Humanity’s First Recordings of its Own Voice [David Giovannoni]


It’s a great privilege to join this distinguished gathering today to explore the French and America origins of sound recording today’s symposium Honors Edward Leon Scott De Martinville and his invention of sound recording on the bicentennial of his birth we congregate in a big tent a metaphor perhaps for our unity of purpose amidst our diversity We sit directly above the vault that Safeguards Edison’s documentary legacy And we strive to harmonize the American invention of sound recording with Scott De Martinville For this we draw on Scott’s manuscripts drawings Recordings and other primary materials that we’ve identified in the venerable institutions that have safely preserved these documents for more than a century and a half the academy of sciences the institute of France the patent office and the society for the advancement of National industry among these documents are several Dozen sound recordings made by Scott De Martinville between 1857 and 1860 we verified that they were in fact sound recordings by playing them back something the French inventor never intended this recording from the archives of the French academy of sciences Was made a full 17 years before thomas Edison invented the phonograph As you might imagine our playback was considered newsworthy as its somewhat disturbed the narrative of Thomas Edison being the first inventor to record sound at midnight on March 27th 2008 the New York times website published Jody Rosen’s article about our discovery and the sound we’d recovered It was midday in Asia Europe was just waking up and within minutes the story was international news Americans awoke the next morning to a revised historical reality in which the French beat Edison The BBC located Lawrence Scott de Martinville and asked how he felt about the resurrection of his great-grandfather’s Recordings, I’m so happy for him because in his last will he asked his children and grandchildren to make sure that his Discovery would not be ignored this is a Moral task that was given to us by his own handwritten will which and this one I have at home and We were particularly proud in the family we tried to restore his memory My father and also my grandfather and they had marble plate that was put on the wall of the national library Vivian in Paris that was the place where he Lived for the last period of his life and when he passed away and on this plate He said that he had invented how to record sounds as a matter of fact when when he passed away he had just been as a Spectator assisting to the presentation of Mr.. Edison’s invention in Paris and His name was not even mentioned And he was so upset that I believe that this is one of the reasons that eventually ended tragically this is going to require The rewriting an awful lot of history books isn’t it? well only a few sentences met Key ones What are those key sentences? That’s the question we’re here to answer today What are the new facts, how do we interpret them? What’s the new narrative the corrected narrative of How Scott made the invisible Force of sound? Visible of How he made the transient echo permanent How do we square this achievement with Edison’s invention of the phonograph 20 years later? How do we reframe the accomplishments of these two men who independently? invented sound recording At this symposium today are people for whom these questions matter You are the ones who will write these key sentences you’re teaching the history of sound recording at universities You’re writing the next great biography of Edison or the ultimate textbook on Humanity’s control over sound You’re studying the interplay between pure science and applied technology You’re writing masters theses doctoral dissertations historical articles. You’re making documentaries You’re interpreting history for park visitors or telling the story to your readers listeners or viewers the facts matter and getting the story right matters To this end the park recently opened an exhibit that asks who invented sound recording and in answer to the question affirms that sound recording was invented twice the first time by Scott in the 1850s and again in 1877 by Edison Scott called his invention the phonograph Edison called his the phonograph each invention was a significant technological Achievement in its day the citizens of France and the u.s. rightly take pride in their inventors accomplishments But today as we prepare to rewrite the history books I hope that we can transcend the overt Nationalism of past narratives which divides sharply along country of Origin fren texts go like this Scott invented the phonograph from which Edison obviously took the idea for the phonograph American texts go like this Scott invented the phonograph and why on Earth he didn’t think of playing back his recordings is Beyond our ability to Comprehend or as Adam sternberg wrote upon reading the times piece De Martinville day of international Glory has been delayed Sadly by the fact that his funada graph was designed to record sounds but not to play them back Making it both the world’s first and most useless recording device Actually the phonautograph was quite useful indeed a genuine technological breakthrough For years. It was used in the nascent field of acoustics for measurement experimentation and teaching as Patrick feaster is Written seeing sound was no mere half measure as a Laboratory instrument the phonograph was no less complete in function than a seismograph that could record earthquakes But not recreate them Let’s take a few minutes now to examine the most important things to know about the first invention of sound recording Edward Leon Scott De Martinville Invented sound recording when he conceived of a machine that would do for the ear. What the camera did for the eye? He first imagined an apparatus to Gather and fixed airborne sounds patterned after the human ear while editing Professor long Js. Treatise on Physiology that was in 1853 or 1854 He cites both Ears Physiology, and photography were two of Scott’s inspirations in the same way an eye or our camera fixes Rays of light Scott’s ought to focus waves of sound through a funnel or horn in the same way the ear Channels sound to the tympanum Scott sought to channel sounds to a diaphragm of rubber paper or gold beaters of skin Here the anatomical analogies have to and because Scott needed a medium unto which speech would write itself just as light wrote itself onto a photographic plate and Here he had to go beyond the analogy of photography because unlike a still camera his instrument had to record over time as Did the stigma graph which recorded the pulse of the heart on a stratum that advanced with time? The phone on a graph similarly inscribed its signals by removing a fine layer of soot from that stratum Leaving behind a white trace of its motion Now you might ask why did they use soot well the thin layer of carbon? offered minimal resistance to the rigid and Super Lightweight stylus Low Friction and low mass are essential to capturing my new details in the signal Scott made his first proofs of concept around 1854 here We see Crude tracings powered by speech on the left a guitar on the right In 1857 when he filed his patent he was still noodling with forms his instrument might take I should note that the phonograph never took a singular definitive form in Scotts hands he reconfigured it between experiments for the task at hand kind of like the modular approach taken by audio component manufacturers a century later in This form from 1857 a plate of lamp black glass Traveled underneath an inscribed instrument in this case a bird’s feather or a boars Bristle driven by the vibrations of a diaphragm Mounted to the bottom of the horn The instrument looked very much like anton stool inder’s replica on display here today Speaking singing or blasting a horn into the instrument Causes the stylus to trace squiggles in the suit for just a fraction of a second Scott left us only two phenomena. Graham’s made with an instrument like this neither of which offers recordings We can play or more specifically recordings that we can interpret when they are played Scott’s methods were unfortunately still imperfect early in 1857 Scott’s phone autograph caught the attention of the society for the advancement of National industry The Society gave Scott funds to register his initial patent and to conduct experiments with members of the society at his side members such as jewel Lisa a jew a Pioneer in the science of acoustics Lisa jew first demonstrated the figures that bear his name while working with Scott in 1857 When we began our quest for Scott’s recordings Lisa’s use dossier on the phonautograph was thought to have been lost Fortunately it was located in the Society’s archives in 2008 as a direct result of our work It preserves a number of experiments made in 1857 and shows us. How sound recording technology? developed over that year For instance here’s an experiment in which two signals were recorded at once One was made by a tuning fork writing directly to the paper another is the sound of a tuning fork of the same frequency Recorded through the funada graphs horn See where Scott or lisa shoe made marks on the paper They were counting the number of vibrations to ensure that they were the same in each recording here. They match exactly the phenomena graph passed the test it could accurately measure frequency of Airborne sounds as You might imagine the short duration of the flat recording stratum was quite limiting due hummels Vibra graph invented several Years before offered an obvious improvement It’s rotating drum would record up to twenty or thirty seconds This is the only image we have of the next-Generation funada graph it was an important Milestone because in the second half of 1857 all of the Elements essential to the recording of sound were in place let’s listen to two recordings made on that machine the first recording was made in the fall of 1857 on the top Lisa you wrote This attempt proves clearly that this apparatus is capable of studying tambor on the bottom Scott wrote sound of a deep voice held in Proximity to the membrane It’s important to remember as we listen this was not a performance It was simply an experiment to test how the instrument would perform Let me play the recording for you as we first heard it Most prominent in our initial playback is the wavering of pitch remember this was recorded on a drum Turned by hand and hand cranking is quite irregular fast in One Direction slow in the other The blue line at the bottom of this graph tracks the changing pitch resulting from the irregular cranking Modern software allows us to compensate for these irregularities and listen to an experiment made 160 Years ago what we hear is a voice changing in intensity and tambor as the recording? progresses Clearly this recording is imperfect It initially captures only the first harmonic of the voice the fundamental appears only briefly at the end however Being from October of 1857 this is the earliest interpretable Recording of the human voice we’ve been able to play as noted it’s obviously pretty distorted and it was exactly this type of distortion that scientists like Lisa who feared Late in 1857 he wrote an evaluation of a Sonata graph for the Society’s committee of economic arts His evaluation began positively acknowledging the originality and utility of scots invention the phonograph Made invisible sounds visible it fixed the fleeting echo onto paper Making sound available for study as it had never been available before Lisa’s who noted the machines ability to accurately inscribe both pitch and timbre he declared the funada graph would prove to be of Great service to science However, Lisa who wrote Mr.. Scott seeks in these traces? information of a higher order He believes his apparatus can indicate Articulation we believe on this point. He is completely in error Lisa jus based his concerns on distortions in the recording chain the movements of the Membrane the motions of the stylus each introducing mechanical distortions that Disfigure the vibration as Lisa’s who wrote as a mirror of irregular surface this figures the features it reflects Lisa’s use concerns were certainly justified We see those distortions here and we hear them in our playback of all Scott recordings but Scott’s recordings got better with time in fact we can hear in their playback the Articulations and nuances and inflection the got truly believed were in the traces Neither he nor Lisa Shue could read the articulations by eye But today we can hear them when we read the recordings by ear So here’s the crux of the problem that would frustrate scot for the rest of his days limitations in the Technology limitations in acoustical Theory and Limitations in his ability to demonstrate to the scientific establishment That there was more information in his phone autographic traces than met the eye Here’s another recording also from late 1857 it’s A test of the Funada graphs ability to capture the tambour of a cornet Will lay a spectrogram over it so you can see as well as hear the changes in pitch And Timbre well first play the recording as it comes to us from the paper Like the earlier voice recordings it suffers from unstable hand-cranking Now as we did with the voice recording let’s correct these errors in the time domain Let’s pause here for a moment to make two observations First we can clearly hear the sonic signature of a cornet So we know that in 1857 Scott was able to graph sounds with a good degree of accuracy Remember there was no way for Scott or his contemporaries to hear this for 150 years no one knew for sure if Scott had succeeded in recording sound now We know he did Because we can play his recordings and recognize what we hear the proof is in the hearing a second observation You might be wondering. How we knew this was an ascending major scale and b flat with a equal to 435 We could talk all day about this But the short answer is we have to make a lot of educated guesses for instance we know that concert 435 had been established in France months before this recording was made and That piston Cornets are typically B-flat instruments, so these decisions are quite defensible But a major ascending scale Why not a minor scale or a series of evenly spaced whole notes? The ear and forensic evidence support this interpretation as well as they support the major scale here We have a very different picture of the historical record So the observation is this the playback of a hand-cranked recording whether it was made on a phone autograph in 1857 Or on a Tin foil phonograph in 1877 Will always be mediated by interpretive choices made by the restoration engineer? The best engineers consider a slew of forensic assessments as guidance but without strict time registration at the moment of recording we face a whole series of interpretive Playback challenges in the time domain The fact is sound and time are inseparable. I noted earlier that Scott Explicitly sought to photograph the word But a snapshot of sound makes no sense Sound is perceived over time so it must be recorded over time? For the phonautograph to become a useful audio measurement tool. It needed to mark the passage of time So we were not at all surprised to find timecode burned into scots recordings from 1860 It blew our minds, but it wasn’t a total surprise Scott’s timecode actually a pilot tone in today’s parlance was written by a tuning fork adjacent to the audio recording the Fork vibrated at a constant and known frequency One could count its vibrations to determine the exact time elapsed at any point along the recording no matter how Unevenly the Crank was turned today in an exercise Undreamed-of by Scott and his contemporaries. We can play back time coded funada grams with extraordinary precision in the time domain without the need to interpret them as we do uncoated hand-cranked recordings a final observation before we hear some of these in 1860 the Funada graph was at the vanguard of instruments that recorded sensory phenomena in real time over time like stigma graphic tracings and zoetrope ik pictures For not a graphic recordings were pioneering forms of time-based instrumentation Edison’s phonograph didn’t require this precision at least not initially His invention succeeded because it odd listeners with playback while an off-speed reproduction were noted but they were forgiven because Playback was blowing their nineteenth-century minds we hope our demonstrations later this afternoon will Impart that sense of wonder and appreciation to this 21st century audience Okay, so let’s listen to a few of Scott’s most technically adept recordings All were made with a tuning fork track which relieves us from guessing about speed all were made on a machine similar to John Paul I ignored replicas on display in the origins of sound recording exhibit all are held at the academy of sciences and the library of the institute of france and All were made in 1860 the last year for which documents survived of scots experimentation This fanatic Graham was made april 17th 1868 Days after the Iconic o’clair, De La Lune, we heard earlier Again remember that every recording was an experiment Scott’s intent in this recording was to demonstrate the acoustic effects of declamation to show the inflections of Intonation the Selection from Othello is a piece Scott used repeatedly it is testing Scott’s His objective was to study the combination of the uvular R. Sound with vowels The same Day Scott recorded this phrase from Phaedra it reminds us of Scott’s Aspirations for his invention his vision was as much aesthetic as it was scientific in his writings He aspired to preserve for future generations the performances of eminent actors those grand Artists who died without leaving behind them the faintest trace of their genius? It also reminds us that Scott was indeed a man of letters note how this line consists of monosyllabic words That’s the same reason Scott’s father used it for stenographic experiments in the 1840s mmm next is an experiment to trace a vocal scale from may of 1862 out in his notes that the diaphragm he was using clearly recorded the harmonics of the principle vibration Although the next recording is undated it too is from the spring of 1860 the opening lines of Takato Tasos Aminta in Italian Intended as a study of the tonic accent meaning pitched based stress rather than volume based stress This recording serves as something of a rosetta stone At the bottom Scott notes. I was wrong. It should be ooh money for me not for me. Ooh money as written above mole Before we heard this we didn’t know what Scott was wrong about Was he referring to the transposed words written on the page or to the words spoken in the recording? Once we heard the recording we knew he transposed both Admitting it was his mistake saying I was wrong is Tantamount to Scott identifying himself as the speaker and as you’ve heard the voice in this recording is the same as in all other recordings from 1860 the Lesson here of course is that we must listen to these recordings if we hope to rewrite the history books? correctly a Fellow Phaedra Aminta The materials Scott chose for his recitations revealed to us a man of letters in mid 19th century Paris This is also true for the songs he chose Here he sings a melody from Cherubi Knees meesa solemnity Finally we arrived at Scott’s last funada Graham, or at least the last we’ve located so far in this recording from late 1860’s Cought sings a lively rendition of song of the bee It’s one of his most melodic Joyful and playful recordings Unlike most others refer today it strikes me as more than just an experiment Like the Cherubini this may be among the very first recorded performances So as we rewrite the history books here are a few key sentences for your consideration Edward Leon Scott De Martinville Invented sound recording in France 20 years before thomas Edison reinvented it in the United States when Scott built his first successful funada graph in 1857 He became the first person to capture his voice for future generations to hear The telephone was the first instrument to send voices over long distances But the funada graph was the first instrument to send voices into the future The recordings Scott sent to us are humanity’s first recordings of its own voice since we first introduced them nine years ago mon ami pierrot has taken his place next to Mary’s lamb as a universal icon of the invention of sound recording Appreciation of Edward Leon Scott De Martinville and his accomplishments is spreading – in 2011 the Library of Congress inducted Scott’s recordings onto its national recording registry a juried list of sound recordings judged to be culturally historically or aesthetically important and worthy of perpetual preservation in 2015 Unesco installed Humanity’s first recordings of its own voice onto its memory of the world register an international Initiative that helps safeguard documentary heritage against collective Amnesia and the effects of time We can debate who took the first photograph Who made the first phone call or who captured the first moving images? But there is no debate about who made the first sound recordings It’s rare to know something with such certainty Thank you you

Tags:, ,

Add a Comment

Your email address will not be published. Required fields are marked *