheno-/geno-song, Panzéra and Fischer-Dieskau, the “myth of breath” — text PDF. First published as “Le grain de la voix”, Musique en jeu, 9 (1972); also translated by Stephen Heath in Image-Music-Text (1977).
  • Andy Hildebrand and the invention of Auto-Tune (1997) — Vice; the Cher “Believe” effect and its concealment — Sound On Sound; ubiquity and the Rubin quote — Wikipedia: Auto-Tune; Melodyne’s note-level editing — Sound On Sound.
  • Billie Eilish’s close-miked vocals — Headliner; the proximity effect (physics of intimacy) — My New Microphone.
  • ANOHNI’s vibrato read through Barthes (Anna Muchitsch) — Cambridge / Popular Music; Bon Iver’s “Messina” rig (Chris Messina) — W Magazine.
  • Tom Waits’s voice (Daniel Durchholz) — Wikipedia; the backlash, Jay-Z’s “D.O.A. (Death of Auto-Tune)” — Wikipedia.

Continue reading

  • #roland-barthes
  • #the-voice
  • #auto-tune
  • #pitch-correction
  • #singing

← Back to the feed