Ancient Sound recordings of historical personalities

[Dancefortwo]

This is probably a sound recording of Queen Victoria's voice. It was known that at one time her voice was recorded by this new technology so tracking down the cylinder it was recorded on wasn't easy.  Through a lot of detective work by several people - this is probably her voice.  Be warned, it's very difficult to hear and it's sounds all mangled up because it is, after all, from 150 years ago. 

 

It's weird to think about but we're hearing the voice of someone born 200 years ago.   Napoleon and Thomas Jefferson were alive when she was born. 

If you're interested here's the account of the detailed search for the cylinder sound impression of her voice in the Smithsonian magazine.


https://www.smithsonianmag.com/history/i...-98809025/

[abaris]

Austro-Hungarian emperor Franz Josef in 1903, subtitled. Still bad quality, but quite understandable. Love the soft Viennese accent of the turn of the century upper classes. If you read the subtitles, he's commending the telephone, but in truth he hated it. The only one in his private rooms had to be on the toilet, so that he didn't have to see it whenever he walked by.

[no one]

Voices carry.

[Bucky Ball]

Amazing ... my great granny was 3 years old then.

[abaris]

Renamed in accordance with @Dancefortwo. It's now a collection of old sound recordings of prominent persons of ages past. So fire away, if you find any.

[Deesse23]

Helmuth von Moltke

[abaris]

Otto von Bismarck. Also recorded by Edison when he was in Europe. Date is 1889.

[abaris]

The voice of Pjotr Iljitsch Tschaikowski, dated to the 1890s. It's in Russian, of course and the singers, if we haven't raised standards siginificantly since then, are below average.

[Dancefortwo]

Thanks for changing the title of the thread.  Great idea.  It opens it up to a wider theme.  

So the first recording of the human voice was on April 9th 1860.  Yes, 1860.

A Frenchman named Edouard Leon Scott de Martinville, was trying to find a way to photograph the vibrations of the human voice so each vibration could be transcribed into letters and then into text.  He came up with a aparatis he called a "phonautograph".  The barrel was made out of plaster of Paris, not exactly high tech but it worked.

 

He took a stiff boars bristle and a piece of glass coated with soot and recorded it and it looked like this.



The flat horizontal lines couldn't be heard or played back until 150 years later when an audio historian realized that computers could be used to hear the scratches on the soot.  There's a whole thing about how the computers were programmed to convert the scratches to sound that I don't understand but anyway, here's what the computers found.  It's the sound, proabaly of Scott de Martinville himself, singing the French song, "Au Clair de la Lune".




I like that someone in the comment sections says, "So much better than today's music."  Everyone's a comic.   

[Minimalist]

Sound recording has come a long way.

[abaris]

Sound recording has come a long way.

They may have been not as bad as they sound today when they were originally recorded. Truth is, it probably took many decades, till someone rediscovered and transfered them to a more modern carrier. They had plenty of time to deteriorate, seeing as the first recordings were on wax roles.

[Cavebear]

This is probably a sound recording of Queen Victoria's voice. It was known that at one time her voice was recorded by this new technology so tracking down the cylinder it was recorded on wasn't easy.  Through a lot of detective work by several people - this is probably her voice.  Be warned, it's very difficult to hear and it's sounds all mangled up because it is, after all, from 150 years ago. 

 

It's weird to think about but we're hearing the voice of someone born 200 years ago.   Napoleon and Thomas Jefferson were alive when she was born. 

If you're interested here's the account of the detailed search for the cylinder sound impression of her voice in the Smithsonian magazine.


https://www.smithsonianmag.com/history/i...-98809025/

As poor the recording was, it is impressive to hear a famous voice from the past. Thank you!

[airportkid]

In the science fiction short story "Farewell to the Master" (made into the film "The Day the Earth Stood Still") an alien accompanied by a robot lands at the US capital, and the alien is instantly shot dead after announcing his name and that he has a message.  The robot recovers the alien's body and tries to resurrect it from a sound recording of the alien's brief greeting, but the resurrections quickly die because the sound recording equipment was imperfect.  The robot then analyzes the recording equipment to identify its imperfections, and is able to finally successfully resurrect the alien.

It seems to me that it should be possible to apply the same principle to old sound recordings and derive a very close reproduction of the original sound.  So far it seems that noise is more or less successfully eliminated, but the distortions caused by the nature of the recording remain.  For example, many old recordings sound as if they were underwater, or muffled.  The aspects of the recording equipment that made that distortion might be measurable so they can be "reversed", and make the recovered sound identical to the original.

[Dancefortwo]

For those who are interested this is a fabulous documentary on the detective work a group of audio scientists did to find the first recorded sounds.  It is really worth the hour spent understanding how they took the tracings on paper covered with soot and tranfered it to sound.  




After watching this it seems that the very earliest recorded sound is from 1857, not 1860.   That's three years before the other recording of Scott singing Au Clair de la Lune.   It's the sound of a trumpet, perhaps played by Scott de Martinville, doing the scales into the phonautograph.  

@airportkid , you might be interested in this.   This is really cool stuff.

[abaris]

It seems to me that it should be possible to apply the same principle to old sound recordings and derive a very close reproduction of the original sound.  So far it seems that noise is more or less successfully eliminated, but the distortions caused by the nature of the recording remain.  For example, many old recordings sound as if they were underwater, or muffled.  The aspects of the recording equipment that made that distortion might be measurable so they can be "reversed", and make the recovered sound identical to the original.

You see what's lying next to the phonograph?



It's wax cylinders. That was the medium they recorded on. The Moltke and the Bismarck records were done by Edison on a European voyage, using that kind of technology. He travelled for thousands of miles carrying wax cylinders with him. They probably already deterioated when he came home to his workshop. And there they stayed, probably for decades, before they were transfered to a more lasting carrier.

While it is entirely possible to filter out the noise, you can't filter out the damage.

[abaris]

This, though only 20 to 25 years later, illustrate what I mean.

It is Charles I, last emperor of Austria and last king of Hungary and Bohemia. It's from 1915, when his uncle was still alive and he was still heir to the throne he took in november of 1916, after the emperor of 68 years finally died.

It's already much clearer than the previous recordings, since it was no longer taken on wx but an actual record. There is noise, yes, but it's nothing as compared to the olders recordings. He talks about having visited all the fronts and having seen the damage and the suffering. Which is true, since Charles was known for going there, before and after he took the throne. He also summons the unity of the front and the homefront and closes with the motto his uncle gave his reign: Viribus Unitis, which means, with combined forces.



The reason why all these speakers sound so woodenly is that they didn't give a speech. They were asked to repeat a speech they gave, in front of a recording device and the team recording, advised the speakers to pronounce as clear as possible, since they knew about the limitations of the relatively new technology.

[airportkid]

... you can't filter out the damage ...

That's not what I was talking about.  Transducing sound into an electrical signal and back into sound with early technology would do the equivalent, say, of making A sharp always sound like F flat, except, of course, the distortions would be many more than that.  I'm suggesting that those properties are identifiable and measurable, and would permit "redistorting" the signal back to close to what it originally was as an acoustic signal.  I haven't looked at what @Dancefortwo  put up yet and maybe what I'm referring to is already being addressed.  But I've never seen it.

There's no replacing what damage removed directly, but some damage could be repaired by contextual interpolation.  Say two bars of Toscaninni's March 8 1953 performance of Beethoven's 5th at Carnegie Hall are irreparably destroyed.  The tempo and style of the whole symphony could be applied to the missing bars such that any departure from the original would be negligible.

Today I'm sure we have the ability to make a film starring John Wayne at age 30 alongside Dallas Bryce Howard with a CGI John Wayne indistinguishable from the real John Wayne in manner and voice.  The only thing preventing it is estate copyright.  And sensitivity to good taste, but that's another argument.