What do you think? Leave a respectful comment.

Digital Technology Helps Researchers Hear Earliest Recordings Better

The recording is just 78 seconds long, featuring a cornet solo and a man reciting nursery rhymes. Dated back to 1878, experts say it may be the oldest playable recording of an American voice. Ray Suarez talks to Lawrence Berkeley National Laboratory’s Carl Haber who helped uncover the significance of this tiny piece of tin foil.

Read the Full Transcript


    Finally tonight: the amazing story of how modern digital technology opened up a window into the beginnings of recorded sound.

    Ray Suarez has our look.


    The sound is just 78 seconds long. It features a cornet solo and a man reciting nursery rhymes, including "Mary Had a Little Lamb." Experts say they have reproduced the sound of the oldest playable recording of an American voice and the first captured musical performance.

    It's a recording made in 1878 on a small sheet of tinfoil, then placed on the cylinder of a phonograph invented by Thomas Edison. A hand crank turned a stylus that moved on the foil, recording sound. The foil was donated years ago to a museum in Schenectady, New York, but its significance wasn't appreciated until this summer, when it was brought to researchers in Berkeley.

    The foil was so fragile it could not be touched. Instead, it was scanned by computer to read the grooves in the foil and create a program to recreate the original sound; 134 years later, it's a little indistinct, a little hard to make out.

    Here's some of what they heard of the cornet.



    For more on the significance of this work and how it was done, we turn to one of the primary experts involved.

    Carl Haber is with the Lawrence Berkeley National Laboratory. He's in New York state, where the recording will be played for the public for the first time tonight in more than 100 years at the Museum of Innovation and Science.

    And, Carl Haber, where's the foil been in the ensuing hundred years since it was last played?

  • CARL HABER, Lawrence Berkeley National Laboratory:

    Well, in the recent years, I think, since about the 1970s, it's been in the collection of the museum.

    Before that, I believe it was in the hands of some private collectors, who I think passed it on to the family. And then it was finally given to the museum, as you said, some time in the 1970s.


    When you first got in the your hands and started to figure out how to hear what was on it, did you already have an idea of what had been recorded, a cornet, a human voice, that sort of thing? Or was it really sleuthing that started from square one?


    Well, we had no information directly about what was on this particular recording, but we have done a number of projects on early recordings from the 1800s.

    And there are certain things that typically show up, like people will do "Mary Had a Little Lamb," because that was kind of iconic at the time, and other such things. But we — it could have been anything, so we didn't really know what it was going to be.


    The machine that was meant to play this foil back is now long gone. Tell us how you reverse-engineered this, so once again we could hear what was on this 19th century recording.



    Well, the machines are still around in antique collections and in museums and stuff. The problem is foil is very delicate and would essentially not withstand an attempt to play it back. And this particular foil was torn in many places as well.

    So, at the Lawrence Berkeley National Lab, we have been doing a project over the past few years, in close collaboration with the Library of College, to develop a set of optimal methods that would allow us to play back essentially any early recording without touching it.

    So, if the recording is torn, broken, too delicate to touch, that aspect would not become — become necessarily a problem. So, what — the idea of restoring such a thing, and particularly this tinfoil, was that we have a kind of camera. It's a scientific camera that has very, very high resolution and can collect information like a normal camera, but it can also measure the third dimension. It can tell us the up and down, the depth of the object that we're photographing as well.

    And we trained that camera over the surface of this foil. And we took large numbers of images that, together, were probably 4,000 megapixels or something of that scale, a very, very large image. But it was large enough of that it contained in minute detail the undulations, the movements of the groove that Edison's machine had cut into this foil.

    Once we had the image, we were able to put it on a computer, and, as you say, reverse-engineer, essentially ask the question mathematically, what would the needle have done to create this shape or what would a needle have done to respond to the shape if it had been drawn through it?

    And from there, it's actually a very simple set of physics to get you to what the sound would have been that they recorded or would have been heard in the process of playing it back. And it's actually a very direct process once you can properly characterize the surface.


    Now that you have done that, do you feel pretty confident that what you're going to hear at the museum tonight, what we played back during the show this evening, is pretty much what people would have heard in 1878?


    I think what people would have heard in 1878 would have been a little clearer, because if you have seen the picture of the foil, it was folded seven times, and then stored in an envelope for 140 years or whatever.

    So those folds — and perhaps it had been unfolded a number of times, because the folds were very ragged and had actual fissures that ran along through the folds. And so that, unfortunately, gives a very characteristic sort of thump, thump, thump, thump sound that you hear when you play it. Every time you cross one of those folds, you hear a noise.

    When it was first played back in Saint Louis in 1878, I presume that wouldn't have been there.


    Do we know who the people are, the person playing the cornet, the people saying the nursery rhymes?


    We don't know who was playing the cornet, but Chris Hunter, who's the curator at miSci in Schenectady, has done a lot of research.

    And he looked through the microfilms, and he found the Saint Louis newspapers from that time announcing the exposition of this device. And it was attributed to a man named Thomas Mason, who was actually a journalist. And he used the pen name "I.X. Peck," so, I expect.

    He was apparently a humor writer. And it is recorded in the records that he purchased this machine and that he made an exposition of it in June of 1878. So, our closest guess is that he is probably the person speaking.


    You know, all kinds of machines, devices have been made to capture and reproduce sounds, pictures.

    And, yet, we can't listen to our view many of those images today. Do we have to be careful that we preserve every old scrap of sound we have got, create a sound library, if you will, so that a century from now people will be able to listen to, for instance, this conversation?


    I think that it's very important to create stable archives of the information that we create as a culture.

    You can debate as the amount of information gets larger and larger how best to do that, but that's a question of, you know, what do we do to preserve what's being created today and in the future, and also what do we do to preserve that which was created in the past?

    For me, I can understand a little bit better about the past, because that's mostly what I have been thinking about. And I think it's tremendously important to preserve the images, the moving pictures and the recorded sounds of the past.

    Early sound is something they're not making any more of. And it gives us a really significant and important window on our history as a culture, as inventors, as innovators, as scientists, as researchers about what people were doing, the insights they had, which really a lot of it underlies the information and communication age that we live in today.

    So, I strongly support the preservation of recorded information.


    Well, in 2136, when people watch this, they will conclude beards were much more popular than they actually were.

    Carl Haber, thanks for your time.


    Thank you very much.

The Latest