July 23, 2012 | 1
If you happened to be a servant in a fine house in Boston, circa 1890, you might have spent a lot of time and excess energy running up and down the stairs between the kitchen and the second floor. Unless, that is, you happened to be a scullery maid in a classic shingle-style home designed by Henry Walker Hartwell and William Cummings Richardson, who saw fit to install a built-in speaking tube between the pantry and the second floor corridor, along with an electrical buzzer system. Considering that Alexander Graham Bell’s patent for a prototype telephone wasn’t issued until 1877, Mssrs. Hartwell and Cummings were on the technological cutting edge.
That speaking tube is still there, and drew the curiosity of acoustician William Elliott, who conducted a series of measurements to determine the quality of the sound transmission, in terms of intelligible human speech. It’s not enough for a scientist to know that something works well; he or she wants to know why it works so well, in this case, why speaking tube systems manage to transmit intelligible speech. Elliott talked about his results at last fall’s meeting of the Acoustical Society of America — and yes, I am only now getting around to blogging about this, months later. It’s been a busy year.
Speaking tubes date back to around 1849, when an article in Scientific American described an “acoustic telegraph” that would enable people to converse with friends “as far as 60 miles away” (!) via a tube made of gutta percha (a latex material derived from trees in Southeast Asia). That proved to be a bit ambitious, but the article also noted that such devices would be extremely useful for communication within factories, foundries and other public buildings.
There were, indeed, many patents issued for various components of speaking tube systems between 1860 and 1890, usually for communication within a single building. And an Italian immigrant named Antonio Meucci, whom many credit with inventing the telephone before Bell, also built an acoustic speaking tube system in his home, similar to the pipes used for communication on ships.
In the early 18oos, Jean-Baptiste Bio had experimented with how sound travels through long tubes, using the water pipes of Paris, and found that the confines of the piping served to keep speech intelligible over a good 1040 yards, compared to how well sound carried in free space. Increase the diameter of those pipes, however, and there would be a corresponding decrease in intelligibility.
So Elliott had some solid science to draw upon when he conducted his own measurements of the speaking tube system in the Cambridge home, part of an electric system for giving household staff the heads-up.
Each end of the tube — one just outside the kitchen pantry, the other in the second floor corridor — is covered by a whistle valve. If someone wanted to communicate, he or she would open the valve and blow through the tube, producing a sound like a whistling tea kettle. Then whoever was on the other end would know to open their valve as well, and the two parties could have a spoken conversation through the tube. Oh, and both ends had flared openings, the better to hear the other party. It is, after all, a waveguide, one that in this case guides sound waves.
Elliott was interested in the question of why speech was intelligible through such speaking tube system, so he set up a small computer speaker near one valve opening to transmit “pink noise” — in which every octave contains equal sound energy, thereby making it a useful acoustic measurement tool — with a microphone at the other end as the receiver. He then measured how well sound traveled through the tube at select specific frequencies, and compared that data to the known range of frequencies deemed ideal for intelligible human speech.
His conclusions: the transmitted signal fell within the 200 hertz to 5000 hertz frequency range, which just so happens to also be an excellent range of frequencies for speech intelligibility. The tube’s geometry — small in diameter, with rigid metal walls — served as an excellent acoustic filer, since those mid-range frequencies could travel through the tube far more efficiently than very high or very low frequency sounds. That’s useful information to have on hand, should anyone feel inclined to design a more modern communication system based on this concept.
Back in the 19th century, it was all part of the communications revolution that began with the invention of the telegraph and the telephone. A telephone’s transmitter contains both a wire coil and a small magnet. Speaking into the transmitter causes the coil to vibrate in response to the sound waves within a magnetic field. This turns the sound wave into an electrical signal, which can be transmitted over the telephone wire. That current is detected by the receiver’s coil, producing a second magnetic field. And this causes a thin membrane, similar to the human eardrum, to vibrate in response to the electrical signal, turning it back into sound
Once transmission and reception of sound waves had been demonstrated, inventors began thinking of ways in which they might record and reproduce that sound — inventors like Thomas Edison, who wanted to design a machine capable of transcribing telegraphic messages by using a stylus to make indentations on a paper sheet in response to sound vibrations.
Edison achieved a similar effect with a steel needle, placed against a strip of waxed paper. Then he shouted “Hallo! Hallo!” and watched as the vibrations moved the needed, scratching a pattern in the waxed paper. Even better, when he later passed the needle over those marks, he heard a faint but audible “Hallo!” in his own voice.
That was proof of principle, and Edison asked his mechanic, John Kreusi, to build an actual device, based on his design. It used a stylus to prick a pattern in a tinfoil cylinder in response to sound vibrations of, say, somebody’s voice. A second needle would play the sound back as the patterned cylinder turned, via a hand crank, with a handy amplifying horn to enhance the intelligibility of the playback.
In 1877 (either August or December; accounts differ), Edison recorded the phrase “Mary had a little lamb” and successfully played it back on his phonograph, even lugging the device over to the Scientific American offices in New York City for a demonstration. It netted him a nit of media coverage. Per the December 22 issue:
“Mr. Thomas A. Edison recently came into this office, placed a little machine onto our desk, turned a crank, and the machine inquired as to our health, asked how we liked the phonograph, informed us that it was very well, and bid us a cordial good night.”
Edison was granted a patent for the phonograph in February 1878 and founded a company to market the device. (French scientist Charles Cros also designed a similar device, described in a paper written in April 1877, but Cros never built a working model.) He later improved that early prototype, and while its initial intended commercial function — replacing stenographers — never caught on, its use for entertainment was undeniable. The Edison Phonograph Works soon started building early prototypes of jukeboxes, and a sibling company manufactured talking dolls that contained tiny wax cylinders for recording and playback — rare collectibles today.
In 1938, composer Richard Strauss combined these two technologies (speaking tube and phonograph) to monitor the gate to his lavish estate. Visitors who range the bell would trigger a phonograph some 50 yards away. The phonograph would play back a pre-recorded message (“Dr. Strauss is not at home”) through the tube to dissuade unwanted visitors. Those who knew Strauss well would ring a second time — this would stop the record and the gate would open. It was the very first answering machine/buzzer entry system.
Looking back on those first phonographs and recordings, the technology seems very crude. But the principles haven’t changed. Even today, old-fashioned LPs rely on variations of the grooves in the vinyl to encode information, which is “decoded” by the tip of a phonograph needle. CDs operate on much the same principle, only they encode data in “pits’ instead of grooves, and the data is decoded by laser light.
In fact, we get most of our sensory information from variations in background: we can read words in a book because the ink encodes meaning onto the paper in a recognizable pattern. A German artist named Bartholomaus Traubeck even created a modified record player that analyzes the growth rings on a cross section of a tree and translates the information into piano music (watch video here).
Today, most of us likely listen to music in the form of digital mp3 files loaded onto an iPod, smart phone, or equivalent device. It’s easy to forget obsolete technologies like 8-track and cassette tapes, old 45 singles, and vinyl LPs. I like the digital music age, but a team of acoustic engineers at the University of Texas, Austin, think we may have lost something precious along the way: an understanding of how information, like sound, is captured and relayed.
“[I]t is increasingly common to find young students who know nothing other than digitally recorded sound, with correspondingly little understanding as to how sound is captured and replayed using their digital devices,” the researchers wrote in a lay language paper of the talk they presented at last fall’s ASA meeting (yep, still catching up). Students like sci-fi author John Scalzi’s daughter, Athena, who went along with this good-humored exchange over encountering an old vinyl LP for the first time, much to the amusement of Internetizens (no, skeptics, she didn’t fake it):
That’s why said researchers — Jason Sagers, Andrew McNeese and Preston Wilson — collaborated on a project to recreate Edison’s original purely mechanical design for the phonograph. They say their recreation “serves as a useful tool for teaching acoustics and providing a hands-on demonstration to people of all ages.” (You can see a video of the replica in action here.)
But recreating the phonograph did present a unique scientific opportunity: they were able to compare Edison’s original 1877 foil recording with their own 21st century version, and were intrigued that regularly spaced dimpled patterns appeared at certain times during the recording process. They hypothesized that “the phonograph does not record all frequencies equally, but tends to prefer particular frequencies.”
Being good scientists, they tested that hypothesis and constructed a mathematical model to predict the acoustical behavior of the device, based on circuit modeling (the electrical circuit elements were replaced by mechanical and acoustic masses, springs, and so forth). They also used a scanning laser Doppler vibrometer (SLDV) to non-invasively measure the vibrations when the phonograph was playing, so they had a data set to compare to the model’s predictions.
As the researchers reported in their ASA paper:
“We found that the most important piece affecting the acoustic behavior of the entire system was the horn, which amplifies the sound as it travels into the phonograph during recording and out of the phonograph during playback. The horn experienced axial acoustic resonances (i.e. the horn amplified certain frequencies more than others) which caused the regularly spaced dimples … to appear in the foil.
The second most important component of the phonograph was the mouthpiece. The mouthpiece contains a small cylindrical hole which is backed by a narrow air cavity before being terminated by the diaphragm. The hole and cavity combine to produce an additional acoustic resonance, which enhances the range of frequencies that the phonograph can record.”
Maybe it’s not as high-fidelity as the mp3s we use today, but without Edison’s phonograph and related technology, we would never have developed the digital technology to create them. Science builds on everything that came before, which is why I love writing about the history so much — lest we forget.
“The talking phonograph,” Scientific American, Vol. 37, No. 25, pp. 384, December 1877.
“Edison’s improved phonograph,” Scientific American, Vol. 57, No. 21, pp. 238, November 1887.