Carol Greene’s New True Book of Robots. Not Kraftwerk’s idea of a robot but good eyebrows nonetheless. (illustration credit 8.14)
With the Kraftwerk single “Numbers,” the child in the machine was Speak & Spell, an educational tool developed by Texas Instruments and used by E.T., the witch-fingered namesake of the 1982 Miami vocoder classic “E.T. Boogie.” Speak & Spell was the first successful commercial application of Linear Predictive Coding, co-conceived at Bell Labs by George Doddington. Michael Noll, a former associate of Doddington’s, remembers being “flabbergasted.” “Doddington took the idea over to Texas Instruments, and up popped that toy. We couldn’t believe someone had made a chip and was using it to synthesize speech … and selling the damn thing for forty bucks! We saw an ad for it in Bloomingdale’s and rushed down. I said, ‘Obviously it doesn’t work.’ ” The man performing the demonstration then pushed the button and the machine said: “Spell ‘house.’ ” Noll responded with ‘Good grief!’ ” Incidentally, the day we spoke, Noll had just sold his own Speak & Spell on eBay for $26.
Florian Schneider had also acquired a device that translated English into Japanese. “The Texas Instruments Language Translator—it was similar to Speak & Spell. That’s what we used on all these Computer World things. At the time, there was no sampling, so we had to record it live off the tape. In real time. Thank you, Texas Instruments.”
The Collector of Artificial Voices moved on. Next acquisition: the DECTalk, a text-to-speech device now operated by the cheek of physicist Stephen Hawking. “These people, they were trying to play Frankenstein and have the homunculus speaking,” says Florian. “It’s really perverse. I got this wonderful machine from DEC [Digital Equipment Corporation]. Then the inventor, Dennis Klatt, got throat cancer and could not speak any more. The ghosts you are waking, they are going to kill you, Doctor Faust.”
Speak & Spell was an educational toy equipped with its own “mystery word” button, also the first commercially successful use of Linear Predictive Coding (LPC), popularized by E.T. and used by Kraftwerk on “Numbers,” a beat perpetuated in the tailgating scene in “The U,” an ESPN documentary about the University of Miami football program in the Eighties. (illustration credit 8.15)
VERBOT SPEKTRALZERLEGER
In the early Sixties, the ghost of Dr. Faust appeared in the Siemens vocoder in Munich as engineers fed Goethe to the machine. “Die Sonne tönt nach alter Weise.” The sun makes music of old. This frequency specter was committed to magnetic tape along with speaking winds, pleading Volkswagens and instructions for children. It was a demo.
Florian Schneider—who used his own voice to admire the sun on “Autobahn”—first heard the Siemens tape in his early twenties, while studying music at the Akademie Remscheid, near Düsseldorf. “There were speaking chimes and bells. Childish songs about making the bed. A quiet whispering. Motors talking. ‘Please don’t put me in the wrong gear.’ All these weird things. It really stimulated my fantasy in those days. Oh, what is this! And where can I see it? But no. No way.”
Not for sale and costing 50,000 DM, the Siemens vocoder made its television debut in 1962 in a commercial for the Swiss chocolate Camille Bloch, with a steam engine chanting “Shok! Shok! Shok!” In 1966, the composer Peter Thomas used the Siemens model for a rocket-launch countdown in the German science fiction serial Space Patrol. “I thought about how I could make a marriage between two instruments, a voice and a cello,” Thomas told The Independent in 1999. “The marriage was in the vocoder. I asked the cello player to play a long note. He asked, ‘How long a note?’ I said, ‘Until Christmas.’ Then I spoke the countdown.”
The vocoder spectrogram I received from Siemens appeared no less paranormal than any of the others (phantasm clouds, ferns, a surprised fish skeleton, etc.) and translated to English as “automatically run local networks.” According to Siemens, its vocoder was involved with automated directory assistance and news satellites. And according to Henrik Teller, a retired officer in the Danish Signal Battalion, a Siemens-made vocoder called Elcrovox had been deployed by the German armed forces (Bundeswehr) in 1968. Said Teller, “In those days, everything was ‘need to know’ and ‘nice to know’—I did not need to know.”
Produced by German composer Peter Thomas in 1966, the soundtrack for the science fiction TV show Raumpatrouille (Space Patrol) may be the first vocoder record. Includes the track “Danger for the Crew.” (Courtesy Fontana Records and Allen “Doc Strange” Goodman) (illustration credit 8.16)
Promotional photo of the Siemens vocoder, circa 1969. Siemens helped fund the early speech ciphony work of Oskar Vierling, a pioneer of electro-acoustic instruments and director of the Laboratory Feuerstein in Upper Franconia during World War II. (Courtesy Siemens Corporate Archives, Munich) (illustration credit 8.17)
Frank Gentges, a Cold War Secure Voice consultant who has heard just about every vocoder this side of the Berlin Wall, thinks the Elcrovox was one of the best versions on the market back then. “The Elcrovox could handle those guttural sounds in the German language that we did little to focus on—those back of the throat noises. It was better sounding because of superior fine tuning of the design that US designers did not get to do because of rigid schedules.”
“I never saw the Siemens studio, actually, but I heard about strange things going on there,” says Holger Czukay of the Cologne group Can. “The vocoders in the Fifties and Sixties were secret units. It was a really sophisticated thing at the time. Siemens manufactured it themselves. They were working with the vocoder and paper strips with holes.”
In 1966, an article published in The Journal of the Acoustical Society of America revealed that Siemens had registered a patent (# 594976) under the name Schmidt, describing “ideas similar to those of [Homer] Dudley.” It had been filed in 1932, a year before Hitler came to power.
GOETHE BUST
In Salomo Friedlaender’s 1916 story, “Goethe Speaks into the Phonograph,” Professor Abner Schorr attempts to “trick” Goethe’s cadaver out of its voice in order to woo a girl. After violating Goethe’s crypt, the professor calibrates the poet’s formant resonances and recreates his voice box inside a head bust. He then uses a phonograph (a popular instrument of woo) and a bellows to inhale and record the leftover acoustic vibrations from Goethe’s study. During playback, Goethe’s vibrations fall asleep in mid-thought and start snoring. The girl swoons. Jealous of his own invention, the professor chucks the fake Goethe larynx in front of an oncoming train. German media theorist Friedrich Kittler would later write, “Goethe’s bass frequencies, vibrating in infinity … remained unmeasurable.”
LABORATORY FEUERSTEIN
According to Manfred Schroeder, the German physicist who designed the first voice-excited vocoder for Bell Labs, Siemens had been developing vocoders prior to the war. “Did the Germans use the vocoder in the Thirties? I think the answer is yes. We knew a little about German efforts by a speech engineer from Siemens named Friedrich Vilbig. They had patents. I’m sure they had several [vocoders] on an experimental basis, used in the lab, so to speak. But I don’t think there was any field use of the German vocoder. We would have known.”
In 1999, a vocoder prototype chucked by the Nazis was dredged from the Schliersee, a lake in upper Bavaria. According to documents declassified by the NSA in June 2009, the vocoder had been under development in Germany in the early Thirties, both at Siemens and the Heinrich Hertz Institute for Vibrational Research, the latter being run by synthetic speech pioneer Karl Willy Wagner, a German contemporary of Homer Dudley’s. Wagner’s employees included Harald Bode, who went on to design vocoders for Robert Moog in North Tonawanda and Dr. Fritz Sennheiser, the name on your headphones and the vocoder that would drive Neil Young fans to mutiny.
In 1942, another Hertz acoustic scientist, Dr. Oskar Vierling, opened the Laboratory Feuerstein in Upper Franconia near Bavaria. Inventor of the Grosstonorgel (a vacuum tube oscillator-based organ), Vierling joined the Nazi party to expedite his speech-privacy research bu
t did not remain in good standing. (Vierling’s staff was reported to have operated “in tolerable harmony,” save for one Dr. Zappe, who was arrested for “irregularities.”) To avoid Allied bombings, the Feuerstein lab relocated deeper into the Alps in 1942, doubling as a hospital. On May 15, 1945, a week after V-E Day, British cryptanalyst Alan Turing made a secret trip to Feuerstein following intercepts of what was believed to be vocoded test transmissions between Vierling’s lab and Hanover during the war. That summer, the Allied Target Intelligence Committee (TICOM) raided Feuerstein, but not before Vierling stashed the bulk of his equipment in a bomb-proof vault hidden behind a fake wall. “This done, he awaited developments.” Among the interrogated were Vierling, Fritz Sennheiser and a man named Wolfgang Martini.
The Nazis wanted the technology destroyed; but the British wanted it saved. TICOM would allow Vierling’s lab to continue its many projects. Under TICOM supervision, the Feuerstein lab worked on “high-speed transmitters for agent use,” a vocoder based on “the Dudley type,” a Wobbler, acoustic torpedoes, anti-radar coating for submarines, a Speech Stretcher and a pocket calculator. Oskar Vierling’s productive summer would end with his arrest late that August, as he was forced to sell his crypto-patents to Switzerland. The Feuerstein Lab became property of the German Catholic Church.
THE ESCAPE OF DR. ZAPPE
When interrogated by TICOM’s Lieutenant Tompkins (no relation) in August 1945, Dr. Zappe discussed the “melody channel in the design of the synthetic.” Zappe also claimed that the Feuerstein Lab in Bavaria had built seven vocoders since 1941. (Only one unfinished model was seized.) The German army had also commissioned Feuerstein’s director, Oskar Vierling, to construct an artificial speech machine named “Anna 43.” Zappe believed Vierling had been dallying on the vocoder to get more money. In February 1945, Vierling had Dr. Zappe arrested by the Gestapo, though Zappe escaped by digging himself out of a jail in Ebermannstadt. Vierling then put a hit out on Zappe for 10,000 DM, accusing him of high treason. Dr. Zappe, who claimed to be the greatest physicist in the world, was never captured.
After the war, Vierling’s associate Fritz Sennheiser opened Laboratory W in a farmhouse previously occupied by British troops near Hanover. There he received an order from the Central Encipherment Office to design a vocoder to guard against wiretapping—billed as “diplomatic service.” Sennheiser’s vocoder habit would survive the postwar denazification (Entnazifizierung), a time when engineers and physicists found themselves either drastically switching careers or recruited by American tech companies for the Cold War. According to Sennheiser’s son Joerg, his father had been “interviewed” by American Signal Corps officers as early as 1944: “The people who questioned him wore uniforms of the American Army but spoke German, with no accent.”
While Sennheiser remained in Germany, Friedrich Vilbig was hired by MIT and performed speech encoding at the Air Force Cambridge Research Center. “He was collared, so to speak,” says Manfred Schroeder. “They may not have had any choice. I think most of them went gladly.”
Dr. Fritz Sennheiser at age 33, photographed at the Heinrich Hertz Institute for Vibrational Research in Berlin. Dr. Sennheiser worked for Oskar Vierling before opening the Laboratory W in Hanover, focusing on voice ciphony after the war. The Sennheiser 20-channel vocoder would later be used by Herbie Hancock and Neil Young. (Courtesy Fritz Sennheiser) (illustration credit 8.18)
This is believed to be part of a vocoder prototype chucked in a Bavarian lake by the Nazis, circa 1945. (Courtesy Klaus Schmeh) (illustration credit 8.19)
RATHER BELONG TO THE REALM OF THE UNREAL
For Dr. Werner Meyer-Eppler, who served as a radio engineer for the Kriegsmarine, postwar “rehabilitation” included linguistics, running barnyard recordings through an artificial larynx, and lecturing on the vocoder. Appointed as Director of Phonetics at the University of Bonn, Meyer-Eppler called the vocoder spektralzerleger, the spectral decomposer. An information theorist, he was interested in how the vocoder could separate speech from the intelligence it conveyed while also multiplying the “channel bed” capacity. Publishing essays with titles like “Observations During the Retarded Feedback of the Language,” Meyer-Eppler was perhaps the first to realize the vocoder had a future in electronic music.
“In a certain sense the vocoder was the initial machine to open the path to composing an electronic studio in Cologne without being used as a musical instrument,” says Elena Ungeheuer, who earned a PhD studying Meyer-Eppler’s influence on electro-acoustic music. “It was the scheme of the vocoder—the functioning of the vocoder. Robert Beyer, then a sound engineer at West Germany’s Radio Cologne (WDR), recognized now we can form sound without being restricted to keyboards. There’s been controversy about the role of composers for the birth of the electronic studio in Cologne. But the phonetician Meyer-Eppler started the process with his report about the vocoder.”
Meyer-Eppler’s 1949 lecture for radio engineers in Detmold would be widely regarded as a pivotal moment in the history of electronic music. Any speech beginning, “It sounds like a homunculus or a robot” showed much promise for the future, as if rescuing the word from the impotent robot dictator featured in Otto Rippert’s 1916 film Homunculus der Führer. In attendance that day was Dr. Robert Beyer, a radio bigwig who would establish Cologne’s fabled Electronic Music Studios in 1951, and Herbert Eimert, a composer and the host of WDR’s chirpily named “Musical Night.” Can’s keyboard player Irmin Schmidt learned of the vocoder through Eimert. “It was one of the first electronic instruments which came from a totally different background,” says Schmidt. “Not for music but military. At that time, the whole electronic studio in Cologne, when they started, the whole science about it came from other sources and not the music.”
Dr. Werner Meyer-Eppler in the Dead Room recording cabin, Institute for Phonetics, University of Bonn. A radio engineer and vocoder pioneer who researched infra-acoustics and noise detection for the Kriegsmarine, Meyer-Eppler literally drove colleagues mad by putting a delay on their own voices inside headphones. “First they get crazy, then they get outside themselves.” (illustration credit 8.20)
During the Detmold speech, Meyer-Eppler assured his audience that what they were about to hear would “rather belong to the realm of the Unreal than to Sober Science.” He played records of animal sounds and “voiced” them, in German, through an Electrolarynx. It wasn’t exactly “Bozo at the Zoo,” but the Unreal quickly had Sober Science chugging a beer-bong tube. Meyer-Eppler then played some Bell Labs vocoder standards, provided by Homer Dudley during a visit to Bonn in 1948. There was “How Dry I Am,” in triplicate, followed by “Suzy Seashells,” “Barnacle Bill” and “Goodnight, Ladies.” A singing tube of toothpaste that claimed to buff the paint right off your car. The vocoder wishing itself happy birthday in the pitch of a man pretending to be a woman pretending to be Mrs. Featherbottom. There were “grotesque scenes with a single speaker” and “Voice of Energy,” a song that Kraftwerk admitted to borrowing in German.
“We pretty much copied the demonstration of the Bell Vocoder and did the whole thing in the German version,” says Florian. “We borrowed this from Siemens and used it as interlude between other pieces.”
“With a means of research like the vocoder, it becomes possible to test out all its possibilities,” said Meyer-Eppler. “Even without a determined purpose.”
PLEASE MAKE A BEAT WITH THIS
Meyer-Eppler’s wedding of vocoder to Melochord in 1951 is a spectral drone rivaled only by the Fantasy Three’s “Biter’s Dub” (thirty-two years later), as if the voices inside your head were controlled by electric toothbrushes, just 4/4 away from the future. Try it with your Philips Sonicare 3000.
ITEM 300001522431
The best way to call upon the ghost of Werner Meyer-Eppler, it turned out, would be through a network of computer spies called “eBay.” In 2006, the year a man auctioned his soul online for $500, “Prototype VOCODER of German 70s Electronic Pioneers” appeared in the
virtual marketplace. To the automated sniper, it was Item 300001522431. Bidding commenced at $3,800, a reasonable figure for a machine that once uttered the word “Autobahn” a few times and coughed.
This large green box was the Barth Musicoder, developed at the Physikalisch-Technische Bundesanstalt and then compressed into an unfinished customized unit for Schneider. (Renamed after World War II, PTB was originally PTR—Physikalisch-Technische Reichsanstalt—where Karl Willy Wagner had designed a vowel synthesizer in 1936.) The Musicoder first saw action on “Ananas Symphonie,” a song featured on the 1973 Kraftwerk album Ralf und Florian. The machine was photographed on the back cover, somewhere behind names spelled out in blue cursive lights, Florian in fluorescent.
“We still had this vocoder sleeping under our studio in a storage place,” says Schneider, who was selling it through a dealer. “So I put it on eBay.
The EMS Vocoder 2000. Recorded in the mid-Seventies, the EMS vocoder demo included stock market reports, a track called “Two teenage girls’ grisly news,” and a woman describing her dreams of little talking animals. (Courtesy EMS) (illustration credit 8.21)
How to Wreck a Nice Beach Page 15