by Trevor Cox
But adding a delay is ineffective if the loudspeakers are turned up too high, because the precedence effect can be overruled by a loud sound that arrives later—a situation that prevails at most rock concerts. Without electronic amplification, though, reflections from walls are usually too quiet to cause a problem. But in the case of the Mapparium or Teufelsberg, where the dome’s focus greatly amplifies the delayed sound, the reflections are so strong that we are fooled by false localization. When I burst my balloons in Teufelsberg, the first reflection from the ceiling was 11 decibels louder than the sound coming directly from the balloon (Figure 5.6). (A useful guide is that an increase in 10 decibels is roughly a doubling in perceived loudness.) When I knelt down to unzip my knapsack, it sounded like someone was opening the bag above my head!
Figure 5.6 Direct sound and reflections from bursting a balloon in the center of the Teufelsberg radome.
Barry Marshall, from the New England Institute of Art in Brookline, Massachusetts, who used to be a guide at the Mapparium, told me how he used the acoustic to play practical jokes on visitors and “blow their minds.” The strong focus meant that he could stand far away from visitors and surprise them by calling “over here,” and they would look in the wrong direction.24 In Teufelsberg, I contented myself with spying on other conversations by finding the point where the speech of other visitors was being focused.
Long-distance whispering and sound focusing tend to unnerve people because they hear something that seems supernatural. If I were to chat with you in a normal room, the low frequencies in my voice would have roughly the same loudness in both your ears, whichever way you faced, because they can easily bend around your head by diffraction. Normally, low frequencies become much louder in just one of your ears only if I get very close, maximizing the “sound shadow” cast by the head. The effect reduces low frequencies in the far ear, making you think I am next to you. But in the Mapparium, the sphere can focus sound so intensely on one ear that it can trick the brain into thinking that I must be close by. Not only can I whisper sweet nothings to a loved one yards away; I can even narcissistically whisper them to myself!
The point of whispering, of course, is to say something so quietly that unintended listeners cannot hear it. Apparently, the original dome of the Capitol allowed members of the House of Representatives to whisper private messages to each other. But the amplification worked both ways: the congressmen could also overhear their colleagues’ secrets. We naturally seem to associate sound amplification by curves with spying, subterfuge, or illicit liaisons. Fellini used this link to dramatic effect in his film La Dolce Vita, in which a concave basin allows eavesdropping on conversations from the lower floor of a villa.25 But the most curious spying legend concerns a large limestone cave near Syracuse in Sicily called the Ear of Dionysius. The story goes that the tyrant Dionysius (ca. 430–367 BC) used the cave as a prison and exploited the acoustics to find out what hapless detainees were whispering to each other.
The cavern is shaped like a tall, pointy donkey’s ear and narrows dramatically at the top. The wedge shape acts like a funnel to sound, as Figure 5.7 illustrates, potentially collecting whispers from the ground level and concentrating them at the roof of the cave, 22 meters (72 feet) above. Legend has it that Dionysius spied on prisoners from a listening chamber at the top, picking up the amplified sound through a small hidden opening at the top of the cavern.
The cave is a popular tourist venue, and in the past, the listening chamber could be visited—as one traveler noted in 1842, “the only . . . mode of access being by means of a rope and pulley, the adventurer hazarding his life in a little crazy chair.”26 Despite the legend being recounted to tourists, some reports cast doubt on whether spying was possible. In 1820 the Reverend Thomas Hughes wrote, “A very low whisper is heard only as an indistinct murmur; the full voice is drowned in the confusion of the echoes. The voices of several persons speaking at the same time are as unintelligible as the cackling of geese, so that if the ancient Sicilians were half as loquacious as the modern, who always chatter in concert, they must very often have put the listening tyrant to a nonplus.”27
Figure 5.7 Sound in the Ear of Dionysius.
With modern health-and-safety precautions, visitors can no longer enter the upper chamber. A listener today is simply left to enjoy the reverberation at ground level, marvel at the legend, and gaze at the cavern’s large, ear-like shape. (A different sound-related tradition that has been stopped is the firing of firearms for tourists: according to another nineteenth-century visitor, “A pistol was fired, and the report was like the discharge of an eight-and-forty pounder.”28)
Recently, Gino Iannace, from the Second University of Naples, and collaborators persuaded the cave owners to let them into the listening chamber to survey the acoustic. Just as my team rates theaters, classrooms, and railway stations, Iannace’s group took a battery of measurements to assess speech intelligibility in the cave. The results were “on the bad side of average,” indicating that the cave’s reverberation makes speech muddy and incomprehensible. Undeterred, Iannace then carried out a series of perceptual tests, asking listeners to transcribe phrases recounted in the cave, but nobody could get a single word right. Disappointingly, the scientific measurements failed to support the legend.
Utterly overcome by pain and grief, I crouched against the granite wall.
I just commenced to feel the fainting coming on again, and the sensation that this was the last struggle before complete annihilation—when, on a sudden, a violent uproar reached my ears. It had some resemblance to the prolonged rumbling voice of thunder, and I clearly distinguished sonorous voices, lost one after the other, in the distant depths of the gulf.
Suddenly my ear, which leaned accidentally against the wall, appeared to catch, as it were, the faintest echo of a sound. I thought that I heard vague, incoherent and distant voices. I quivered all over with excitement and hope!
“It must be hallucination,” I cried. “It cannot be! it is not true!”
But no! By listening more attentively, I really did convince myself that what I heard was truly the sound of human voices.29
This is the moment at which Professor Hardwigg and Harry, the heroes of Jules Verne’s A Journey to the Centre of the Earth, miraculously reconnect via a whispering wall formed from a granite labyrinth. It is a prodigious structure, and Harry works out that he is hearing Hardwigg 8 kilometers (5 miles) away.
Outside Jules Verne’s imagination and aboveground, the largest real whispering wall I know of is 140 meters (460 feet) long—a mere baby in comparison, and also less poetic; it is the concrete dam that withstands the Barossa Reservoir in South Australia. For some reason the dam was built to be a precise arc. This vast, gray slab of concrete has turned into an unlikely tourist attraction, with visitors chatting with each other from opposite ends of the dam.
This wall does not focus sound as happens with elliptical ceilings and domes; the listener and speaker are too far from the focal point of the arc. What happens is that the sound hugs the inside of the concrete wall and is transported with surprising loudness to the other side of the dam.30
Whispering arches behave in a similar way, and they also show up in the most unlikely of places. On the lower level of Grand Central Terminal in New York City, outside the famous Oyster Bar & Restaurant, sweeping tiled archways, designed by Rafael Gustavino and his son in 1913, support the ceiling. If you whisper into one side of the arch, the sound follows the curve of the tiled ceiling before coming back down the other side. For the best effect, the whisperer and listener need to get close to the stone, like naughty children standing in opposite corners of a classroom.
This scene does not immediately make me think of marriage proposals, but still the location is a popular place for popping the question (the jazz musician Charles Mingus was supposed to have done so there). The sound effect has also inspired literature and films; the author Katherine Marsh uses the whispering arches as the starting point in her children
’s books The Night Tourist and The Twilight Prisoner, describing the arches as “one of the coolest places in New York.”
I have found a dozen documented cases of whispering arches. Very few, if any, seem to have been designed for their sonic quirkiness. The arch in St. Louis Union Station, Missouri, is adorned with a plaque that begins: “The Whispering Arch, an architectural accident or the sharer of secrets?” (a curious question, since presumably it could be both). Apparently, the sound effect was discovered in the 1890s, when, as the plaque says, “a workman dropped a hammer on one side of the arch and a painter on the other side, nearly 40 feet [12 meters] away, heard him.” This whispering arch was thus an accident of geometry.
No doubt there are many other whispering arches to be found. Elaborate architraves around doorways that contain sweeping arcs help channel sound from one side to another. Acoustician and Mayan pyramid expert David Lubman measured one in West Chester University in Pennsylvania. The doorway had an arc rather like a piece of upside-down, bent guttering within which sound propagates. People are so used to hearing sound become quieter as they get farther away that they find whispers emerging from the half pipe of the arch surprisingly loud. Lubman wonders whether this whispering feature was deliberate, because the half pipe that carries the sound seems to have little other purpose.31 But it may just have arisen by accident—a by-product of the door design. Sadly, the feature is largely ruined by traffic noise.
My favorite whispering arch is at the ancient monastic site of Clonmacnoise in County Offaly, Ireland. (How can a collector of sonic wonders resist that name?) An ornate Gothic doorway dating to the fifteenth century has carvings of Saints Francis, Patrick, and Dominic above and opens into the roofless remnants of the cathedral. Like the Oyster Bar archway in Grand Central Terminal, it is a popular spot for wedding proposals. Folklore has it that the doorway once had a very unusual use: Lepers would stand at one side of the doorway and whisper their sins into the half pipe in the architrave (Figure 5.8). The priest would stand at the other side of the arch, far enough away to avoid infection, listening to the confession emerging from the architrave. I spent an afternoon watching busloads of foreigners having fun whispering in the arch, despite rain and howling winds. How do these whispering arches work? They behave like whispering galleries.
Figure 5.8 The architrave half pipe you whisper into at Clonmacnoise.
The whispering gallery of St. Paul’s Cathedral in London gave me, as a young teenager on a scout trip, one of my first acoustic memories. The cathedral is built in the shape of a cross, with the dome rising above the intersection between the arms. It is such an important landmark of London that, during the Blitz in World War II, Prime Minister Winston Churchill ordered it to be protected at all costs, to boost morale.
Visitors climb 259 steps from the main cathedral floor up to the base of the dome and then emerge onto a narrow floor only a couple of yards wide running around the inside of the dome’s circular walls. At this point the dome has a diameter of 33 meters (108 feet). Metal railings line the inside edge of the gallery to keep people from falling off as they look up to the top of the dome or down to the cathedral’s floor while admiring the opulent splendor. I remember having great fun calling out to friends around the dome. It was busy and noisy, yet I could still hear my friends’ rude whispers carry a remarkable distance.
Whispering galleries have fascinated many famous scientists, such as Astronomer Royal George Airy, best known for his work on planetary science and optics. In 1871 he published a theory of how whispering galleries work, but it explains what happens only in perfectly spherical rooms such as the Mapparium. The Nobel Prize–winning physicist Lord Rayleigh was also intrigued, writing that “Airy’s explanation is not the true one” for St. Paul’s. To prove his point, Rayleigh made a scale model of a whispering gallery from a semicircular strip of zinc 3.6 meters (12 feet) long.32 He used a birdcall whistle at one end to produce a chirp skimming along the inside of the metal strip; at the other end the sound was remarkably strong, powerful enough to make a flame flicker. But when a narrow barrier was placed anywhere along the inside wall of the zinc strip, the flame remained undisturbed. This result showed that the sound waves were hugging the inside surface of the curved strip.
Sound clinging to and following the inside of the gallery walls is a pleasing scientific finding, but that alone does not explain the startling effect of the whispering gallery. Visitors often hear peculiar sounds, as reported by C. V. Raman in his 1922 paper:
In response to ordinary conversation, strange weird sounds and mocking whispers emanate from the wall around. Loud laughter is answered by a score of friends safely ensconced behind the plaster. The slightest whisper is heard from side to side, and a conversation may be easily carried on across the diameter of the dome, in the lowest undertone, by simply talking to the wall, out of which the answering voice appears to come.33
The sound hugging the wall creates an aural illusion because it is much louder than expected. In addition, both whisperer and listener need to get close to the wall. When listeners move their ear a short distance away from the wall, the sound suddenly gets much quieter. When the brain tries to work out how far away a sound source is, it uses loudness as a clue. Normally whispers are loud only when you are close to the talker. Furthermore, only when a source is close by does the loudness decrease rapidly with small movements of the head. The brain misinterprets the rapid quieting of the whispers as the ear moves away from the wall, and it thinks the source must be in the stonework.
Raman received a Nobel Prize for his work on light scattering, but he also carried out extensive research in acoustics. In the early twentieth century he documented five different whispering galleries in India, including the vast seventeenth-century Gol Gumbaz mausoleum in Bijapur. From the outside, Gol Gumbaz is an imposing building, testifying to the power of the Adil Shahi Dynasty; it rises impressively from the surrounding plain. It looks like a giant cube, with slender octagonal turrets in the corners and a vast dome nearly 38 meters (125 feet) in diameter on top, some 30 meters (100 feet) from the floor. In the words of the acoustic engineer Arjen van der Schoot:
If you walk into the place the internal size of it is humbling, but you soon forget because you’ll be taken by the acoustics as soon as you start making sound. The reverb[eration] in Gol Gumbaz is so staggering that Indian people travel for days just to hear it. And when they arrive they find a hundred people inside screaming at the top of their lungs.34
With children enjoying yelling and listening as their voices repeat over and over again, the atmosphere is like a crowded day at a swimming pool. Van der Schoot was carrying out acoustic measurements and so had the rare pleasure of enjoying the mausoleum when empty: “It took us two years to get the proper permissions and clearing the place for a couple of hours. Bus loads of people were held at the gate so we could do our work in silence in the astonishing whispering galler[y], where, when it is silent, you can count 10 echoes to your whisper.”35
As much pleasure as Gol Gumbaz gives its visitors, the whispering gallery was an accident of design. The decision to put a dome on top of the hall was made only after building had begun. I can find evidence of only one whispering gallery constructed deliberately. According to a 1924 edition of Through the Ages Magazine, “The whispering gallery in the Missouri State Capitol [Jefferson City, 1917] was carefully, mathematically laid out by a celebrated expert on acoustics, and this is undoubtedly the first time on record that such a thing has been done successfully.”36
For an acoustics conference, I produced some animations of how sound waves move around a whispering gallery. Using the latest algorithms on a fast computer, the movie showed how whispers are carried around by hugging the walls. In an idle moment while preparing my talk, I dashed over to the library to borrow a copy of the nineteenth-century acoustic bible The Theory of Sound, by Lord Rayleigh, which, remarkably, he wrote while recovering from rheumatic fever in Egypt. His description of how whispering gallerie
s work is simple to sketch out—much easier than using my complex computer models.
Figure 5.9 Sound in a whispering gallery.
Imagine cuing a ball on a circular pool table so that it goes off almost parallel to the side wall. The ball then maps out how sound moves around the gallery when someone whispers along the wall. An unexpected effect becomes apparent: the ball hugs the wall, circulating close to the curve without ever going into the middle of the circle. The same thing happens with sound in a whispering gallery, as shown in Figure 5.9.
When I visited the listening station at Teufelsberg, I demonstrated the whispering-gallery effect to my tour guide, Martin. He had encouraged others to try out the acoustics in the center of the room but was unaware that the voices could skim around the edges. Later, in brief moments when the radome was empty, I made a measurement by bursting a balloon on one side of the dome, while my sound recorder was balanced next to the wall on the other. With such a loud bang, the sound can make many circuits around the edge of the dome before dying away to silence. For one burst I counted eight clear bangs. A plot from one of the recordings (Figure 5.10) shows four or five spikes at times when the bang passes the microphone.
Figure 5.10 Sound created by bursting a balloon in the Teufelsberg radome being used as a whispering gallery.
But why is St. Paul’s a “whispering” gallery and not a “speaking” gallery? I recently returned to the cathedral and made some covert recordings. The gallery is best visited early in the day, when there are not too many other people making noise. It is also best visited with a friend who can do the whispering, but I was on my own. Fortunately, the attendant was particularly adept at producing just the right whispers. Back in my laboratory I analyzed the recordings, which showed one good reason for whispering rather than talking normally: The background noise leaking up from the main floor of the cathedral is quite loud across the bandwidth of normal speech. But in the higher frequencies at which the attendant was whispering, the background noise is much quieter, so the ghostly sounds are not swamped by the background hubbub.
