In early 2000, south of Oakland, California, a physicist stuck in traffic was listening to the radio. He heard Mickey Hart, a drummer for the Grateful Dead, say that the archives of the world’s aboriginal musics were deteriorating and needed attention. The bulk of the archives had been assembled between 1890 and 1940 by ethnographers using antique devices that recorded mainly on wax cylinders and aluminum disks. Many of the recordings had not been played for a number of years and had grown so fragile that the pressure of a stylus might destroy them.
Anything you can embed sound in you can scratch, crease, crumple, bend, break, tear, warp, or melt. Oil and dust and dirt get in the path of a stylus and deepen, widen, or distort a groove, making the stylus skip. Something, possibly mold, attacks the grooves on wax cylinders, leaving gaps that sound as if someone were banging a drum. On cylinders of Native American music, which typically features singing and chanting, this sometimes sounds like a phantom accompaniment. Wax cylinders were meant for businessmen to dictate letters on. After a letter had been typed, the cylinder was scraped clean with a tool the manufacturer supplied, so that another letter could be dictated, until the cylinder was used up. No one regarded the cylinders as permanent, and permanence wasn’t what the ethnologists sought. They wanted a version to transcribe—they wanted to read the text more than hear it.
The physicist, whose name is Carl Haber, works at the Lawrence Berkeley Laboratory, in Berkeley, California. There are essentially two kinds of physicists—experimental and theoretical. Experimental physicists test the ideas of theoretical physicists. Sometimes they observe peculiarities that theoretical physicists then name and explain. Haber is an experimental physicist. At Lawrence Berkeley, he is a member of the ATLAS Group, an international alliance of labs and universities that conduct experiments at the CERN collider, in Switzerland, where they discovered the Higgs boson, the “God particle,” in 2012. The ATLAS Group is huge. When a paper is published, Haber says, “it has more than two thousand names on it.” Haber specializes in the silicon detectors that line the collision chamber at CERN. The detectors are a little bigger than a passport, there are thousands of them, and they are arrayed like shingles around the chamber. They track the paths of subatomic particles in the aftermath of their impacts. Forty million collisions take place each second.
Haber regards himself as an instrument builder—that is, as someone who, after considering what device an experiment requires, designs the device and makes it or sees to its being made by a custom shop. To align the detectors at CERN, he had used a machine called a SmartScope, which “photographs the detectors in microscopic detail, then analyzes the images and the placement over and over again,” he says. This type of measuring is called optical metrology. The SmartScope had worked so well that Haber was drawing sketches and writing notes for new ways to use it. “When I heard Mickey Hart talking about sound recordings, I thought, Maybe I’ve been considering the wrong problem,” he says.
With the SmartScope, Haber had precisely positioned the detectors at CERN without touching them. He wondered what might happen if he used the machine to photograph the grooves of a cylinder or a disk that was too fragile to play. “I wondered, Could we somehow use optical metrology in some brute-force way to have the images interpreted as sound?” he says.
Haber went to Down Home Music Store, in El Cerrito, and bought a couple of 78s, “Goodnight, Irene,” by the Weavers, and “Whispering,” by Les Paul. The records were made from shellac, which is brittle; one of his colleagues dropped “Whispering,” and it broke. Thirteen people at Lawrence Berkeley have won Nobel Prizes, most recently in 2011, when Saul Perlmutter won “for the discovery of the accelerating expansion of the Universe through observations of distant supernovae.” Haber told colleagues that he wanted to measure the grooves on the phonograph records. A postdoctoral student named Vitaliy Fadeyev volunteered to help. “I remember colleagues who were just dismissive,” Fadeyev says. “Like basically not giving it enough thought—the actual value of it, the way you can save history and culture. How to do it, once you are told, and you are a physicist, it’s almost obvious. I find this is the case with all big discoveries. Once you are told the solution, it’s almost too simple. But, when you are trying to find the solution, you are scratching your head.”
Fadeyev scanned the groove of “Goodnight, Irene” as it unfolded over several of the record’s revolutions. Each revolution took close to forty minutes to complete. In all, he scanned about twenty seconds of music. A computer, following the path of the groove as if it were a stylus, converted the images to music.
“When I finally told Carl, ‘We can scan the record,’ he was shocked,” Fadeyev says. “He brought our senior engineer into the room, and that engineer laughed out loud. He saw an unexpected use of the machine—not measuring the object but getting the sound.”
Fadeyev’s scan “demonstrated in a very crude way that you could use a tool like this to measure a record and get sound, so I knew the idea had some merit,” Haber says. “It was laborious; it took hours; it was not practical. But I knew that, if we could make a machine that did this in a purposeful way, it could work. In physics, there are tons of brilliant ideas, but you have to make them work.”
Silence is imaginary, because the world never stops making noise. A sound is a disruption of the air, and it doesn’t so much die as recede until it subsides beneath the level of the world’s random noise and can no longer be recovered, like a face that is lost in a crowd. In past times, people sometimes thought that all sounds that ever existed were still present, hovering like ghosts. Guglielmo Marconi, who sent the first radio message, in 1902, believed that with a microphone that was sufficiently sensitive he could hear Jesus delivering the Sermon on the Mount, and in 1925 a writer for the Washington Post speculated that a radio was capable of broadcasting the voices of the dead. A radio transmits vibrations, he wrote, and the voices of the dead “simply vibrate at a lower rate.”
Except for an echo, only a sound that has been recorded can be heard again. A recording is a pattern of sound waves embedded in a medium. A person speaks, the sound moves a diaphragm, the diaphragm stirs a stylus, and the stylus traces a pattern analogous to the sound: an analog recording. (A digital recording converts the pattern of the waves into numbers.) Among audio preservationists, the few people who search for antique sounds and try to play them are sometimes called archeophonists; their field is archeophony. The oldest sound archeophonists have recovered is a proofreader of medical texts singing a folk song in Paris in April of 1860. As far as anyone knows, the proofreader, whose name was Édouard-Léon Scott de Martinville, recorded sound before anyone else did. Scott embedded his voice in soot on a piece of paper. He never heard it. The means for playing back sound weren’t invented until 1877, when Thomas Edison recorded himself on tinfoil. No one heard Scott until 2008, when Haber managed to play Scott’s recording. For his work in audio preservation, Haber received a MacArthur Fellowship in 2013.
Haber is fifty-five. He is small-framed, with a long face, short, wiry gray-brown hair like Brillo, and a beard. His manner of dress is informal, and his appearance is slightly rumpled-looking. His expression is alert and inquiring and placidly intelligent. He hardly ever raises his voice. He grew up in Queens, got his undergraduate and graduate degrees at Columbia, and began working at the Lawrence Berkeley Lab in 1985. For more than fifteen years, he was part of a project that discovered the top quark, a subatomic particle, at the Collider Detector at Fermilab, in Illinois, a smaller version of the collider at CERN. (The Fermilab collider is roughly four miles around; the CERN collider is seventeen.) Haber’s team built a device called a silicon vertex detector. “You work twenty hours, then you go off and cook pasta and drink wine with your friends, who are also working, then you go back to work,” he said. “It’s like being in a band. It was tons of fun.” When that work ended, Haber was added to the ATLAS team.
In late 2003, Haber and Fadeyev wrote a paper called “Reconstruction of Mechanically Recorded Sound by Image Processing,” in which they described how their work might be used to make digital copies of fragile artifacts. They sent the paper to places that might be interested, among them the Library of Congress, where it reached Peter Alyea, who was involved in digitizing the library’s sound recordings. “You had to understand some of the physics lingo, but my father is a physicist, so I had grown up with physics talk,” Alyea says. “All these light bulbs were going off for me, though, because I realized that it applied to all these possibilities.”
Meanwhile, Haber had measured his first wax cylinders and was deciding what capabilities his new machine should have. Analog recordings—that is, cylinders and records—have two kinds of grooves, lateral and vertical. Records have lateral grooves and cylinders have vertical ones. Lateral grooves move side to side, like a river through a plain; vertical grooves rise and fall like hills. The SmartScope took two-dimensional images of reflected light. For lateral grooves, this worked fine, because only two dimensions are necessary to show side-to-side movement, but vertical grooves require three dimensions, in order to convey depth.
In three dimensions, what is measured is not the light reflected from a surface but the distance of the surface from the lens. Three-dimensional scanning registers more of the groove, and is more precise: two dimensions is a map; three dimensions is a topographical map. To depict depth, Haber used a device called a confocal microscope, which measures depth across a line parallel to the axis of a cylinder. As the cylinder turned, the microscope could acquire up to eighteen hundred lines a second. Haber called his machine IRENE, after “Goodnight, Irene.” IRENE stands for Image Reconstruction Eliminate Noise, Etc. He built one for his lab and one for the Library of Congress, which gave money for his research. Each machine can measure both two and three dimensions, and each cost about two hundred thousand dollars in parts and labor. There are now five of them. In addition to Haber’s IRENE, in Berkeley, the Library of Congress has two; there is one in Andover, Massachusetts, at the Northeast Document Conservation Center; and there is one in India, at the R. Muthiah Research Library.
With IRENE, archeophonists could now listen to cylinders and disks that had been considered unplayable. Many of these songs and speeches and musics, held at institutions such as the Smithsonian, the Library of Congress, the University of California, Harvard, and the British Museum, had been dormant for more than a century. The Smithsonian has a number of artifacts from Alexander Graham Bell’s laboratory, the Volta Lab, in Washington, D.C., where he had worked on recording sound, mostly in wax. A curator named Carlene Stephens had overseen the collection for nearly thirty years without having heard any of the recordings or, in some cases, even knowing what the disks contained. She assumed that she would retire having carried out her responsibility to preserve them. In 2010, from a wax disk, Haber recovered Bell saying, in 1885, “Hear my voice, Alexander Graham Bell.” It was the first time anyone alive had heard what he sounded like.
Before Haber built IRENE, an archeophonist could choose only between not playing a fragile cylinder, for fear of destroying it, and playing it in order to convert it to a digital form and then accepting the possibility that the original would be ruined. Mickey Hart told me that, playing a recording in order to copy it, perhaps for the last time, he is “filled with sorrow and total joy—we’ve gotten it before time has taken it from us, but it’s also gone, too.
“There are a lot of great musics that are dying every day, falling off the edges of the world,” he went on. “Some have disappeared entirely, of course. The chances of a recording making it are slight. The mortality is great. The cultures they belong to may be dying, too. Haber has come up with this jewel that will allow us to preserve this music and give it back to the cultures that spawned it.”
Not every culture wants its music restored, however. “Some of them think it’s too sacred to hear,” Hart said. “They want it locked up forever.” Haber has been asked by a California tribe to preserve cylinders that have songs and speeches on them. These include creature and spirit songs, dancing songs, and gambling songs. For the Burke Museum, in Seattle, he has preserved music collected in 1930 by the anthropologist Franz Boas. The music was made in a village on Vancouver Island called Fort Rupert by members of the Kwakiutl tribe. The occasion was the filming of a silent movie. On the film, you can see figures dancing, and behind one of them a man singing into a phonograph, a recording machine of the period. Many of the songs and rites involved have been forgotten. If they are remembered, it is often only partly—in the case of a song, a first verse, perhaps. Katie Bunn-Marcuse, a curator at the museum, was able to take the film to Kwakiutl towns and play alongside it the music that Haber had made available, allowing the people to hear their tribal songs. “It was very emotional,” Bunn-Marcuse says. “Elderly men and women heard their ancestors.”
Haber described IRENE to anyone who invited him to speak. In 2006, he began giving an annual talk on sound preservation at the Library of Congress. After the one in 2007, a man named David Giovannoni spoke to him. Giovannoni was part of a collaborative that called itself First Sounds. Its other founding members were Patrick Feaster, a lecturer at Indiana University, and Richard Martin and Meagan Hennessey. Martin and Hennessey have a company called Archeophone Records, which received a Grammy in 2007 for “Lost Sounds: Blacks and the Birth of the Recording Industry 1891-1922.”
When the four of them discussed what their next project might be, someone suggested finding the oldest retrievable sounds. “Patrick knew that there was this guy in France named Édouard-Léon Scott de Martinville, who had made phonautograms on what he called the phonautograph,” Giovannoni says. “We did not yet know how many phonautograms had survived.”
Having proofread a passage describing the anatomy of the human ear, Scott had designed the phonautograph to approximate it. A phonautograph looks a little like the kind of early phonograph that had a turntable and a horn, except that, instead of a turntable, the phonautograph had a cylinder about the size of a coffee can mounted sideways between upright brackets. To make a phonautogram, Scott fastened a sheet of paper around the cylinder, then blackened the paper in the smoke of a burning lamp. Pressed to the cylinder was a stylus made from pig’s bristle. The stylus was connected to a diaphragm stretched across a narrow opening at the bottom of a small barrel. Turning the cylinder with the handle, Scott stood at the open end of the barrel and talked or sang.
Recording one’s voice in soot seems like a screwball idea, until you know that Scott wasn’t interested in hearing recorded speech. He was interested in the pattern made by the sound waves representing the words, which he called a “natural stenography,” and which he believed a person could learn to read. What people said was more revealing of character, he thought, and more vital than what they wrote; a phonautogram was a visual record of the pitch of someone’s voice, how loudly he spoke, and with what emphasis. It was, Scott wrote, “living speech; our manual or printed calligraphy is nothing but dead speech.”
A phonautogram would enable a singer to record a phrase precisely, leaving “an imperishable trace of those fugitive melodies which the memory no longer finds when it seeks them.” Two people could record a conversation. A phonautogram could also preserve the performances of “those grand artists who die without leaving behind them the faintest trace of their genius.” Furthermore, a deaf person could imitate the patterns on the paper rather than trying to speak by imitating the movements of someone’s lips. When he could produce the same patterns, he would know that he was speaking correctly. Scott regarded Edison’s invention of the phonograph, in 1877, as pointless, because it reproduced sound instead of writing it.
Scott started experimenting with phonautograms in 1853; the first phonautograms Giovannoni turned up, however, were in New Jersey and had been made by Thomas Edison in the summer of 1878. Edison had been hired by the Metropolitan Elevated Railroad, in New York, to determine why the trains made so much noise. He had modified a phonograph to trace sound waves on paper. Giovannoni scanned some of these and gave them to Haber. Haber and a colleague, Earl Cornell, managed to produce a few seconds of a sort of muddy, rushing sound that you could tell was a train only if you already knew that it was a train. “I kept hoping someone would say, ‘Here comes a train,’ ” Cornell says, “but mostly it was pretty boring.”
In the meantime, translating a book written in French by Scott, Patrick Feaster deduced that there were Scott phonautograms in Paris, at the French Academy of Sciences. Giovannoni took a scanner to Paris and scanned a few phonautograms. The phonautogram that he gave Haber was on heavy paper, about a foot and a half long. The paper had pairs of scratchy, wavy lines in rows across it. There were seventeen pairs in all, each of which represented a turn of the drum. As Haber and Cornell worked on recovering one line at a time, the sound emerged in increments. At first, Cornell said, they thought they heard someone groaning. Then they began to discern a pitch rising and falling, and they realized that someone was singing.
Haber and Cornell sent the individual lines to Feaster, who stayed up all night assembling them. Beside the track of the singer’s voice, Scott had recorded a tuning fork, so that someone studying the phonautogram could measure the duration of a specific sound. The tuning fork, he wrote, was vibrating at “500 simple vibrations.” When Feaster set the phonautogram at five hundred vibrations a second, he heard a woman with a warbly voice, as if she were singing on a record that was rotating unevenly.
“I was listening in bits and snatches, removing what fluctuations I could, so it really shaped up over the course of the night,” he said. By the early morning, he realized that the woman was singing “Au Clair de la Lune,” and that he “was the first person to hear someone singing a recognizable tune before the outbreak of the American Civil War.”
A couple of months later, Feaster was working on another Scott phonautogram. A man was speaking, but at five hundred vibrations his voice was clearly too high-pitched. Feaster discovered that five hundred simple vibrations in the middle of the nineteenth century, in France, meant two hundred and fifty modern vibrations. When “Au Clair de la Lune” was played at the right speed, the singer was Scott. His voice sounded lugubrious and spooky, like a communication from the underworld. No one knows why the recording is so deliberate. He may have been trying to have each phrase last one rotation of the drum. Or perhaps he wanted the shapes of the letters and syllables to be very clear, so that they would be simple to read. In any case, he was not giving a performance; he was conducting an experiment. He was concerned with making a visual record, not an aural one.
Haber spent a day recently with Peter Alyea at the Library of Congress, in a subbasement room where the library has one of its IRENEs. They were planning to do some maintenance on it. Haber sat on an office chair while Alyea knelt on the floor, examining the computer that operates IRENE. The computer wasn’t functioning the way they expected it to. As Alyea removed parts and unplugged wires, Haber said that many of the cylinders that interest him were broken. The breaks interrupt the sound, leaving gaps, which IRENE can repair. “There’s a mathematical procedure called interpolation, which means, basically, that you connect the ends of the gap with a smooth curve,” he said. “Whether you hear the fix depends on how long the gap is. If you did nothing, you might hear a click. With the fix, your ability to understand in certain contexts will be enhanced by not having the distraction of the clicking. You might not even hear the repair, but it has nothing to do with the truth. For example, if the recording is forensic—if there’s a gunshot in the gap—you’re lost. But if you’re trying to understand a Native American language it might help.”
IRENE wasn’t responding to Alyea’s instructions. He typed several commands, then got up and left. I asked Haber how a gap was filled. “If the missing section is music, you can stick in silence,” he said. “Or, instead of silence, if you’re trying to make it listenable, you can take music from another part and patch it in there. It’s like detecting a skin tone on either side of a blemish and using that to conceal the blemish, like Photoshop. Psychologically, your mind may not react if the contrast is not so great. If it’s a very big contrast, it’s more obvious.”
Alyea appeared, walking awkwardly and carrying a computer the size of a small suitcase. “You should know that, in the preservation sense, that’s a no-no,” he said for my benefit. “It’s called a listener’s copy.”
“It’s part of the ethics of preservation,” Haber said. “There’s a notion of doing as little as possible, not adding interpretations and leaving the material in a form that others can work with.”
“When you add information, it becomes ambiguous what you’ve done,” Alyea said.
“Restoration implies choices,” Haber said. “Preservation implies stabilizing. With preservation, you want to leave the data clean.”
Alyea sat on the floor and began removing parts from the computer to see if they were connected properly. “If you take a cylinder and you break it, particularly if it broke fifty years ago, everything on the pieces moves,” Haber said. “They absorb moisture, they expand or contract, they deform and don’t come back to where they were. Imagine there are parallel lanes on a road, and there’s been an earthquake.”
Alyea, sitting cross-legged, had parts of the computer arrayed around him. I asked if he was making any progress.
“We’re always making progress; it’s a question of what direction,” he said glumly.
Haber turned the pages of the computer’s manual. I could see headings such as “Convolution and Deconvolution” and “Helix Development.”
Alyea began connecting wires to the computer. The computer made a sound like a squeal. “That’s bad, that’s worse than before,” he said. “That’s an unhappy sound.” A red light on the computer came on. “It was green before,” Alyea said.
They tried several fixes, then Haber said, “I think we’re up against the wall on this one now.”
Alyea took off his glasses and rubbed his eyes. He spent some time on the phone with Earl Cornell, who was in California, but they weren’t able to fix the problem. He and Haber then discussed whether they should call the help line for the computer’s manufacturer.
“What question are we going to ask?” Alyea said.
Haber dialled the number for the help line and said, “This is Carl Haber, from the Lawrence Berkeley Lab.” He described the problem, and was put on hold.
Alyea was on his knees, working with a screwdriver. “I think the capacitors are the problem,” he said. He and Haber switched places, and Alyea began talking to the technician. Alyea put the technician on speakerphone.
I said, “I guess they can’t just tell you to reboot.”
Haber looked at me soberly. “There is no script for this,” he said.
The next day, IRENE worked fine, but they didn’t know why. Alyea said that happens sometimes.
There are approximately seventy Native American languages in California, far more than in any other state. In 2007, Haber scanned cylinders of Native American speakers from the Phoebe Hearst Museum, selected for him by Andrew Garrett, a linguistics professor at Berkeley. Garrett is interested in Yurok, which is spoken along the lower Klamath River, in northwestern California. For Harvard, Haber scanned aluminum disks of singers of epic poetry made by Milman Parry, in Yugoslavia, between 1933 and 1935. Parry and a student of his, Albert Lord, formed the Parry-Lord hypothesis, which proposed that epic poems such as the Iliad and the Odyssey were made up of formulas that involved descriptions and characters and scenes that were assembled for an occasion by poets who were improvising, much as musicians improvise according to a song’s chord structures and the rules of harmony. Recently, Haber was visited by some British scholars who are interested in a collection, in Iran, of Persian music that is on cylinders that are broken into many pieces.
One afternoon, I drove with Haber from Berkeley to Palo Alto to visit the Stanford University Archive of Recorded Sound, in order to meet a musician who was interested in preserving piano rolls. Piano rolls were used on player pianos between about 1900 and 1930, when records replaced them. The rolls have holes in them, which act as instructions to a player piano, like the rolls in a music box. The musician, a pianist who also lectures at Stanford, was named Kumaran Arul.
At the Stanford library, we went downstairs to a windowless room. Haber was holding one temple of his glasses and spinning them like a pinwheel when Arul arrived. On a table were seven file boxes of piano rolls, which looked like scrolls. Haber sat at one end of the table and said, “What have you got?”
“Let me just explain,” Arul said. “I’m interested in rolls from the early part of the twentieth century. Measured by time, they are the longest recordings we have from the period.”
The librarian unrolled one. There were dots and long lines cut into it.
“This is a code,” Haber said.
The librarian showed us a roll of a Chopin étude played by Arthur Rubinstein and one of Maurice Ravel playing his piece “Sonatine.”
“The question is: how well do we need to know the exact shapes of those holes to get the nuances of the sound?” Haber asked.
The rolls preserved playing styles popular in the period which had subsequently fallen into disuse, Arul said. “Hearing them virtually revises how we understand what nineteenth-century music sounded like,” he said. “It’s that radical.” The styles conflict with the way piano is taught, he told me later. “The ideal example is probably Carl Reinecke. He’s nearly the earliest-born person we have any recordings of. He was born in 1824, he knew Schumann and Mendelssohn, and he was known for the way he played Mozart, which would violate all the principles we are taught today. His style was rhythmically much more irregular, and he didn’t align vertical parts—if there are two notes written simultaneously, he will not play them together. He does it so often that it sounds practically chaotic. Most professionals who listen to it think there is something wrong with the roll.”
Arul told Haber that the rolls were not as fragile as cylinders. The paper dried out, but most of the rolls that were in the range of a hundred years old could still be played. “A number of composers wrote for piano rolls,” Arul said. “Stravinsky wrote his ballet ‘Les Noces’ with parts for piano rolls.” There had been “a cottage-enthusiast industry copying them over the last fifteen years,” he went on. “But the quality varies, so we are trying to start the process of having an archival standard.”
Scholars often dismissed the rolls, Arul said—“ ‘It’s a robot,’ they say.” Haber said he liked the idea of a conference devoted to music by automatons, which would feature the rolls, but “that opens a whole new can of worms.”
Haber and I said goodbye to Arul, then went to a café on the campus. Haber said that he was interested in the opportunity to work on a system different from grooves. “I’m enthralled,” he said. “I like the idea that it’s a code.” He took out a pen and, on a napkin, made a drawing of how he imagined a piano roll worked, with each line representing one of the keyboard’s eighty-eight keys, and each hole an instruction about how forcefully to strike the key and how long to hold it.
He put the drawing aside. He said that what intrigued him about recovering relic sounds was the period and the figures who inhabited it. “Roughly toward the end of the nineteenth century, there were these early guys—I like to call them the heroic inventors,” he said. “Edison, Bell, Muybridge with his time studies, Marconi. They were not particularly well established academically; they were not trained as engineers, mathematicians, or scientists; they were very creative; and they did intuitive, seat-of-the-pants, trial-and-error experiments, whereas once you get into the twentieth century, and you have an understanding of the physics and chemistry involved in these original scientific gestures, you get engineers and academics doing this kind of work. They’re more cautious. No scientist would have thought you could hear Jesus. It violates the Second Law of Thermodynamics.”
He shook his head.
“Anyway, they were the first to record the world as it was actually happening,” he continued. “Before Scott, people wrote things down. They took photographs, they captured moments, but nobody was freezing the world in a sequence, as it unfolded. Sound, for example, is a phenomenon that changes as a function of time. Electrical signals, the brightness of stars—they’re changing as a function of time. Anything we measure today in a laboratory—radio signals, meteorological equipment, subatomic particles, the ability to divide time into smaller and smaller increments—it all comes from the outline of this work.”
We had bought tea, and Haber drank some. “Once the telegraph got installed, suddenly it mattered to keep track of time in New York and Pittsburgh,” he went on. “It made a difference when the train left. One telegraph network needed to be able to speak to another. These rapid breakthroughs in the measurement of time were very important.”
A young woman asked to sit in the seat beside Haber, and he moved to make room for her. “Thirteen years ago, it did not occur to me, stuck in traffic and listening to Mickey Hart about lost musics, that we would have succeeded in playing all these iconic recordings of the nineteenth century,” he said. “I only thought, Here’s a way we can be of help to a community of serious people. They were custodians and stewards of world treasures.” ♦
