Cocktail Party Physics

your birdsong stays on my mind

"Officers, life doesn't have to be ugly. See, look at the birds out there. Listen to their call: 'Oo-wee! Oo-wee! Oo-wee! Oo-wee!" -- Beverly Sutphin, Serial Mom

Alas, as fans of John Waters' masterpiece know, Beverly's love of her feathered friends didn't extend to homo sapiens. She even killed her dentist and his wife after spying on them eating dinner and pulling roasted bird apart with their fingers. Those monsters! Clearly they had to go. But she's absolutely right that bird calls are something special, which is why bioacousticians and other scientists spend a lot of time in the field recording ornithological utterings and than analyzing those recordings back in the lab.

I was reminded of this by Ed Yong's recent post on the intricate song stylings of the lowly plain-tailed wren -- namely, that the males and females of the species engage in lilting duets with one another. That's the finding of Johns Hopkins researcher Eric Fortune, who has studied the birds in the wild to learn more about how each bird's brain circuitry encodes the entire song, although when they sing alone, they still include the same gaps in the melody that a partner would normally fill in. Read Ed's post for the gory details.

Birdsong -- or more technically "vocalizations" -- is also one way that diehard birdwatchers identify different kinds of birds in the field, along with the more traditional visual markers, especially since some of the rarer varieties prefer to be heard rather than seen. It has always been thus for birdwatchers, many of whom prefer the term "birding" because of the aural element involved in the practice.

Jen-Puc Piquant recalls a line from the Merry Wives of Windsor: "She laments sir... her husband goes this morning a bird-ing." I suspect that context indicates hunting, however, since Wikipedia informs me that the first written record of watching birds because they're, like, pretty and stuff (as opposed to shooting them for food) dates to the late 18th century, in the writings of Gilbert White and John Clare, among others.

Even then, it was as much about collecting specimens from around the world for the enthusiasts' private collections as anything else. The term "birdwatching" doesn't appear in print until 1901, when Edmund Selous published a book by that title, although the first field guide dates back to 1889: Florence Bailey's Birds Through an Opera Glass.

Today, birdwatching is more than just a hobby, it is a vital industry. Per the U.S. Fish & Wildlife Service, some 46 million enthusiasts spent $32 billion on their hobby in 2001 alone -- and that's just in the United States. Birdwatchers invest in a lot of technical toys, although it does make for carrying a lot of unwieldy equipment into the field: binoculars, a recorder, a field book, CDs, and MP3 player, cameras, tripods, not to mention food and water for the human observers.

The science has also exploded in this area. Between 2005 and 2006, bioacousticians published more than 1700 papers on a wide range of animal sounds, including birdsong. Thanks to that avalanche of research, these days there's lots of audio resources available to birdwatchers, including catalogs of hundreds of different bird songs, so that no one person needs to memorize all of them, like those poor Victorians.

But how do you take written notes on the sounds that birds make? Sure, you could use conventional musical notation, but many birdwatchers aren't musicians. Acousticians use sonograms, a kind of graphical representation that's great for research purposes, but doesn't really capture the essence of what the birdsong really sounds like. We're left with attempting to represent the songs phonetically, as in the quote from Serial Mom at the top of this post ("Oo-wee! Oo-wee!"). Field guides often resort to vague phrases like "far-carrying melancholic song," or the mysterious "tee-do-do-eet."

Enter two enterprising Ecuadoran researchers who think they may have a solution to both problems: there might just be an app for that. Hugo Andrade and David Puente have developed a software system that "transforms sound into a sequence of numbers that can be readily converted and printed in a Quick Response (QR) barcode."

They debuted a prototype version of the software at last week's meeting of the Acoustical Society of America, where attendees were able to see firsthand how the system imports recorded sounds and transposes into a QR barcode, and then "decodes" it using a conventional barcode scanner.

The next step for Andrade and Puente is to adapt their software into an app for smart phones, a kind of mobile catalog in searchable "e-book" format. That would be the ultimate pocket field guide for birdwatchers, wouldn't it? Instead of lugging around lots of pricey and heavy equipment, enthusiasts could happily go about their birdwatching business armed with just a field guide and a mobile phone with a built in camera. If they hear a bit of birdsong, they can look it up on the smart phone app, which will contain a picture of the bird (for visual identification) and a barcode that can be played so watchers can verify the birdsong matches what they've just heard. "No more confusing tweeting sound descriptions!" Andrade declared in his lay language summary of the work.

Ideally the software could also be adapted for recording purposes in the field as well, to capture the sounds of various animal species -- birds, sure, but also frogs, insects, bats, and so on. So perhaps the next time Eric Fortune goes into the field to capture the dulcet tones of his beloved plain-tailed wrens, he'll have far less baggage to lug around.

a shot in the dark: the acoustics of gunfire

Who shot first, Han Solo or Greedo? If you're a diehard Star Wars fan, you know the answer is Han Solo. It's a popular bit of geek trivia, referring to a classic scene in what is now known as Star Wars Episode IV: A New Hope. Fans have engaged in heated online debates, signed petitions, and broken off budding friendships over this pressing issue. And now, thanks to advances in the acoustics of gunshot forensics described at last week's Acoustical Society of America conference, those in the "Han Shot First" camp -- and Jen-Luc Piquant reminds you that there is no other acceptable answer -- can bring in the mighty fists of science to make an even stronger case.

You remember the scene. Han meets bounty hunter Greedo at the Mos Eisley Cantina. Greedo is there to collect on behalf of Jabba the Hut, who wants payment for a load of smuggled cargo that Han dumped to avoid getting caught. The two -- well, one man, one strange-looking alien -- sit across from each other. Greedo has his blaster pointed at Han, who, as they chat, positions his own blaster beneath the table. In the original theatrical release, Greedo says, "Ive been looking forward to this for a long time," and Han replies, "Yes, I'll bet you have." He then blows Greedo away before the bounty hunter manages to take his shot. (Pic below via Mental Floss.) Except fans were outraged when director George Lucas modified the scene for the film's re-release in 1997. In the modified scene, Greedo fires first, missing Han Solo despite the point blank range (Greedo is a pretty horrible shot for a bounty hunter), and Han only fires and kills Greedo in retaliation. Apparently Lucas didn't want there to be any moral ambiguity whatsoever surrounding Solo's ultimately heroic character -- but that ambiguity has always been a big part of the character's charm. (Think the redemption of Spike in Buffy the Vampire Slayer for a similar character arc.) You don't mess with the Star Wars canon, people -- even if your name is George Lucas.

But what if you were a cantina patron whose view of the table where Han and Greedo was sitting was blocked? When the authorities asked for your account of the shooting, how much could you really tell them about what happened, when you only heard the blaster shot? You might be able to tell them quite a bit, actually assuming you'd boned up on the latest acoustics research on gunshot forensics.

Okay, granted, nobody's done a bona fide study on sci-fi blasters, but acousticians Steven Beck (BAE Systems in Austin, Texas) and the FBI's own Hirotaka Nakasone have collaborated on a number of studies of gunshot recordings of various types of firearms, wit microphones placed in many different configurations and angles relative to the gun being tested. They've combined those findings with visual evidence (photographs of gunshots) to develop more accurate waveform models for forensic audio analysis of actual gunshot events.

A waveform is a kind of acoustic signature of a given sound. So waveforms can contain a lot of useful information for someone who's trained to interpret them, once there's a reliable catalog of waveforms associated with specific sounds, of course -- in this case, different kinds of gunshots at different ranges and angles. For instance, a trained analyst might be able to determine whether a sound is a gunshot or not, or how many guns were used, or -- most relevant for Star Wars fans -- who fired first. It turns out that a gunshot "acoustic event" (i.e., "sound") has more than one component. First there is the muzzle blast (the loud bang that propels the bullet forward), which sounds like a short pop in laboratory conditions, but is louder with more reverberation in a more real-world setting. Factors that can affect the resulting waveform from this part of the gunshot include the make of firearm, model, length of the barrel, and the type of ammunition used.

The muzzle blast is followed by a ballistic shock wave, i.e., the crack as the bullet zips through the air, traveling faster than the speed of sound. It's basically a supersonic boom, as the shock wave forms a cone behind the speeding bullet. You don't always hear it either, if you happen to be standing at more than a 90 degree angle to the line of fire. How big the shock wave is, and how long it lasts, depends on such factors as the bullet's speed and its size, while the usefulness of a recording of the ballistic shock wave depends on where the microphone is located, relative to the line of tire. Per Beck and Nakasone's lay language writeup for the conference:

"Currently analysts use simple blast waveform models and cross correlation techniques to study acoustic gunshot waveforms and try to answer basic forensic questions." However, "There are many variables that can affect recorded gunshot signals and forensic audio analysts need to be aware of their influence on the resulting waveforms."

That's why the two acousticians performed so many different kinds of studies of recordings made in both the lab and the field, trying to take into account the many, many factors and conditions that could change a waveform in significant ways, and make it less accurate (and hence less useful) for forensic purposes.

The resulting waveforms matched closely with their theoretical models, at least for the high-quality recordings they made themselves in the lab, with top-notch equipment. Beck and Nakasone acknowledge that for real-world purposes, the kinds of gunshot recordings an acoustic analyst encounters are likely to be done with handheld devices like cell phones, with a lot of background noise. "Waveforms recorded under more realistic conditions can vary significantly -- even to the point of being unrecognizable and forensically unusable," they admitted in their lay language paper.

The culprit is reverberation, or the echoes that distort the waveform, not to mention from the recording equipment itself. Most gunshots occur in urban settings, so the sound will ricochet off pavement, buildings, even nearby parked cars. Beck and Nakasone focused their efforts on recordings of gunshots in the field, and then compared those with those made in the lab, but there are also "acoustic event locating systems" used by military and law enforcement officials.

Such a system was implemented in parts of Washington, DC, back in 2006. It was even featured on an episode of Bones, in which the female serial killer known as the Gravedigger is assassinated by a long-range sniper outside the courthouse for an appeals hearing. The mighty crime-fighters at the Jeffersonian use the sensor system to pinpoint the likely source of the gunshot via acoustic triangulation.

There's more than one kind of acoustic sensor used in these kinds of systems. Per How Stuff Works (a.k.a., "The Enormous Time Suck for the Eternally Curious"), some sensors detect the sonic boom a speeding bullet produces as it moves faster than the speed of sound, while others detect the optical effects produced by a muzzle blast. The DC system uses the former variety, spacing a dozen or so sensors per square mile. Those sensors can each detect the sound of gunfire within a two-mile radius. Since the speed of sound is well-established, all you need to do is measure the difference between the time it takes for the sound of the gunshot to reach three different sensors to figure out the exact location using standard GPS. That's acoustic triangulation.

Then this information gets relayed to the nearest 911 call center, and the police are on it, people. At least that's the main idea. It's pricey -- DC spent millions of dollars on the system -- but it can cut the response time for a "shots fired" 911 call in half, according to DC police officials. If only we had the film equivalent of such a system, or highly trained forensic audio analysts, to settle once and for all what every Geek and Gamer Gurl knows: that Han Shot First. (FYI, video below is a parody of Katy Perry's "California Gurls.")

 

drunken masters of lingua franca

We're having a bit of a thematic installment of Friday Fodder this week. See, I regrettably missed this year's Acoustical Society of America meeting in Seattle, but I still managed to rifle through their impressive collection of lay language papers covering talks presented at the meeting. I already covered the watershed work on muting the noise of flushing low-flow toilets a couple of weeks ago. But there were also tons of papers relating to speech: researchers at looking at the voice physiology of transgender people; how men and women imitate each other and what this might tell us how gender, emotion and just plain old acoustics impact those imitations; and how to identify the sexual orientation of speakers based on just their voices.

There was even a paper on "understanding casual speech," which I guess must be, like, hard and stuff for some folks. And now I'm happy to report that Abby Kaplan, who works in the linguistics department at the University of Utah, had some interesting things to say about drunken speech patterns -- namely, whether it's harder to pronounce certain sounds or words when intoxicated.

This might strike many readers as "Duh!" science, but as Kaplan discovered, actually demonstrating in a rigorous fashion what we all instinctively feel to be true -- based on personal experience -- isn't quite as simple as it seems. (Nor is she the first to be intrigued by this question. Via Language Log, I learned that there is a 1902 treatise by a guy named Edward Wheeler Scripture, The Elements of Experimental Phonetics, that toys with the notion of studying the phonetics of drunken and other unusual speech patterns through recordings.)

Let's start with the assumption that certain sounds and words are more difficult to pronounce than others (an effect that is exacerbated if, like me, your vocabulary was acquired through extensive reading -- I habitually put my emPHAsis on the wrong syllABle). "Intuitively, it seems obvious that some sounds are harder to pronounce than others," Kaplan writes. She points to the "th" sound, as in "think" as being rather rare in the world's languages, and a stumbling block for many ESL students.

Linguistic theory holds -- so far -- that some phonetic speech patterns involve substituting easier sounds for harder ones. The most common example is how [p] is ponounced [b] in some languages (Korean, Malaysian) when it occurs between two vowels, while in other languages [b] is pronounced [v] when between two vowels. But how do we know if these quirks occur because one sound really is easier than another? Well, you can always conduct a controlled experiment with participants reading a list of words first while sober, and then while intoxicated, since the latter sate impairs verbal (and other) functions.

And here is where the fun begins, because in order to test this hypothesis, Kaplan had to get her study participants quite drunk. I'm guessing she had quite a few volunteers. They were plied with equal parts vodka and orange juice until they reached a blood alcohol content level of between .10 and .12, just above the legal driving limit in the US of .08) -- until, one assumes, they started to sound like the title character played by Dudley Moore in the original Arthur from the 1970s:

Kaplan's findings: (a) drunken speech is indeed different from sober speech (the "Duh" aspect); and (b) "the overall range of sounds that people pronounce is smaller in drunken speech than in sober speech," which is consistent with Kaplan's original thesis, with one caveat: the participants didn't always choose the easier sounds when they were drunk. Sometimes they pronounced more of their [p]'s like [b]'s when drunk, but then, they also pronounced more of their [b]'s like [p]'s. So you can't really say, based on her study, that one sound is easier than the other. And this, says Kaplan, has some interesting implications for the prior assumptions made by linguists:

"If drunken speech really does involve saying `easier' sounds, then these results are a challenge for what linguists have traditionally believed about the sound patterns described above.  These results suggest that neither [p] nor [b] is easier to pronounce between vowels, but rather that some sound intermediate between the two is easier than both.  If it's not true that [b] is easier to say between vowels than [p], then there must be some other reason languages replace [p] with [b] in this context.... One important goal for future research is to try other methods of studying 'easy' and 'hard' sounds, and see whether the results of those methods are consistent with the results of this experiment."

Arthur's slurred speech is made more difficult to understand, one presumes, by his pronounced British accent. Over at The Last Word on Nothing, Sally Adee has an amusing post on why a British accent makes phrases like "chuffed" and "can't be arsed" -- even random dialogue from Jersey Shore -- sound posh, whereas when Adee herself attempts those phrases, her native-German-laced-with-Long-Island-and-a-bit-of-Baltimore accent simply isn't up to the "can't be arsed" challenge: "When a Brit delivers the phrase, those three little words are transformed into gleaming pearls of wit. But in my mouth, that 'r' is like hitting a ten-foot pothole in a clown car.”

I think we should all be able to say "chuffed" and other amusing Britishisms if we feel so inclined, even if it makes our friends roll their eyes -- possibly even with a fake British accent. C'mon, it could be funny, as well as pretentious. Jen-Luc Piquant deliberately puts on airs for her fellow avatars in Cyberspace; she thinks it makes her sound more sophisticated to affect zee 'eavy Gallic accent, n'est-ce pas? I don't have the heart to tell her that it just makes her seem like a prat. Who am I to judge?

Heck, sometimes I find myself imitating (badly) people's regional accents in casual conversation in spite of myself. I recall as a teenager, during church fellowship hour, finding myself in conversation with a well-spoken, articulate high school senior girl, who was making a valiant effort to reach out to the shy, socially awkward sophomore (i.e., me -- people don't believe me today, but I was painfully shy through much of college), fidgeting before her and struggling to maintain eye contact. I did my best to match her liveliness, her vocal inflections -- except then she got a strange look on her face and abruptly walked off. She thought I was mocking her (I wasn't). It was one of many failed attempts to fit in with my church's social structure.

The tendency of people to imitate tone of voice, gestures, and other elements of conversational style is something that is well-documented by linguistic studies, according to Sara Phillips, a linguist at Ohio State University. Phillips and her collaborators decided to explore one commonly imitated factor in particular: regional dialects, most notably, how certain vowels are pronounced. Northerners, for instance, might flatten out the vowel sound in "block" such that it seems closer to "black," and make a subtle distinction in vowel sound when they say "caught" or "cot," compared to Midland speakers. "We wanted to see if speakers imitate only idiosynchratic properties of the talker's voice, or if they imitate properties of the talker's dialect, too," writes Phillips.

So they had participants repeat words and phrases of certain regional dialect talkers, recording the utterances of both. These were designated A, B and X, in which B was the speaker's normal pronunciation, A was the speaker prnouncing the same word in imitation of a dialect speaker, and X was the original dialect speaker's pronunciation. Then they played the recordings back for a control group of listeners who were asked to determine whether A or B sounded more like X. (There are sound files so you can hear for yourself.) Anyone who chose A seemed to be picking up on the imitated accent of the dialect. Or were they? Phillips et al. coupled their study with an acoustic analysis, and came to a somewhat different conclusion:

"It appears that speakers were imitating aspects of pronunciation that are not known to mark dialect, such as the duration of the vowel or final consonant of the word. In our experimental materials, the Northern talkers produced longer vowels and consonants than the Midland talkers.  This difference may have resulted in the appearance of dialect imitation, even though participants were really only imitating idiosyncratic features of the talkers’ speech. If they were imitating dialect features, we would have expected to find effects of vowel pronunciation instead.  While we did find weak evidence for vowel imitation, the effect was not significant.

Taken together, these results suggest that while speakers imitate characteristics of individual talkers’ speech, they do not necessarily imitate well-known dialect markers. It may be that our effect of dialect on imitation was due to peculiarities of our speakers rather than real differences between the Northern and Midland dialects.

Okay, but has anybody bothered to study what it is about foreign accents that makes the speakers so difficult to understand? Mais, oui! I give you graduate student Alison Trude and her faculty advisor, Sarah Brown-Schmidt, both of the University of Illinois in Urbana-Champaign's psychology department, who conducted a series of studies looking at a listener's ability to understand a native English speaker and a native Quebecois speaker both speaking English words and phrases.

A classic illustration of the difference is the respective pronunciations of "bee" and "beet." While a native English speaker will pronounce the "ee" in both the same, the Quebecois will pronounce "beet" with a short "i" sound, so that it sounds like "bit." It was an elaborate experimental setup involving tracking eye movement as participants viewed computer images of various objects while listening to accented and unaccented words describing those images. Supposedly this told the researchers how quickly they understood the words.

The prediction: "Previous research has shown that participants spend more time looking at pictures whose names sound more alike, so it was predicted that participants would look at the picture of the beet more when hearing the English speaker say 'bees' and would look at the picture of the beet less often when hearing the French speaker say 'bees,' because in the French accent, 'bees' and 'beet' sound less similar."

The actual result was just the opposite. Participants looked at the wrong picture more often when listening to the French speaker, and adapted more easily if they heard each of the accented words pronounced with the same ending ("beet", "seat", or "feet"). It also helped if the accented words didn't sound like another English word, since there was less room for further confusion. So Trude and Brown-Schmidt concluded from their experiments that (a) if you're talking to someone with  foreign accent, having that person utter a string of words that sound similar will help you adapt to their accent; and (b) you'll have more trouble if the accent causes you to confuse two real words in English, i.e., the "beet" vs "bit" example (or "beep"). Per Trude and Brown-Schmidt:

These findings show that we as listeners have very particular expectations about how speech should sound. Some theories say that when listening to a foreign speaker, we relax the rules of pronunciation and are more accepting of things that don’t sound quite right because we understand that he or she cannot speak the language perfectly. However, these results suggest the opposite: native speakers of a language are listening for well-formed words, regardless of who’s talking.

This might have a useful application in bringing foreign actors to Hollywood, for example. Arnold Schwarzenegger notoriously kept his lines to a bare minimum in Conan the Barbarian in his early acting days, all uttered in heavily Austrian-accented English. And Hong Kong martial arts master Jackie Chan's move to American films was hampered by his broken English and heavy Chinese accent; ditto for Jet Li, who had, like, three lines in Lethal Weapon 4. And thus we come full circle back to our theme of drunken speech with a look at a unique style of martial arts: Zui Quan, which loosely translates as "drunken boxing." (Yeah, it's just an excuse to bring up Jackie Chan and include an awesome video clip. What of it?)

Chan immortalized the style in two films, Drunken Master I and Drunken Master II, wherein his protagonist gains the upper hand in climactic fight scenes by becoming inebriated --although it's important to strike the right balance: earlier in DM-II, Chan's character becomes a bit too drunk after a fight with his father, slurring his speech and warbling a little ditty called "I Hate Daddy," before getting the crap beaten out of him by thugs because he's too inebriated to fight at all.) In reality, Zui Quan practitioners only imitate the physical movements of drunkards; they are not actually drunk.  Per Wikipedia:

"The postures are created by momentum and weight of the body, and imitation is generally through staggering and certain type of fluidity in the movements. It is considered to be among the most difficult wushu styles to learn due to the need for powerful joints and fingers. ... Zui Quan techniques are highly acrobatic and skilled and require a great degree of balance and coordination, such that any person attempting to perform any Zui Quan techniques while intoxicated would be likely to injure himself.

Even though the style seems irregular and off balance it takes the utmost balance to be successful. To excel one must be relaxed and flow with ease from technique to technique. Swaying , drinking, and falling are used to throw off opponents. When the opponent thinks the drunken boxer is vulnerable he is usually well balanced and ready to strike. When swigging a wine cup the practitioner is really practicing grabbing and striking techniques. The waist movements trick opponents into attacking sometimes even falling over. Falls can be used to avoid attacks but also to pin attackers to the ground while vital points are targeted."

You can see Zui Quan in action in this eight-minute choreographed fight scene from Drunken Master II, when Chan was at the height of his athletic prowess. (And remember, he does all his own stunts, so he really is falling backward onto a bed of hot coals. Ouch!) Definitely one of the Best. Fight. Scenes. Ever.

hop, skip and a jump

If invention is the mother of necessity, it's not surprising that times of war should give rise to innovative technological breakthroughs. Such was the case in the 1930s, when Europe (and much of the Western world) was on the brink of a second world war and military leaders from many different nations were scrambling to find advanced weapons technologies to give their "side" an edge in the escalating hostilities. One place nobody thought to look was Hollywood, which might explain why an obscure patent filed in 1942 failed to garner much notice.

Fans of classic film know the name Hedy Lamarr for her memorable silver screen performances. But this lovely actress also made a small contribution to wartime technology with her co-invention of an early form of spread spectrum communication technology, or frequency hopping, in which a noise-like signal is transmitted on a much-larger bandwidth than the frequency of the original information.

Born in November 1914 as Hedwig Eva Maria Kiesler in Vienna, Austria, Lamarr came from "Jewish haute bourgeoisie" stock (Wikipedia's words, not mine). Her father was a bank director and her mother was a Hungarian pianist, who made sure Hedy studied ballet and piano as a child. As a teen, the young Hedy attended a famed acting school in Berlin headed by director Max Reinhardt. She dropped out of school to be Reinhardt’s production assistant and had bit parts in two films before starring as a love-starved young wife married to a much older man in a Czech film called Ecstasy. Not only did Lamarr appeared nude on screen, frolicking through the woods and crouching behind convenient bushes, but there were numerous close-up shots of her (ahem) simulating orgasm in quite explicit sex scenes. (She later recalled that far from being sexually transported, her expressions were in response to the director poking her in the derriere with a pin to elicit them.)

Even back then, all it took was a sex tape to launch a stunning young woman into a life of celebrity. Lamarr leveraged her beauty and sudden notoriety into what seemed like an advantageous marriage just before she turned 20, to a man 30 years her senior. Her (first) husband was Friedrich Mandl, an arms merchant based in Vienna who sold munitions and manufactured military aircraft. Mandl forbade her to continue acting -- in fact, he tried to buy up all the prints of Ecstasy, feeling that the expression on his wife's face during those love scenes was indecent. 

Lamarr initially took the restrictions on her freedom in stride: she dutifully presided over her husband’s lavish parties, attended by Hitler and Mussolini among others, and was often present at his business meetings. None of the men in the room gave a second thought to the presence of Mandl's beautiful young wife -- how could she possibly follow their manly discussions about wartime strategy and weapons? As a result, despite her lack of formal education, she acquired a great deal of knowledge about military technology, most notably guided torpedoes and the vulnerability of radio-controlled weapons to jamming and interference.

Disillusioned with married life –- especially her husband’s increasingly controlling behavior and shady business dealings with Nazi industrialists –- Lamarr disguised herself as one of her maids and escaped first to Paris in 1937, where she obtained a divorce from Mandl, and then moved to London. She would marry five more times before giving up on the institution. (One colorful version of her legendary escape has her attending a party with Mandl decked out in all her expensive jewelry, then drugging him with the help of her maid before fleeing.)

After meeting Louis B. Mayer in London, he signed her to MGM as Hedy Lamarr. Her debut was in the 1938 film Algiers. She went on to play many more roles -- including a cringe-inducing turn as a native femme fatale in White Cargo, announcing ominously to each victim, "I am Tondelayo. I will come and make tiffin for you." She also co-starred with Lana Turner and Judy Garland in the musical extravaganza Ziegfield Girl, and showed an unexpected flair for comedy opposite Clark Gable in Comrade X, in which she played a Russian streetcar conductor named Theodore.

In a town filled with stunning women, Lamarr stood out. Actor George Sanders once said that she was “so beautiful that everybody would stop talking when she came into a room.” But Lamarr was more than just a pretty face: she had a natural mathematical ability and lifelong love of tinkering with inventions. One of those ideas bore fruit when she met her Hollywood neighbor, avant garde composer George Antheil, in the summer of 1940.

Born in New Jersey to Prussian emigrants, Antheil studied music in Philadelphia and toured Europe as a concert pianist, before turning his hand to composing. His signature piece was called “Ballet Meanique,” a complicated score originally written for Fernand Leger’s 1924 abstract film of the same name. It called for mechanically synchronizing sixteen player pianos, as well as xylophones and percussion. He returned to the US in 1933 to compose for film, and also became a syndicated romance advice columnist, writer for Esquire, and author of a book entitled, Every Man His Own Detective: A Study of Glandular Endocrinology.

Legend has it that Lamarr approached him for endocrinological advice on increasing her breast size, but the two soon began chatting about weapons, particularly radio controlled torpedoes and how to protect them from jamming or interference. She realized that “we’re talking and changing frequencies” all the time, and that a constantly changing frequency is much harder to jam.

This became the basis for their design for a torpedo guidance system. Lamarr contributed the idea of frequency hopping, while Antheil drew on his experience with “Ballet Meanique” to devise a means of synchronizing the rapidly changing radio frequencies envisioned by Lamarr.

Their joint invention used a mechanism similar to piano player rolls to synchronize the changes between the 88 frequencies –- not coincidentally, this is also the standard number of piano keys -– and called for a high-altitude observation plane to steer a radio-controlled torpedo from above. They submitted their patent on June 10, 1941, and the patent was granted on August 11, 1942.

This was not an entirely new concept. Nikola Tesla alluded to frequency hopping in 1900 and 1903 patents, filed after he demonstrated the first radio-controlled submersible boat in 1898. Realizing he needed to shield the radio signals from interference, he devised two different techniques that relied on altering the carrier frequency to eliminate interference.

A similar patent for a “secrecy communications system” was granted in 1920, with additional patents granted in 1939 and 1940 to two German engineers, who also held German patents for their work. And evidence came to light in the 1980s that during World War II, the US Army Signal Corps worked on a communication system dubbed SIGSALY that used the spread spectrum concept as well.

Lamarr and Antheil had less success convincing others their idea was feasible. Examiners at the National Inventor's Council questioned the robustness and accuracy of the internal clockwork mechanism responsible for moving the perforated tape through the system, while the U.S. Navy felt the clockwork mechanism was too bulky and unreliable to use with a torpedo, although Antheil argued it should be possible to miniaturize it to fit inside a watch. As Antheil later recalled:

In our patent Hedy and I attempted to better elucidate our mechanism by explaining that certain parts of it worked like the fundamental mechanism of a player piano. Here, undoubted, we made our mistake. The reverend and brass-headed gentlemen in Washington who examined our invention read no further than the words "player piano." "My god," I can see them saying, "We shall put a player piano in a torpedo."

In fairness to the US Navy, that wasn't their only objection; frequency hopping was a bit too far ahead of its time -- technology had to catch up. It wasn’t until 1957 that engineers at Sylvania Electronic Systems Division adopted the concept, using the recently invented transistor for an electronic system rather than the original clockwork mechanism. The Sylvania system was installed on ships sent to blockade Cuba in 1962, three years after Lamarr's and Antheil's patent had expired, and its primary use was for secure military communications rather than remote control of torpedos. The same basic concept is still used in US defense communication satellites -- and in modern cell phone technology.

Antheil died in 1959, no doubt still bitter that his work hadn't been taken seriously by the military. He fared better than Lamarr, who wanted to join the National Inventor's Council and was told she could be most useful to the war effort by exploiting her celebrity status to raise money -- which she did, once selling kisses for $50K a pop and raising a whopping $7 million at a single event. Lamarr went on to make more than 20 more films, most famously Cecil B. de Mille’s 1949 Samson and Delilah. She played the title role of the Biblical temptress, of course.

But somehow, true greatness in Hollywood eluded her -- partly due to the dearth of meaty roles for women in general, but also because casting agents couldn't see past her looks to find her substance. (Her fellow bombshell actress Jayne Mansfield had a similar fate. Mansfield purportedly had an IQ of 163 -- or 149, depending on which source you believe -- spoke five languages, and was a classically trained pianist and violinist, but admitted her public didn't care about her brains. "They're more interested in 40-21-35," she said.)

Lamarr retired from film completely in 1957, settling in Altamonte Springs, Florida, where she painted and dabbled in the odd invention -- like a pocket on the side of a Kleenex box in which one could deposit used tissues. It was a quiet existence, apart from the occasional lawsuit, marriage or divorce, or shoplifting scandal (she was arrested in 1966, and again in 1991 at the age of 78, although the charges were later dropped). Lamarr has a star on Hollywood’s Walk of Fame in honor of her film career, but she took particular satisfaction in being awarded the Electronic Frontier Foundation Award in 1998, more than 50 years after she and Antheil received their patent. “It’s about time,” she reportedly said upon hearing the news. Her son accepted the award on her behalf.

Lamarr died on January 19, 2000, in her Florida home. In addition to her own memoir, Ecstasy and Me, published in 1967, two new biographies appeared last year, in response to a resurgence of public interest in this forgotten film star when her wartime patents came to light. Despite this, she will likely always be remembered more for her spectacular beauty than for her technological contributions, which are usually treated as an intriguing footnote to a life that was not exactly devoid of drama. “My face has been my misfortune,” Lamarr once observed, describing it as “a mask I cannot remove. I must live with it. I curse it.”

Hedy Lamarr - Ecstasy - The funniest videos are a click away

flushed with pride

There is a marvelous French film from the 1970s called The Tall Blond Man with One Black Shoe, starring Pierre Richard as a hapless concert violinist who unwittingly becomes entangled in an elaborate game of one-up-manship between two rival factions in the French secret service. Specifically, one group -- knowing they are being bugged -- pretends to be meeting an international James Bond-like super secret agent man at the airport, except really, they just randomly pick someone off the flight to Munich and pretend he's a spy. The person they pick is the violinist, who stands out because he is wearing one black shoe and one brown shoe that particular day (due to a mixup with the hotel concierge in Munich). This being a screwball comedy, wackiness ensues, since the violinist is clumsy and accident-prone, plus is carrying on an affair with his best friend Maurice's rapaciously adventurous wife, Paulette, a harpist by trade (they're all in the same orchestra, natch). Let's just say she likes to play "horsey" -- or fait le cheval -- and the rival faction captures that particular moment on audiotape (per the video clip below).

Anyway, believing him to be a spy, the rival faction springs into action, bugging the violinist's apartment -- including placing a bug in the toilet -- and monitoring his conversations from a white florist's van on the Paris streets. The problem is, they set the volume control too high on the bug in the toilet, which means every time the violinist flushes, it roars through the speakers in the van at very high decibels. This sets up the perfect sight gag: the van driver stops at a light, his window rolled down, and a man in a convertible pulls up beside him. Suddenly there is the loud sound of a toilet flushing from inside the van. Convertible Man stares. The van driver stoically looks straight ahead. And mercifully, the light changes.

All of which is a roundabout way of saying, the sound of one toilet flushing can be very loud indeed. It's even louder if one happens to install a fancy low-flow toilet, common in places like southern California, where water is less plentiful than in, say, Seattle. Seattle is the location of this year's Acoustical Society of America (ASA) meeting, coming up in a couple of weeks, and among the highlighted papers being presented is one on reducing toilet flushing noise in adjacent offices, by an acoustical consultant from North Carolina named Noral Stewart. It's stuff like this that makes the ASA meetings so much fun.

(This title of this post, by the way, is an illusion to a 1969 "biography" of Thomas Crapper, a 19th century sanitation engineer rumored to have invented the flush toilet. He didn't, although he did take out several patents relating to plumbing. The author of Flushed with Pride was a satirist named Wallace Reyburn, and his "biography" of Crapper was mostly fictional, despite being based on a real person.)

So, apparently Stewart was hired to solve a knotty office problem: overly loud flushing sounds from new low-flow toilets that were carrying over into seven rows of five offices (each row on a different floor), in which the three middle offices were right next to the rest rooms. You really didn't want to be the poor sods who got the middle offices on each of the floors: those shared a wall with both the men's and women's restrooms. Talk about Flush Central. Preliminary measurements of the sound levels showed many of the flushes were louder than 60 decibels (especially in that middle office), and almost all topped 50 decibels. That's roughly the same level as a typical conversation, which might not seem all that loud, until you realize that every time someone flushes a toilet, those in conversation would be momentarily drowned out by the sound. Noise from ventilation systems, for comparison, is usually around 35 to 45 decibels. (Inside the restroom, sound levels are more like 80 decibels; the loudest Stewart measured was 96 decibels in one of the women's stalls, a good 20 decibels higher than your average toilet's flush should be.)

 The recommendation of an initial study was to reduce the flushing sound levels to a more tolerable 40 decibels. How to accomplish this? First, they tried installing an extra wall out of gypsum, filling the cavity with foam, and this brought the noise down to between 40 and 50 decibels. That's because those materials are great at absorbing sound waves.

But there's a catch: the flushing sound has a particularly high frequency -- kind of a hiss, in fact -- and for whatever reason, gypsum isn't as effective when it comes to blocking those particular frequencies.

The solution: use fiberglass in the cavity instead of foam -- or invest in new gypsum panels with a damping agent in the middle to block the frequencies gypsum can't. By doing this, Stewart managed to bring the sound levels down to below 40 decibels. I'm sure the folks in the middle offices were well pleased with these improvements. Nothing kills a business conversation faster than an ill-timed flush -- especially when you're trying to play it cool over speaker phone.

All the adjacent walls were modified in this way, and you'd think Stewart's job would be done. But never underestimate the acoustical complexity of office space (or any space, for that matter). See, while this solved the noise problem for most of the offices, there was one troublesome office right next to the women's restroom where the sound levels of the flushes still topped 40 decibels, despite the modifications to the shared walls. Why? I'll let Stewart explain:

"It was evident that structural flanking was now controlling the remaining sound reaching the offices in this and the other louder cases. That is, the dominant sound in the offices is no longer the sound heard in the rest room coming through the wall. Instead, the sound in the offices is due to direct contact of the toilets and their piping with the walls and the floor. This very efficiently introduces the sound into the structure."

And believe me, this is a case where "efficiency" is not desired. Nor was the ventilation system much help in this respect: for some reason, this office building had a very quiet system, on the order of 30 decibels. That would generally be a good thing, except a louder background noise would help mask the constant sound of flushing toilets seeping into the affected offices. Those are tougher issues to resolve after the fact, which is why Stewart sums up the lessons learned from his case study thusly: "Extra care is required in the building design when [low-flow] toilets are used."

Specifically, you need to make sure the walls are well insulated against the correct frequencies, and that plenty of rubber is used around the toilets and pipes to mute sound waves. Balance that with just the right level of background noise to mask any flushing noise that does seep through, and you'll have far happier office workers. The same goes for designing restaurants; you never want to put tables too near the restrooms, and if you do, make sure the flushing sounds are well insulated to avoid putting diners off their food.

Who knew there could still be complicated science and engineering involved with the lowly toilet? I leave you with more pleasant sounds, however: another classic scene from The Tall Blond Man With One Black Shoe, this time a botched classical performance because Maurice now suspects his wife's infidelity (although still clueless that his best friend is her partner in crime).

babble-onia

The Time Lord brought home the latest issue of Scientific American last night, just so I could peruse Graham Collins' article on "Solving the Cocktail Party Problem" -- namely, studying how our brains separate various auditory streams when in a crowded room, like a restaurant or a cocktail party. (Personally, my brain has never been especially good at this. I find myself having to really concentrate when the noise levels reach a certain critical threshold.) Scientists have been pretty successful at studying how the brain accomplishes this amazing feat. They've been less successful at devising computer algorithms to do the same thing.

We've all experienced this phenomenon. Walk into a crowded bar, with music blaring, and your first impression is likely to be a shudder at the sudden wall of sound -- which you will interpret at first as a single loud noise. But very quickly, you adjust, and different sounds begin to emerge. We navigate by tuning our neurons to specific voices, thereby tuning out others -- like that irritating, leering would-be Lothario at the other end of the bar, or all that ambient noise.

A few years ago, at an acoustics conference, I chatted with Shihab Shamma, a researcher at the University of Maryland, College Park. He believes this ability arises from auditory nerve cells in the brain that re-tune themselves to specific sounds as part of the adaptive process. It's kind of an auditory feedback loop that enables us to sort out confusing incoming acoustical stimuli.

He's surprised, however, by how quickly this process happens: auditory neurons in adult mammal brains make the adjustment in a few seconds. To Shamma, this suggests that the developed brain is even more "plastic" or adaptable than previously realized. We're literally changing our minds.

Scientists are still a bit in the dark in terms of understanding the mechanisms that cause this rapid tuning, but Shamma says that if we can mimic those abilities, it could lead to the development of more effective hearing aids and cochlear implants. In the shorter term, it might help improve automatic speech recognition systems by teaching them to filter out moderate levels of background noise and other acoustical "clutter."

And that brings us to the latest SciAm article (subscription required, sorry). Apparently a team of researchers at IBM's TJ Watson Research Center have managed to create an algorithm for the "cocktail party problem" that outperforms human beings. Why is it so hard, and therefore such a bit deal? It comes down the number of possible sound combinations which quickly becomes unwieldy. Here's how Collins phrases it:

"Whether one person is talking or many, the sound contains a spectrum of frequencies, and the intensity of each frequency changes on a millisecond timescale; spectrograms display data of this kind. Standard single-talker speech recognition analyzes the data at the level of phonemes, the individual units of sound that make up words... Each spoken phoneme produces a variable but recognizable pattern in the spectrogram. Statistical models ... [specify] the expected probability that, for instance, an "oh" sund will be followed by an "n". The recognition engine looks for the most likely sequences of phonemes and tries to build up whole words and plausible sentences."

In other words, speech recognition works a bit like Auto-Correct -- and we all know what can happen when Auto-Correct goes horribly, horribly wrong.

Collins continues:

"When two people talk at once, the number of possibilities explodes. The frequency spectrum at each moment could come from any two phonemes, enunciated in any of the ways each person might use them in a word. Each additional talker makes the problem exponentially worse."

The good news is that such algorithms can simplify the search by focusing on the dominant speaker -- c'mon, we all know there's at least one Loud Talker in any given crowd. A number of shortcuts have been devised in recent years by exploiting this kind of thing. A "bottom-up" approach looks for segments in a spectrogram without a dominant speaker, and sets those segments aside, literally removing them from the equation so the algorithm can focus on finding phoneme sequences in the "clean regions" -- i.e., where there is a dominant speaker. That approach has been adopted by scientists at the University of Sheffield in England, apparently.

Alternatively, you can use a "top-down" approach, devising an algorithm that analyzes trial sequences of the most likely phonemes for all speakers in a given spectrogram. Finnish researchers at Tampere University of Technology exploit this approach by switching between each of two speakers. As Collins explains, "Given the current best estimate of talker A's speech, search for talker B's speech that best explains the total sound." Context is everything, baby. The IBM achieved their "superhuman" automated speech separation by tweaking a "top-down" approach and devising an algorithm to seek out areas on the spectrogram where one talker was bellowing so loudly s/he masked the voices of the other(s).

But you really shouldn't worry too much just yet about secret agents eavesdropping on your party guests: the new algorithms aren't that good. Maybe someday. In the meantime, please to enjoy this classic party scene from Breakfast at Tiffany's to illustrate just how tough the cocktail party problem is likely to be. As one of the YouTube commenters remarked, "It's not a party until someone is laughing and crying at themselves in the mirror."

REPOST: pardon my reverb

Note from your errant blogger(s): Life is currently kicking our collective asses, both professionally and (for some of us) personally, hence the eerie quiet of late at the cocktail party. So much cool stuff, so little time to blog about it! For my part, I've just finished copy-editing The Calculus Diaries manuscript, and am heading off to the San Diego Science Festival, where I will be on a panel on science and science fiction (with David Brin and Scott Sigler, no less!). But I've been blogging for enough years now, that some of you may have missed cool posts from yesteryear. So to tide you over until I get back to blogging this weekend, here's one of my favorite early posts, on concert hall acoustics, inspired by a Jordi Savall performance. And hey, if you hunger for fresher fare, I'm still blogging over at Discovery News.

Back in March, I visited my beloved New York City, having snagged tickets to see a concert performance by the incomparable Jordi Savall, world-renowned master of the viola da gamba and two-time Grammy nominee, whose soundtrack for the French film Tous Les Matins du Monde has sold more than a million copies worldwide. For this event, Savall performed with the Hesperion XXI ensemble and La Capella Reial de Catalunya, both of which he founded. Jen-Luc Piquant is as big a fan of Jordi Savall's music as I am, and we were thrilled at the opportunity not only to hear him perform live, but also by the featured pieces: early Renaissance and Baroque compositions from the Iberian Peninsula region. These are unusual, little-known compositions that Savall has taken upon himself to revive and introduce to classical audiences around the world.

A bit of history: the first Jesuit missionaries in Brazil and Paraguay adapted traditional Amerindian melodies of that region to Christian doctrinal texts translated into the local languages. They even recruited local musicians to take part in the ceremonies, clad in their native garb and playing their own culture's instruments. Inevitably, the two musical traditions merged, combining (according to the program notes) the "European model system, formal structure and contrapuntal techniques... with their own melodic and rhythmic patterns," particularly those associated with native dances. This resulted in some odd, yet affecting hybrids, such as Juan Garcia de Cespedes' Ay qye me abraso -- literally, "I am burning." It features a rhythm based on a popular Mexican dance. The secular aspect is offset by the lyrics: the characters are "panting and sighing because of the excessive heat generated by their emotions at the sight of the newborn Christ." Our own more sonorous, politely composed Christmas carols seem downright emotionally tepid in comparison.

If only the organizers of the event had paid more attention to the venue, the evening would have been an unqualified success. There's no shortage of performance spaces in Manhattan, after all, with decent enough acoustics for this type of music -- the gloriously Gothic St. John the Divine would have been ideal. Alas, the concert was held at St. Paul the Apostle on the Upper West Wide, a small Romanesque-style house of worship whose architectural details added up to a veritable acoustical "perfect storm." Unless one happened to be sitting in the first few rows, the sound was muted, muddy, and each separate part of the complex compositions seemed to reach the audience at different times, so that the percussion always seemed a few beats behind the singers, and/or the instruments. 

Needless to say, the evening was an unmitigated disaster. Several audience members left in disgust, and those who remained to the bitter end were forced to wander forlornly about the sanctuary in hopes of finding one of the "live" spots where the acoustics weren't quite so ruinous. Since it was impossible to fully appreciate the performance, our little group instead found ourselves discussing the scientific aspects of what might be happening to cause the horrific effects.

Composers have always recognized the importance of the acoustics of a given performance space, and many tailored the music they composed to fit those spaces. For instance, Gregorian chants fare well in medieval cathedrals, known for long-reverberation times; ditto for organ music, such as Bach's "Toccata in D Minor." In contrast, Mozart and Haydn composed music to be played in highly furnished chambers, for smaller, intimate audiences. Such pieces lose their clarity when played in highly reverberant spaces. But no one really thought about how to design a concert hall or opera house to achieve optimal acoustics; it was done through trial and error, with little acoustical theory to provide much of a framework to centuries of experiment.

The father of modern architectural acoustics is an American physicist named Wallace Clement Sabine. In 1895, he was a lowly faculty member of Harvard's physics department, who was handed the knotty problem of improving the infamously bad acoustics of the university's Fogg Lecture Hall, part of the recently constructed Fogg Art Museum. Sabine didn't have any particular expertise with sound -- he didn't even hold a PhD (the horror!) -- but he doggedly tackled the challenge  as he would any other physics experiment. He spent several years studying the acoustical qualities both the museum's lecture hall, and the Sanders Theater, widely considered to have excellent acoustics, in order to determine  what might be causing the difference in sound quality.  Specifically, he was attempting to find some objective formula or standard by which to measure and assess the acoustics of performance space designs.

It wasn't an easy task because so many variables had to be taken into consideration. He and his assistants tested each space repeatedly under varying conditions, moving materials back and forth between the two halls -- such as hundreds of seat cushions from the Sanders Theater -- and making careful measurements armed only with an organ pipe and a stop watch. He timed how long it took for different frequencies of sounds to decay to inaudibility under those varying conditions: with and without Oriental rugs, various numbers of people occupying the seats, and so forth.

Ultimately, he was able to determine that there was a definitive relationship between the quality of a room's acoustics, the size of the chamber, and the amount of absorption surfaces that were present. And he came up with the formula for calculating reverberation time, still the critical factor for gauging a space's acoustical quality.

Sabine concluded that the Fogg Lecture Hall's reverberation time was too long -- a spoken word would remain audible for 5.5 seconds, as opposed to the optimal reverberation time of 2.25 seconds -- so there was too much resonance and echo. He solved the problem by outfitting the space with sound-absorbent materials to reduce the "echo effect." His success cemented his reputation and he went on to serve as acoustical consultant for the design of Boston's Symphony Hall, still considered one of the best halls in the world in terms of sound quality. And oh, yes: the unit of sound absorption, the sabin, was named after him.

The field of concert hall acoustics has advanced far beyond Sabine's rudimentary first measurements, although there are still purists who believe that there will always be a subjective element that eludes attempts at strict mathematical description. Nonetheless, using just those sorts of quantifiable tools, Leo Beranek, one of the most eminent acoustic engineers, has identified three basic aspects to achieving a sufficiently good sound in a concert hall: (1) Listeners should be as close to the orchestra as possible; (2) Listeners should have a line of sight to the orchestra so the sound can travel unobstructed; and (3) the interior surface of the hall should be made of a hard material so that sound energy is not absorbed or lost. So an acoustical consultant needs to balance strength, reverberation and clarity requirements when designing a performance space.

Computer modeling has become one of the modern acoustician's most important tools. It turns out that the sound diffusing through a performance space can be modeled as particles of light bouncing around that space, much like a billiard ball bounces around a table in response to being hit by the cue. Acousticians can use those models to calculate the impulse response of the sound. Once the sound field of a modeled space has been determined, it's possible to simulate the sound as it would appear within the real space -- a technique called auralization, the acoustic equivalent of visualization techniques. Auralization software can help acoustical consultants predict, and perfect, how a proposed peprformance space design will sound, based on a wide range of parameters: room size and shape, surfaces, and materials, among other design features.

Auralization is still in its infancy when it comes to concert hall acoustics, but a team of Korean scientists has developed a way to paint "sound pictures" in space: specifically, a technique for controlling the sound in a selected region of listening space. They compare it to painting a picture on a canvas using "acoustical paintbrushes." They mix and overlap sound strokes by playing multiple loudspeakers simultaneously to create interference effects. The end result is a "photocopy" of, for example, a symphony hall's sound picture. The technique is called sound field reproduction.

Getting back to that infamous Savall concert: there were clearly many different factors at work on the night of the botched acoustics. For instance, different musical styles require different reverberation times. If reverberation time is too long, the notes blend together too much, so that it becomes difficult to distinguish individual notes in fast, complex passages. I had been especially looking forward to hearing my first villancico: a religious song performed in the vernacular, incorporating percussive African dance rhythms and a highly theatrical call-and-response effect between soloists and the chorus (tutti). All the pieces that night featured quick tempo changes, unusual rhythms and complex vocal combinations. Throw in the odd percussionist and a woman with castanets and high heels stomping her way through some of the pieces to add even more of a staccato beat, and the result was a hopeless muddle of reflected sound -- because of an overly long reverberation time.

The shape of the performance space plays a vitally important role in all of this, because it affects how sound is distributed in that space. Fan-shaped halls, favored for theaters and early cinemas, lack sufficient side reflections to achieve good sound quality (specifically, the pleasurable sensation of being "enveloped" in the music), while the circular and elliptical shapes favored by many architects tend to focus sound on certain "hot spots," leaving much of the remaining space acoustically "dead." St. Paul the Apostle's main sanctuary features a long gallery topped by a high curved roof -- I believe the technical term is a "barrel vault" -- lined on both sides by decorative archways. I'm not sure if that qualifies as a "shoe box" shape, but the poor acoustics would indicate that it is not. Making matters worse, the floor is mostly wood, with marble tile borders, so too much sound is absorbed. Add in poor line of sight, and you've got that muted, muddied quality to the music in question.

Frankly, the performance by Savall's ensemble du jour would have been better served if it had been given in the subway station. There's a reason so many buskers favor specific spots in New York City's vast subterranean transit network. Alex Case, director of Fermata Audio and Acoustics in Portsmouth, New Hampshire, has conducted extensive studies of local musicians who play regularly in subway stations. I interviewed Case -- who wears a second hat as a professor at Berklee College of Music in Boston -- back in 2004 during a meeting of the Acoustical Society of America about his NYC subway work. Basically, he recorded numerous subway "performances" with portable digital recorders and then analyzed the data he'd collected.

The walls of those stations are made of rigid heavy materials like tile, stone, steel and concrete, all of which are better at reflecting sound waves. This allows sound pressure (volume) to build up naturally, with no need for extra amplification -- that all-important reverberation. The same sound, heard up close, has much less reverberation. (This is also the same reason why so many people enjoy singing in the shower.) In fact, Case likes to mention that there's a very famous tile-and-concrete stairwell outside a professional recording studio in Manhattan that serves as an excellent "reverberation chamber" and has featured on several classic rock recordings.

Not surprisingly, he found that those stations with acoustical qualities in keeping with the established numerical criteria embraced by professional sound consultants, also proved most popular with street musicians. For instance, buskers instinctively seek out locations near hard walls and under low ceilings to amplify their music above the routine din. They also tend to avoid stations with lots of announcements, or major transfer hubs, since there are too many trains running through the station to provide enough time for an extended performance period.

Of course, the very qualities that make subway stations so good for the street musicians can also be blamed for the notorious incomprehensibility of the routine service announcements over the MTA's public address system. (Jen-Luc Piquant firmly believes the audibility of such pronouncements is inversely proportional to the importance of any given announcement.) Amplifying speech requires far less reverberation than music. What sustains musical notes by building up sound reflections causes speech to become mushy and unintelligible. The reflections all mix together so that individual words can't be deciphered. Add in the electronic amplification, and it's little wonder the announcer sounds like he/she is speaking with a mouthful of marbles.

We love subway buskers, and we love Alex Case for choosing to study the acoustics of those performances rather than (or in addition to) the more hoity-toity, elitist performance spaces -- conventional opera houses and symphony halls, for example, which are quite nice in their own right, don't get me wrong. They just exclude a large percentage of the population. Case loves the broad accessibility of the subway, pointing out that some 7 million people ride the NYC subway system every day. Even if only 1 in 10 passengers pay attention to the music, that still adds up to about 700,000 listeners per day, from every conceivable social demographic. In comparison, Carnegie Hall and Lincoln Center, combined, would have to sell out 54 shows every day, on all six main stages, just to compete. So the subway is the concert hall of the huddled masses.

Savall himself has something of Case's "common touch." The Spanish and Portuguese composers featured in the program that night combined high and "low" art, seeking to embrace the "pop culture" of their time. This might explain why there were so many young people in the audience. Classical musical aficionados are constantly bewailing the absence of youth in their audiences. Perhaps it's because the mainstream programs are so predictable: great music, yes, but always the same pieces by the same composers. This isn't an approach likely to appeal to younger, hipper audiences, who are far more likely to gravitate towards fresh alternatives, such as lost or rediscovered historical gems from, say, the 10th to the 18th centuries. That's Savall's specialty. His bio in the program notes offers his success as "proof that early music does not have to be elitist or of interest only to a minority, and that it can and indeed does appeal to an increasingly large and young audience." Savall is making classical music cool.

On the way out after the performance, I picked up a church bulletin, hoping to glean a little information about the building's history and design. Mostly it contained the usual church announcements, but on page 2 there was an item about the new sound system being installed in the sanctuary to address the chronic acoustical problems. It made mention of the need to eliminate certain "dead spots" in the sanctuary, not to mention increasing clarity in the spoken word in the nave. (The church leaders declined, however, to carpet the marble floors and stone pillars to better absorb sound to eliminate the echo effect completely, purely for aesthetic reasons.) Given the fact that Savall is such a terrific ambassador for early music, in addition to being a world-class musician, it's doubly appalling that the organizers would deliberately book his ensemble in such a well-known acoustically challenged space. Yes, it was malice aforethought. Savall et al deserved better. And so did his audience.

all in the family

Jen-Luc Piquant was tres desole in mid-August when she learned (via The Perfect Silence) that pioneering violin maker/acoustician Carleen Maley Hutchins had died at age 98. For those who follow the acoustics of musical instruments, and violins in particular, Hutchins was a legend -- unorthodox, innovative, and completely unapologetic about her work, which occasionally met with skepticism, if not outright hostility. As Margalit Fox wrote in the New York Times obituary for Hutchins: "Working intently and noisily in her home in Montclair, NJ, she helped reimagine the idea of what a violin could be. In the process she designed and built an entire family of violins, eight instruments proportional in size and pitch known collectively as the new violin family or the violin octet." The eponymous Hutchins Consort performs exclusively with her instruments.

My favorite description of her personality comes from Gabriel Weinreich, a professor of physics/acoustician at the University of Michigan: "No respect for authority; a long attention span; scrupulous honesty; enthusiasm for intellectual collaborations; and the willingness to spend a lifetime beating a path through the jungle." I had the honor of interviewing Hutchins (then a spry 94) by phone at her New Hampshire home in 2005 for the American Institute of Physics' Inside Science News Service, and Weinreich's description fits my impressions of her. I never did get around to meeting her in person, which I regret.

It's the no respect for authority and stubborn perseverance that impressed me most: you can't be cowed by the The Man if you're going to be a trailblazing pioneer, and Hutchins had the deck stacked against her from the start. It's easy to forget how far we've come in the last 100 years in terms of women being welcomed in the professional/academic sphere. Hutchins was born in 1911 in Springfield, Massachusetts, grew up in Montclair, New Jersey, and stood out from her peers even as a child. For instance, she loved woodworking and played the trumpet -- both highly uncommon pursuits for girls of that era.

Even more unusually, she attended Cornell University, where she studied entomology and earned a degree in biology in 1933, followed by a master's degree in education from New York University in 1942. This enabled her to land a decent job as a teacher at a private girl's school on Manhattan's Upper West Side. She married chemist Morton Hutchins in 1943, a man who shared her love of woodworking: the couple built their own house together. At the Brearley School, Hutchins made friends with several other faculty members who invited her to join their chamber group. She showed up, trumpet in hand, but it was far too loud for a chamber group. They really needed someone who played the viola. Rather than bow out, Hutchins learned to play the viola, buying an instrument for $75.

While she enjoyed the musical interludes, Hutchins became increasingly frustrated with the limitations of her bargain viola, and asked her uncle, a violin maker, to build her a better one. I guess he didn't want to enable her or something, because instead, he suggested she build one herself -- she did have all that woodworking experience, after all -- and gave her the name of a Russian violin maker in New York City who could help her get started. The Russian man wasn't exactly encouraging, but Hutchins wasn't easily dissuaded. "He didn't think much of a woman making an instrument, but at least he sold me a blueprint and a book and told me how to get started," she told me. And gosh darn it, she succeeded in building her very first viola over the next two years.

From the start, she was serious about the science behind building violins, striking up a long association in the late 1940s with Harvard physicist Frederick Saunders, who studied violin acoustics. She specifically designed instruments he could analyze and pick apart to determine just what gave rise to the distinctive quality sound associated with Stradivari violins, for example. Nor was Hutchins above the occasional scavenging. The New York Times obituary relates my all-time favorite Hutchins story. In 1957, she became enamored of a maple shelf in a phone booth in Columbia University's medical school; it was just the thing for the back plate of a new viola she was building, and she had an "in" at the school: her good friend Virginia Apgar, a doctor at Columbia and herself an amateur maker of violins. As Fox tells it:

One night [Apgar] and Mrs. Hutchins stole into the building with some tools and a replacement shelf stained to match. As Dr. Apgar stood guard, Mrs. Hutchins set to work. To their dismay, the new shelf was a quarter-inch too long. Mrs. Hutchins had a saw and there was a ladies' room nearby. As the New York Times reported afterward, "a passing nurse stared in astonishment at the sounds coming through the door." Dr. Apgar could think quickly. (She had, after all, devised the Apgar score, used worldwide to measure the health of newborns. "It's the only time repairmen can work in there," she said. Spirited out of the hospital, the shelf made a magnificent viola back.

Hutchins excelled at finding the critical balance between two key resonances when making her instruments. First and foremost is the natural wood resonance which can be tuned when the instrument is in pieces; the unattached wooden top and back of the instrument are known as "free plates" during this stage of the process. Traditionally, violin makers would "tune" the plates by carving away the wood underneath to specific thicknesses in order to get the desired natural resonances, flexing the wood plate in his/her hand and tapping them with a fingertip close to the ear to detect the telltale "ring." But it as much an art as a science; Hutchins thought there might be a better, more precise way to tune the plates to the natural resonance, and invented "free plate tuning." She used a loudspeaker to get the plates to vibrate, and spread glitter on top of them, watching where the tiny particles settled in order to find the exact lines that would give her the desired resonances. You can see the process firsthand in this nifty YouTube video featuring Hutchins herself, hard at work, talking about her process and her abiding love for wood:

The other critical resonance for a violin is the air resonance that is produced once the instrument has been assembled. That's the resonance that arises from the body cavity -- kind of like what happens when you blow air across a glass bottle with a narrow neck to produce a tone (and that tone will be higher or lower depending on how much fluid is in said bottle). A violin has two holes that serve a similar purpose. After studying hundreds of master violins, Hutchins and her collaborators concluded that the best resonances always occurred within a very narrow range: the two open middle strings. The D string, second from the bottom, provided the fundamental air resonance, while the wood resonance fell one-fifth higher, on the next highest violin string (violins are tuned in fifths).

Another of Hutchins' innovations came in 1974, when she and collaborator Daniel Haines developed a graphite-epoxy composite to replace the usual spruce wood favored for the top of violin bellies. Choice of wood, and how that wood is treated, has long been a hot area of study by acousticians eager to uncover the "Stradivari secret" -- what gives a Strad that unique quality of sound. (I've written about this topic before here.) Granted, it's a highly subjective analysis -- the BBC organized a blind listening test in 1974, where not even world-famous violinists and violin dealers could all correctly distinguish between a Strad and a modern instrument -- but there does seem to be evidence that treating the wood can enhance an instrument's acoustical qualities.

This was bolstered by the news that made the science headlines a couple of weeks ago during an annual conference in Germany on, um, forest husbandry. Empa scientist Francis Schwarze and a Swiss violin maker named Michael Rhonheimer collaborated on building violins with wood treated with fungus for up to 9 months, and pitted their modern instrument against a Stradivarius in a blind listening test. (Star British violinist Matthew Trusler did the playing honors.) Of the 180 or so people in attendance, 90 chose the fungal violin (dubbed Opus 58) that had been fungally treated for nine months to be the best of the lot, and fully 113 actually thought Opus 58 was the Strad.

Per Horst Heger of the Osnabruck City Conservatory, this could make acoustically superior instruments available to talented young musicians who would never be able to afford their own Strad; now they could have the same tonal quality for less. It all comes down to the quality of the wood, he explains. The fungal attack changes the cell structure of the wood, reducing its density and simultaneously increasing its homogeneity, thereby giving an instrument made with that wood a "warmer, more rounded sound." Antonio Stradivarius might not have benefited from fungus, but he did get help, apparently, from the "Little Ace Age" that hit Central Europe from 1645 to 1715. Trees grew more slowly and uniformly as a result of the longer winters and cooler summers, and that makes for wood with terrific acoustical qualities.

I don't know what Hutchins would say to that, with her instinctive feel for trees and wood. But she knew a little something about re-learning the lessons of the past. One of the most fascinating things she told me back in 2005 was her rediscovery -- while building the first violin octet at the request of composer Henry Brant in the late 1950s -- of lost musical principles once described bby the 17th century German composer Michael Praetorius. (The week Hutchins died, I listened to Praetorius on my morning commute in her honor. It seemed fitting.) Another musically inclined colleague told her about the composer's three-volume treatise, Syntagma Musicum, published in 1619, which detailed musical practices and instruments of the period. 

In particular, Praetorius described an octet family of violins tuned to the same ranges Hutchins was developing. For instance, Hutchins' baritone violin -- similar to a large cello -- featured the same dimension and string length as that pictured in the 17th century text. Ditto for the contrabass violin. "He was writing about something that was common knowledge at the time. That knowledge was lost," she told me. She figured that the tone of these early octets was not up to snuff acoustically, and thus they were replaced when the great 17th century violin makers began building instruments with exquisite resonances ideally suited for the performance needs of the era. Those earlier designs were abandoned, only to be rediscovered over 300 years later by a stubborn violin maker named Carleen Hutchins.

The traditional violin family includes the violin, the viola, and the cello, each different quite a bit in timbre. In contrast, the instruments in Hutchins' octet family range from the seven-foot contrabass to a tiny treble violin tuned a full octave above a standard violin. The sizes are graduated at each half-octave, filling in the gaps in timbre. The result (quoting the New York Times again) is "an even, densely woven tissue of sound, almost like choral music without the words."

An unexpected consequence of the redesign is that the placement of the key resonances shifts in the violin octet instruments. (Fans of the octet claim the instruments correct the acoustic imbalances of traditional violins.) Classical composers wrote music specifically tailored to the strengths of traditional instruments on those key resonances. For instance, Hutchins told me that in Mozart's two-viola quartets placed the first viola through the upper of the two main resonances and the second viola through the lower. With the octet, those resonances shift, which means most classical pieces must be transposed and rearranged before, say, the Hutchins Consort can perform them on this new family of violins. (There are modern composers now writing specifically for the violin octet instruments.)

Similarly, even world-class musicians like cellist Yo-Yo Ma must reacquaint themselves with these new instruments. Ma recorded Bela Bartok's violin concerto using one of Hutchins' alto violins (it has the register of a viola but is played vertically, like a cello), and found he had to play the instrument a bit differently in order to achieve the resonances he needed, for example. Robert Spears, a violin maker who trained under Hutchins (who I also spoke with back in 2005) thinks that the shifted resonances give the octet violins more tonal uniformity across the strings than traditional instruments. "A common complaint among musicians is that as one goes from string to string,  each one can sound like it's on a different instrument," he explained. "Skilled players do a terrific job of minimizing these effects. When suddenly they don't have to compensate for it anymore, it's almost a handicap at first until they realize that particular problem is gone."

Not everyone is a diehard fan of Hutchins' work; I think it's safe to say that she offended her share of colleagues over her long, distinguished career; even her fans would admit she could be abrasive. But even the skeptics admire her fortitude, her perseverance, and there's no doubt she made lasting contributions to the ongoing study of violin acoustics and instrument building. Hutchins herself made seven full octets over the years -- no small feat, since it requires a good 2000 hours to do so. Her "disciples" have built even more. Several of her creations are now housed in museums collections, at the Metropolitan Museum of Art, the Musik Museet in Stockholm, the University of South Dakota, and at the Edinburgh University Collection of Historical Musical Instruments.

The remaining three octets she built are in the capable hands of the Hutchins Consort. Hopefully we can look forward to their continued use in performance (and in the lab). After all, Hutchins said that her instruments "still need to be explored in depth and played seriously by a group of dedicated musicians who will give them the same treatment that is given to learning any stringed instrument." Carleen Maley Hutchins may have shuffled off this mortal coil, but her spirit lives on with the violins she built -- particularly every time they are played.

standing stones in the (inland) sea

Though I don't like to admit it, I grew up in Michigan a quarter mile from Lake Huron. I also have thing for shipwrecks and stone circles like Stonehenge and Avebury. So what do the three factoids have in common? Water. The Great Lakes, shared by Minnesota (barely), Wisconsin, Illinois, Indiana, Ohio, Michigan, Pennsylvania, New York State and the provinces of Ontario and Quebec, are five of the top twelve largest bodies of freshwater in the world, and hold something like 22% of the world's fresh water. They're a little boggling to land-locked folks; you can't see across them in most places and they act more like seas than lakes, but without the daily tides and the salt; 350 million years ago, they were seas that left fossilized coral colonies called Petoskey stones. What you see now, though, is the remnant of glaciers that covered most of North America 10,000 years ago and gouged out those lake beds before filling them with meltwater.

The Great Lakes are deceptively deep and each has a personality of its own. Huron, the one I lived on, is a known ship swallower of oceanic proportions. Of the approximately 4,700 known shipwrecks (with an estimated six to ten thousand total) in the Great Lakes, more than a thousand ships, from ocean-going iron ore freighters to pleasure boats (and airplanes), have been documented as sunk in Lake Huron alone, a large number in the area known as Thunder Bay, just to the north of where I grew up. Lake Michigan, though, gets some of the fiercest storms. It's estimated that 3,000 ships lie in the sands of Lake Michigan.

But ships aren't the only things the waters of the Great Lakes have drowned. The lakes' mean levels rise and fall (PDF) in an approximately 20-year cycle that has to do more with annual snowfall than gravity, so houses and docks and boathouses have a tendency to disappear cyclically as well. Shoreline appears and disappears, is reshaped and eroded and built back up again; wetlands come and go. And as the glaciers pulled away, the lakes rose even more drastically than during this cycle until they reached their present levels, leaving being a number of isolated islands.

One of these islands is Big Beaver Island in Lake Michigan, home to Michigan's very own stone circle. Now there may be another, this one discovered underwater by Dr. Mark W. Holley of Northwestern Michigan College in Traverse City during the course of surveying the shipwrecks and other debris in Grand Traverse Bay. Ah, see? It all comes together: Michigan, shipwrecks, stone circles. Maybe I should clarify before you take a look at the photo of it at right. Don't expected Stonehenge, capital S. It's as much a stone alignment as a henge, which are circular or oval in nature. But standing stone arrangements come in numerous configurations, most of which seemed to have some kind of astronomical significance. Many, like the familiar Stonehenge near Salisbury, UK, were both grave sites and astronomical markers. One acoustics scientist, Rupert Till,has discovered that Stonehenge, at least, reflects music almost perfectly, "making the stone circle an ideal setting for listening to repetitive trance rhythms," as the Discovery News article points out. Rave on, Dude! Of course, whether that's actually what went on there is impossible to tell.

(That circle in the middle of the sonar photo above is an artifact of the imaging tool, a Kongsberg-Mesotech MS 1000 sonar, which produces some eerily vivid images of underwater objects (including a very spooky image of a downed plane on their home page) by pinging soundwaves off them. From watching old  submarine movies, I'm used to seeing a sweeping green line with little blips on it as sonar images, not something this high-def. That high definition is, of course, thanks to software processing programs rather more sophisticated than the echolocation sonar used in WWII subs. Unlike the side-scanning sonar that's towed behind a surface vessel or sub that can give wide, sweeping views of the seafloor, this is a more or less stationary sonar dropped overboard to scan a circular area. Sound waves blanket an area and the time they take to bounce back to the transmitter/receiver tells you how high off the surface the objects returning the sound are, as well as how dense they are, since different materials absorb and reflect sound differently. The software is calibrated for the type of surface that's being pinged, so harder objects, like stones, form clearer pictures than the background sand, for instance.)

Just what function North American stone circles had is still up for grabs; there are very few of them. What there are a lot of is medicine wheels, which may have served a similar function but have a very different design. What you're seeing here looks very similar to the remains of other medicine wheels across the Great Plains (left). What's doubly interesting about this particular set of standing stones is that one appears to have a picture of a mastodon with a spear in its side carved into it. Mastodon remains have not previously been found that far north, so this could be a significant find. There's some question about whether the stone really shows a mastodon, and it's hard to verify the fact since it's in 40 feet of water and covered in algae.

The truly mystifying thing about most stone circles is how the heck they got where they are. North American stone structures are generally made of rocks that could easily be carried by hand from the surrounding area, but European megaliths often weigh from several hundred pounds to several tons, and some of the stones were quarried hundreds of miles away. It's cases like this that really show up human ingenuity and give the lie to the idea that people in history were less sophisticated or dumber than we are. First, they've chipped this slab of bluestone (which describes the color more than denoting a particular type of stone in this case) out of a quarry 250 miles away in Wales. What now? Using gravity, bits of wood, and some rope, a former construction worker in Flint, Michigan shows you how to move two ton rocks with your bare hands:

This isn't exactly an unknown technique; the Egyptians probably used something like it to erect their obelisks, like the one now residing in London, or the one in New York's Central Park. When the Vatican wanted one moved from behind the sacristy of St. Peter's to the piazza in front, it took "900 men, 75 horses and untold numbers of pulleys and lengths of rope," and a lot of scaffolding. So Wally Wallington's feat is pretty impressive. The stones he's moving weight as much as a minivan. Of course, you could just go directly to the minivan itself.

I shudder to think what future archaeologists might make of Carhenge. Maybe it'll wind up underwater too, with a little luck, and rust can take its course. In the mean time, I wonder how a rave there might sound?

[h/t to Geoff Manaugh at BLDGBLOG]

I hear the mermaids singing

Ever since I first heard Brian Eno's Music for Airports (hear a sample here) I've been into electronica and ambient music. WNYC radio show New Sounds' John Schaefer defines ambient music as "music of little or no rhythm, with background sounds and effects as foreground music." As my tired old arena-rock damaged ears get older, I find it immensely soothing and much better to write to than, say, the Stones or The Who, or even, God help me, Van Morrison. I was a band nerd in high school and though I didn't play an instrument, I occasionally "roadied" for our Class AA band because two of my best friends did. One was first chair flautist and the other played piano and percussion. Hanging out with them (and the influence of my classical and opera-loving mother) and another friend's hippie parents got me completely hooked on music of just about all kinds. It wasn't until grad school that I first heard music like Eno's though, and that got me interested in all sorts of things, some of which might be classified as noise by the undiscerning: like found music.

Yeah, you know what I mean: the car horn symphony like this one by Dennis Báthory; the folks whacking out percussion riffs on up-ended 10-gallon paint buckets, the sound of wind in the wires, even whale songs. It's ambient music in its literal sense: Take what's already in your environment and arrange it a little, or use objects that we don't tend to think of as musical instruments to create music. Paul Winter was one of the first musicians to do this with natural sounds, working with whale sounds as early as 1967. Twenty years later, that culminated in his recording of Whales Alive, an extended improvisation with whale song recordings recorded live at the Cathedral of St. John the Divine (where the acoustics, with their seven-second reverb, are by turns magnificent and muddy, depending on the instrument and tempo).

We don't know much about the sounds that whales make (more samples here), except that they are definitely some sort of communication, rather than just echolocation (though they are that, too). Male Humpback whales seem to sing more than females and usually during breeding periods and in known breeding grounds. There's no Rosetta stone for Whale-to-human yet, so we're not sure what they're communicating, or if they're just serenading their girlfriends. They're haunting enough to make some of us want to sing along though, as are some other ambient sounds. T.S. Eliot is far from my favorite poet, but I've always loved this line in the Love Song of J. Alfred Prufrock: "I have heard the mermaids singing, each to each." Since whale songs can be heard for up to 100 miles, I've often wondered what sailors in wooden ships made of the sounds of whales through the hulls, and if that wasn't where the notion of mermaids first came from.

The urge to make music seems to be a universal human characteristic. It's a form of pattern recognition (or Gestalt psychology) as hardwired into our brains as language seems to be. We like to make sense of our surroundings, so we have a hardwired tendency to try to organize the random input from our senses into patterns, whether it's visual input, like Rorschach blots (right), stains on the wall, or sound input. Recognizing music in sounds or sounds in music is often part of musical training for sight reading and composition. This ability to make patterns crops up in odd places and can really come in handy at odd times too. I remember lying in bed listening to a loud building alarm going off for what seemed like maddening hours before I managed to turn it into a sort of white noise rhythm pattern in my mind, at which point I fell asleep to it. We commonly see it as a behavioral tic indicating boredom too. How often have you started to drum your fingers or tap your foot in boredom or impatience? Thus is born many a percussion track, like this one from the British Antarctic Survey’s Rothera Research Station, where the ambient sounds are human-made, and a couple of real musical instruments are added on to round out the composition, as Winter did with whale songs and his clarinet. In this piece, in addition to running water, you can hear lots of percussion on found instruments, wine glasses humming and a different technique that's almost like LP scratching, along with violin and guitar.

Srsly, what else do you do in Antarctica when you're not researching? You form a band!

Engineer and musician Noah Vawter, who's currently a Ph.D. candidate at MIT's Media Lab, takes found music one step farther with Ambient Addition,"a Walkman with binaural microphones" Like Báthory's car horns, Vawter's portable device takes the existing noises around the listener and reinterprets them. In this case, that reinterpretation involves a non-random overlay of sound, something with harmony and rhythm, two key components of music. He accomplishes this with the use of a Digital Signal Processing (DSP) chip, which does pretty much the same thing our brains do: overlays order on chaos. the result is that, instead of isolating yourself from the world around you with a set of headphones plugged into your iPod, you're actually highly tuned into your environment and experiencing the literally found music of it firsthand. Every walk becomes an original composition.

Another example of the intersection of electronics and music, and my new obsession, is Tomoko Sauvage's  water drip performances. Sauvage first came to my attention through a YouTube video of her performing on waterbowls in Paris (It's kind of a crappy quality video, so I won't embed it; here's a better one of her rehearsal with Scottish pop star Momus). She has a set of graduated-sized porcelain bowls filled with varying amounts of water that she plays like a percussion instrument with a couple of wooden kitchen spoons, accompanied by an electronic drone and drum track or electronic shruti box. Same principle as playing a glass harmonium, harmonica or harp (like the wine glasses in the video above): fill a receptacle with water and make it vibrate. The water acts as an amplifier as well as determining what note the receptacle "plays" by how much liquid you make resonate. The water bowls, Instead of being rubbed to make them resonate (as you also do with Tibetan singing bowls), are struck like a xylophone. The cool thing about this method is that the tone can be varied a little by stirring the water, which adds a vibrato. You can hear and see this in the Momus video I linked to above. The struck bowls have a bell-like tone similar to struck singing bowls, one that's deeper and more resonant than glasses. Don't try this at Thanksgiving with the jello salad bowl.

Where Sauvage's work gets back to the ambient is in her waterdrop performances, like this one:

In this case, she's using hydrophones to transmit the sound of water dropping into the bowls and hitting the bowls themselves, the water in the bowls moving, the sounds of her pouring water in, and her disturbing the water and flicking the bowls with her fingers. That kind of plooping sound the water makes pouring into the bowl is due in part to a process called cavitation (the making of a cavity), where air bubbles created by changes in pressure in the water oscillate and explode, creating teeny shock waves. Usually it takes a marked change in pressure for cavitation to occur, but fast-flowing water can do the same thing on a smaller, quieter scale than, say, a submarine or ship propeller. On that louder and larger scale, cavitation can actually erode rock and damage metal. In Sauvage's bowl though, it's more like blowing water through a straw: noisy but harmless. These are normally sounds you wouldn't hear well, if it all, without amplification, and a normal mic wouldn't help much.

Hydrophones were originally developed to collect sounds underwater and transmit them the way land-based mics do, to amplifiers and recording media. This is done with pressure-sensitive transducers "tuned" to the same impedance (how fast sound moves or propagates through a medium) as water, rather than air. Transducers turn the pressure of sound waves into electrical signals that are then decoded by the amplifier. Water is a great acoustic medium because of its density, which gives the sound waves more particles to push around, creating more pressure over the same surface area for the transducers to pick up. Because of this, even faint sounds, like shrimp clicking their little claws to stun fish (remember that shock wave?), can be easily heard underwater with a hydrophone. Sadly, now that we have the tools to hear them, some of those noises and natural songs are being lost in the noise pollution of ship traffic, which is mostly more cavitation noise when it's not sonar or drilling.

Hydrophones are used in a number of different research areas, from studying the aforementioned sea mammal communication and echolocation and estimating their populations and feeding and migration patterns, to hunting subs, studying sound propagation and visualizing sound wave fields (pdf), and monitoring underwater earthquakes and volcanoes. Whitlow W. L. Au, Chief Scientist in the Marine Mammal Research Program at the Hawaii Institute of Marine Biology, University of Hawaii at Manoa, was recently elected President-Elect of the Acoustical Society of America, which gives you some idea of how closely the two fields are intertwined.

But this is the first I've seen hydrophones being used to make music. Is it mermaid music, or music for mermaids?