Humanity's First Recordings of its Own Voice - this Saturday

[[email protected]]

Thomas Edison NHP News Release

Contact: Karen Sloat-Olsen
Phone: 973-736-0550 x17
Reservations: 973-736-0550 x89

Humanity's First Recordings of its Own Voice
Historian David Giovannoni Presentation

WEST ORANGE, NJ - On Saturday evening, November 6, 2010, at 7:00 pm, Thomas Edison National Historical Park welcomes historian David Giovannoni who will give a 75-minute illustrated presentation titled ???Humanity???s First Recordings of its Own Voice.??? The program will be held at the Laboratory Complex at 211 Main Street. Admission to the program is free. Seating is limited and reservations are required. Reservations can be made by calling 973-736-0550, ext.89.

Thomas Edison???s tinfoil phonograph of 1877 is rightly considered one of the marvels of the nineteenth century. But in mid-nineteenth-century France, amateur inventor ??douard-L??on Scott de Martinville conceived of a rather similar machine. Between 1854 and 1860 he experimented with focusing airborne sounds of speech and music onto paper. His phonautograph bore a striking resemblance to Edison???s phonograph of 20 years later. But his recordings, unlike Edison???s, were meant to be read by the eye, not heard by the ear.

For a century-and-a-half his experiments lay quietly in the venerable French archives in which he deposited them. Then in 2007 a few audio historians hypothesized there was a real possibility that modern technology could develop these experimental recordings like dormant photographic plates. Instead of exposing images, however, these would bear sounds ??? perhaps even humanity???s first recordings of its own voice!

In this presentation David Giovannoni recounts how he and his colleagues have identified dozens of these forgotten documents and coaxed several to talk and to sing. A principal in their discovery and recovery, Giovannoni is the first person since Scott de Martinville to personally examine every recording. He???ll explain how they were made and how they are played. He???ll discuss Scott de Martinville experiments, his reception in established scientific circles, and his early descent into an unmarked grave.

For more information or directions please call 973-736-0550 ext. 11 or visit our website at www.nps.gov/edis.

-NPS-

National Park Service
U.S. Department of the Interior
Thomas Edison National Historical Park
211 Main Street
West Orange, NJ 07052

[hillndalefan]

I have a 45 rpm record of this recording which I got from Kurt Nauck. It's fascinating to hear Leon hum "Sur le Pont d'Avignon" in 1866. The visual track traced in lampblack soot was transferred [somehow] to a groove or digital file and reproduced. If there are others, it would be interesting to hear, even though they are only laboratory experiments.

[Swing Band Heaven]

I remember reading about these lamp black recordings years ago and wondering if there would be a way of playing them back with modern technology. Its fantastic that this has now been realised. I should think that a discussion on how it has been done would be fasinating - too me at least.
S-B-H

[TinfoilPhono]

This has actually had a lot of press since the first "digitized" phonautogram was unveiled in 2008. It was even front page news in the NY Times at the time. You can find the full story, and many sound clips, at www.firstsounds.org.

[phonophan79]

I just got back from this event... and what a wonderful, informative and enjoyable presentation it was! I will try to re-cap as best as I can, many apologies if I get anything wrong.

Unfortunately the request was made that there be no video taping, so I have nothing to show.. but Mr. Giovannoni did an excellent job in telling the story of the work done with his fellow researchers to tell the story of Édouard-Léon Scott de Martinville, even working with Scott's great grandson.

As many of you know and have heard, Scott recorded the French folksong "Au Clair de la Lune" on April 9, 1860... becoming the earliest recorded human voice.

The release of the original 2008 recording (as can be heard at http://www.firstsounds.org/sounds/scott.php labeled Au Clair de la Lune Original March 2008 release) made international headlines with the voice of a french woman breaking through the barriers of time with the mysterious and eerie French folk song.

It was very interesting to learn that the recording was... as rookies in the phono hobby often do ... played back at the wrong speed!

So how did they learn that their original release was played back at the incorrect speed? Well... Scott used a tuning fork during the phonautogram recordings as a constant, a reference. Obviously, the manual hand-crank of the phonautogram recording is not consistent due to simple human capability.

So when the researchers working with the visual scan software developed at Berkeley... they were scanning the visual representations of sound made by the phonautograph. The tuning fork's constant and unchanging tone / frequency was separated into its own "track". The tuning fork became 19th century "timecode".

Mr. Giovannoni played an audio clip of the tuning fork pitch as recorded "by hand". It had a consistent "human error" wave, a change in pitch, up and down, up and down. The software compensated for the human error, and then played a separate audio clip with a constant pitch.

What science today refers to as one frequency cycle, in 1850 could have been measured (by today's standards) as half a cycle or even a quarter of a cycle. So when the software took the original 500Hz "Au Clair de la Lune" recording and slowed it down to the true speed of 250Hz, it becomes the voice of a man. Not only the voice of a man, but the voice of the inventor - Édouard-Léon Scott performing an experiment watching the different frequencies and waves from his phonautograph as he sang long and exaggerated tones. (also on the page http://www.firstsounds.org/sounds/scott.php labeled Au Clair de la Lune May 2009 restoration)

There were also two world premier's at this event. Mr. Giovannoni's presentation included the story of his trips to Paris and establishing connections and acquiring permission for research and eventual scanning of the phonautogram recordings. He eventually found himself discovering the original French patent submission paperwork by Scott himself for his phonautograph and the discovery was made for the first ever sound recording in 1857, 3 years earlier than Au Clair de la Lune. The recording was that of a cornet player playing a simple scale of notes.

So while it is now the earliest sound recording ever... Au Clair still remains the earliest human sound recording ever.

The second world premier was a treat in that his colleague sent him an e-mail at 7pm, right when the presentation was about to start. The e-mail included an audio file of, again the inventor himself, Scott reading a potion of Shakespeare's Othello in his native French tongue. This recording is estimated to be only a few weeks after Au Clair... and again an experiment of inventor with his invention.

Mr. Giovannoni expressed that Scott was adamant that his phonautograph recordings contained the information of speech, of articulation... thet they could give a glimpse into the soul of the speaker, male, female, emotions, etc. While playback was never considered by Scott, the French Academy of Science was doubtful of Scott's claim. The phonautogram recordings were seen as a visual representation. That perhaps a caligraphy-frequency or some form of hieroglyph... in where the eye could look at these waves and transliterate it into words and text.

During question and answer time, the question was raised as to whether Edison was aware of Scott's work. It was Mr. Giovannoni's opinion that Edison did not know of Scott's work; that Edison came to invent the phonograph via his experiments with the telephone - and that Edison had gone off on all sort of tangents from the telephone before arriving at the phonograph. So if Edison had known of Scott's work, he certainly would have arrived at the phonograph sooner.

[Starkton]

Let us give credit where credit is due. Mr. Giovannoni's fellow researchers not mentioned in your excellent overview are: Patrick Feaster, Meagan Hennessey, Richard Martin and Rene Rondeau.

Edison had subscribed to many dozens of technical journals worldwide. Of course he knew the phonautograph, just like anybody else with a broad general education, but he had a completey different approach when he conceived the phonograph. Besides, Edison's opinion in relation to the phonautograph was low. On March 2, 1878 he wrote „The plan you suggest [has been] in use for many years in physical laboratories for investigating the larynx. It is called Scotts Phonautograph but it is almost useless as it never gives a true record and the same word may [be] imprinted [?] 100 times and no one will be alike. The amplitude of the vibrations due to the hissing consonants are so small as to be almost invisible [...].“

Nevertheless, during July and August 1878 Edison worked with a modified phonautograph and employed it during his experiments to identify the cause of the noise on the Metropolitan elevated railroad. Quite a number of phonautograms resulted from the experiments and are still stored among the documents in the Edison collection. Experiments to make them audible finally led to the "deciphering" of the Leon Scott phonautograms.

By the way, the "earliest sound recordings ever" are the "acoustic figures" of Ernst Florens Friedrich Chladni, first published in 1787 ("Discoveries in the Theory of Sound").

[frenchmarky]

What's coolest to me about all this is that the French guy made the recordings without even the slightest notion that these would or could ever be played back.
There wasn't even the concept of recording in his mind, it was just "hey look, when I talk into this thing, it makes squiggles!" I'm actually surprised that it took this long to figure out how to play these things back, there must be many things about these 'squiggles' that make them different in nature than a groove in a record.

[Starkton]

it was just "hey look, when I talk into this thing, it makes squiggles!"

This is by no means true.

Édouard-Léon Scott conceived of the phonautograph as early as 1853 while he corrected proofs for a book in which the author, a French physiologist described the mechanism of the ear and asked the question, is it not possible to record an impression of what the ear hears? Scott immediately saw that if one wanted to create an “image” of sound, one should build a replica of the inner ear and connect this contraption to some form of inscription device.” The “sublime artist, God,” led him to his goal by displaying “the marvel of all marvels, the human ear."

Scott was interested in the physics of sound, had studied under Regnault at the College de France, and knew other notable scientists who did research in acoustics. Scott, who had been a typesetter, wanted to reform stenography so as to record thoughts more efficiently in a natural, permanent form. He had been influenced by the idea of photography and wanted to create a similar record for sound. “Sound, just like light,” he wrote, “can provide a lasting image at a distance.” Scott wished to create an apparatus that would inscribe nature's own language of sound, a universal language unencumbered by conventions, and connected directly to the physical production of speech.

(The above is loosely cited from the unpublished dissertation of David Alexander Pantalony, Rudolph Koenig (1832-1901), Hermann von Helmholtz (1821-1894) and the Birth of Modern Acoustics, Toronto 2002)

What surprised me much more is that no French expert had taken care of these recordings. It was Patrick Feaster, a researcher from the US who first came up with the idea to make Scott's phonautograms audible.

[noedison]

Absolutely amazing endevor and wonderful that the tracings could be converted back to sound. I'm sure that we have not heard the final transcription of these as perhaps in a few years there is an even better technology that could further improve the conversion?

[phonophan79]

Let us give credit where credit is due. Mr. Giovannoni's fellow researchers not mentioned in your excellent overview are: Patrick Feaster, Meagan Hennessey, Richard Martin and Rene Rondeau.

Yes, indeed! Thanks for this. It was indeed Mr. Feaster who sent Mr. Giovannoni the new audio file mentioned in my re-cap of Scott reading Othello. Mr. Giovannoni used his iPhone to call Mr. Feaster at the end of his presentation. It was one of the lighter moments when Mr. Giovannoni turned around and picked up the narrow handheld horn that accompanied Mr. Edison's tinfol phonograph reproduction. He held the horn to his iPhone speaker... and it did indeed help the audience hear Mr. Feaster better through the iPhone speaker!

Edison had subscribed to many dozens of technical journals worldwide. Of course he knew the phonautograph, just like anybody else with a broad general education, but he had a completey different approach when he conceived the phonograph.

It really is quite amazing how long it took mankind to realized that sound could not only be recorded but played back. It was mentioned during the presentation that Alexander Graham Bell once said that he had the invention of the phonograph "slip through his fingers" via his telephone.

During July and August 1878 Edison worked with a modified phonautograph and employed it during his experiments to identify the cause of the noise on the Metropolitan elevated railroad.

Yes! This was a portion of his presentation that I did not mention. A section of Mr. Giovannoni's presentation was titled something like "The Quest for sound during the silent decade"... that is after Edison's invention of the tinfoil phonograph in 1877... there is a span of about 10 years in which sound recording is "silent"... mostly due to the fact that Edison's tinfoil phonograph recordings WERE meant to be played back. Since they were actually played back, the lost a huge amount of its fidelity.. since every tinfoil playback caused great wear to the grooves.

...but Edison did not use tinfoil during the Metropolitan elevated railroad tests, he used paper instead. It was said Edison wanted visual representation of the sound for the experiments. ...and in fact one of the very first audio clips Mr. Giovannoni's played was one of Edison's Metropolitan elevated railroad "phonautogram" brought back to life! Mr. Giovannoni did mention that there are other Metropolitan elevated railroad recordings that will be image scanned. As a side note, he said that it is commonly known that Edison liked to use colorful language... these railroad recordings might possibly have some never heard sound of Edison's voice or the previously unheard sounds of New York City over 100 years ago... only time will tell.

Going back to the tinfol issue for a moment... Mr. Giovannoni explained that the image scan software was not designed to work with the tinfol because the software needs the light to be absorbed and not reflected. The tinfol would overload the sensors... and what is needed to bring the tinfols back to life would be un-invasive 3D image scanning... which the technology is beginning to come together and become available.