The Birth of Electrical Recording – Part 1
By Allan Sutton
The following is a revised and expanded version of several chapters that originally appeared in the author’s Recording the ‘Twenties (Mainspring Press, 2008)
Radio’s popularity posed a technological, as well as a commercial, challenge to the recording industry. Even the primitive radio loudspeakers of the early-to-mid 1920s delivered greater volume, wider frequency range, and a more accurate rendition of studio ambiance than the best acoustical phonographs and records. For the first time, listeners were hearing music reproduced with a relatively high degree of accuracy, and performed without the sonic contortions required by the acoustic recording process.
Although the acoustic process had been refined over the years, it had undergone little fundamental change since the nineteenth century. It was an entirely mechanical process, employing a simple horn to focus sound waves on a circular diaphragm of mica or other material, which vibrated in response to those sound waves to drive an engraving stylus. The results were captured on a wax master disc, which was then plated to produce a permanent matrix from which sub-masters and metal stampers were generated.
No microphone or amplification was involved in the acoustic process, nor was there the ability to edit or modify the finished recording except by primitive mechanical dubbing methods. Control over input was limited to the physical placement of performers in the studio, or to trial-and-error experimentation with different horns, diaphragms, and cutting heads. The state of the sound-recording art peaked in 1912, with the introduction of the Edison Diamond Disc, then stagnated.
Singers — crowded around metal recording horns and performing at full voice, with a studio orchestra huddled just a few feet away — sometimes complained they were unable to hear themselves above the din of the accompaniment. The acoustic method’s low sensitivity and erratic frequency response required that adjustments be made for some instruments. Violins and violas were replaced by Stroh instruments, horned contraptions sporting a metal resonator in place of the wooden body. Low woodwinds were substituted for cellos, tubas for stringed basses. Bass and snare drums, which could cause over-cutting of the wax, were moved to the far reaches of the studio, if not banished altogether. A full symphony orchestra was not recorded in the United States until 1917, and even then, the results barely hinted at the size of the ensemble.
The use of horned Stroh violins, like those seen in this 1920 photograph of J. C. Beck’s Orchestra, was one of many work-arounds necessitated by the insensitive acoustic recording process. (Bain Collection, Library of Congress)
The acoustic process provided no means for the engineer to monitor what was being recorded, and instantaneous playback of the wax master was impossible without inflicting damage on the master that rendered it unusable. The recordings systems suffered from multiple resonant points that could be corrected only to a limited extent, by laborious trial and error. Photographs taken in the Gramophone Company’s studio in London, and Columbia’s studio in New York, show recording horns wrapped with cloth strips to damp some of the resonances.
Even when such primitive corrections were successful, they were likely to be negated in playback by yet another set of resonances inherent in the acoustic phonographs of the day. Victor’s recording and reproduction systems in particular were plagued by marked mid-range resonances that produced a disconcerting “honking” effect.
Perhaps the public might have continued to accept acoustic recordings indefinitely, had it not been for the advent of radio and the consequent awareness that more accurate sound reproduction was indeed possible. As Bell Laboratories’ Stanley Watkins later observed,
“The fight [between radio and phonograph] was an uneven one as long as the quality of the recording was limited to the possibilities of the old acoustic method. The radio broadcasting technique with its sensitive microphone pickup allowed the artists freedom of action, permitted the use of full symphony ensembles, and made possible great improvement in quality through an ever-increasing knowledge of the use of studio acoustics.”
The initial interest in electrical sound recording, however, came not from the record companies, but from the telephone industry. Many late nineteenth-century experimenters had attempted to make direct electrical recordings using telephone parts. The technology proved to be of no practical use to the commercial recording industry because of the telephone’s intentionally limited frequency range, coupled with the inability at that early date to amplify the electrical signal. Emile Berliner experimented with telephonic recording in 1896, as reported many years later by his associate, Fred Gaisberg. “The result,” Gaisberg recalled, “was a thin metallic thread of sound. The experiment was years ahead of its time.”
The Early Western Electric Experiments
The amplification problem was solved with the advent of Lee De Forest’s audion tube. By 1915, the Bell Telephone system was employing Dr. Harold D. Arnold’s vacuum-tube amplifier in long-distance telephone transmissions. At the same time, Arnold proposed that systematic research into electrical sound recording and reproduction be undertaken by Western Electric, where Henry C. Egerton had already patented an experimental electromagnetic disc-record pickup.
Henry Egerton’s patent for an electromagnetic pickup,
filed in November 1914
As might be expected of a telecommunications company, Western Electric’s early experiments in electrical sound recording and reproduction were applied largely to telephony. The company’s first commercially produced electrical recorder was Henry Egerton’s 1918 telephone answering machine. The cutter, which employed a principle similar to Egerton’s electromagnetic loudspeaker of 1917, recorded vertically cut wax cylinders. Although the machine was suitable for recording telephone calls and office dictation, it was neither intended for, nor capable of producing, commercial-quality musical recordings.
In 1919 Henry B. Wier, another Western Electric engineer, filed a patent application for a complete electrical sound-recording and playback system. Wier employed an obvious holdover from the acoustic process in his use of a recording horn to focus sound on the microphone. He was able to eliminate much of the distortion that plagued the acoustic recording process by using electrical wave filters to correct resonances in the system — the first practical application of frequency equalization.
Other components of Wier’s system, including the single-button carbon microphone, multiple microphone inputs and mixing controls, vacuum-tube amplifier, master gain control, electromagnetic disc cutter, and switchable loudspeaker and headset monitors, were adapted from the prior work of Egerton and other Western Electric engineers. However, Wier made the mistake, from a business standpoint, of specifying that each performer be confined to an individual, fully enclosed booth. Each booth was to be equipped with a widow through which to view the conductor, and was topped by a conical roof with a microphone inserted at its apex. Whatever its merits from an engineering standpoint, Wier’s concept was utterly impractical for commercial use.
Henry Wier’s proposed system of isolating individual performers in separate booths, whatever its merits from an engineering standpoint, was impractical
for commercial use.
Other shortcomings in Wier’s system were quickly addressed by Western Electric engineers Edward Craft and Edwin Colpitts, who filed a patent application on an electrical recording process in November 1919. Wier’s specification of individual musicians’ booths was immediately discarded. The use of relatively insensitive carbon microphones, another weakness in Wier’s system, was overcome by substituting multiple condenser microphones. 6 Many other components, however, were carried over from Wier’s process. In their patent application, Craft and Colpitts discussed at some length the advantage their system offered over the acoustical process:
“In making records for reproduction in the well-known types of sound reproducing machines, it has been necessary to take great precautions, particularly with respect to the relative location of the artist and the recording mechanism, and to employ artists who are specially trained in record making in order to obtain a record which will reproduce sound with any degree of faithfulness. Thus it has been common for the artist in the case of a voice record to sing or talk into a horn or mouthpiece and to vary the separation of the artist and horn to obtain the desired tonal effects. In the case of instrumental music or in the case of duets or an ensemble of singers, great care has been necessary in grouping the singers or artists relative to the recording point in order to obtain the desired result. In view of the difficulty of training artists and also in view of the difficulty of grouping a large number of instruments for efficient recording, it has been proposed to intercept or pick up the sound waves at a plurality of points and conduct them either acoustically or electrically to a common recording point… The artist or artists merely enter the room or auditorium in which the sound receivers are located, and without regard to the recording apparatus, proceed with their performance.”
Craft’s dispersed placement of microphones and use of multiple channels clearly had the potential to produce stereophonic recordings. Unfortunately, that possibility was not explored at the time. Instead, the multiple signals were mixed to a produce a monophonic recording.
The Craft–Colpitts system saw no use in the commercial disc record industry, but it was briefly adapted to provide synchronized sound to motion picture shorts in 1922. On Friday October 27, 1922, Craft demonstrated his system, synchronized to accompany an animated film, to an audience of electrical engineers at Yale University — the first public demonstration of an electrically recorded phonograph record reproduced by a fully electronic phonograph. Further attempts to develop the system for commercial use were soon scuttled, however. In early 1923 two Western Electric sales executives, George Evans Cullinan and Elbert Hawkins, decided that potential profits from licensing the system were likely to be insufficient to justify further development of the Craft–Colpitts system.
Charles Hoxie, General Electric, and the Pallophotophone
At General Electric, Charles A. Hoxie was also developing an electrical recording system, refining some work he had undertaken for the U.S. Navy during World War I. Unlike Western Electric’s electromagnetic system, Hoxie’s was an optical system. He filed a patent application for a basic photoelectric recording device on April 13, 1918, following up with an improved device in May 1921.
Charles Hoxie (top photo, right) with unidentified assistant in General Electric’s Schenectady laboratory. A complete Pallophotophone setup is pictured, with the recording unit to the right. The lower photo, from 1922, pictures only the projection unit; the system had not yet been adapted to disc recording.
Although the original invention was designed to record radio signals on photographic film, Hoxie began to adapt it for commercial applications after the war, at first for motion pictures, and then for disc recording. On December 27, 1921, a patent application was filed on his behalf for a complete electrical disc-recording system employing a photoelectric microphone, amplifier, and electromagnetic disk cutter. By 1922, experimental Pallophotophone recordings were being made on film, and development of disc-mastering capabilities was also under way.
Charles Hoxie (center) demonstrates the Pallophotophone to RCA executives James G. Harboard (left) and David Sarnoff (right) in May 1923.
Hoxie named his system the Pallophotophone — literally, “shaking-light sound.” It was an apt allusion. The sound-collecting device, or Pallotrope, was a photoelectric microphone employing a light beam focused on a tiny, spring-mounted mirror that vibrated in response to sound waves. A short flared horn, attached to the front of the device, served rather inefficiently to collect and focus the sound.
A simplified explanation of the Pallophotophone system, published by Brunswick-Balke-Collender after it adopted the process in 1925.
By late 1922, it was clear to General Electric that Hoxie’s system had potential in the commercial recording market, and he received their backing to make refinements. In 1925, the Pallophotphone system would be adopted by Brunswick-Balke-Collender, with less-than-satisfactory results.
Merriman and Guest’s Electro-Mechanical Hybrid
While work progressed at Western Electric and General Electric, many independent inventors were experimenting with electrical recording processes on their own, in the United States and elsewhere. The first publicly issued electrical recordings were made in England by Horace O. Merriman and Lionel Guest, although the process was not entirely electrical. On November 11, 1920, they recorded portions of the burial ceremony for the Unknown Warrior at London’s Westminster Abbey via a cable link to carbon microphones placed inside the building.
Announcement of the first issued electrical recordings,
Merrriman, as an officer in what would soon become the Royal Air Force, had been assigned in 1917 to develop a loudspeaker with sufficient volume to be heard from ground to air. When the R.A.F. abandoned loudspeaker research at the end of World War I, Merriman stated that he and Guest “considered what peace-time use could be made of the findings already made in the research for an electrical speaker. We decided to develop a method of making phonograph records by electricity using the Fessenden vibration motor.”
The Fessenden vibration motor was an electro-mechanical hybrid, driven by a microphone and amplifier, but activating a mechanical cutter. The cutter proved to be the weak link in the system. Lacking the sophisticated damping that would become the hallmark of Western Electric’s all-electric cutter, it produced recordings with high levels of distortion, particularly in the louder passages. Nevertheless, the improved frequency response provided sufficient impetus to pursue the process.
An illustration of the Merriman-Guest system, showing the Fessenden
Guest and Merriman designed the first self-contained recording van and set about making test records, initially only of speech. The team was soon experimenting with musical recordings as well, setting up in Columbia’s London studio, where acoustic and electrical recordings were made simultaneously. Comparing the two version, Merriman recalled, “The range of tone was greater on the electrically made records, but there was considerable distortion.” The process was soon judged unsuitable for Columbia’s use, and the relationship was terminated.
For the Westminster Abbey recordings, horns were attached to the carbon microphones, which were placed throughout the abbey and connected to the recording van by cables. In the end, only two musical selections were deemed acceptable for release. Pressed by Columbia and issued privately as part of a fund-raising project for the abbey, the record enjoyed modest sales, and a copper matrix was donated to the British Museum.
Guest and Merriman then spent a month recording organist Marcel Dupre at Notre Dame Cathedral in Paris before departing to the United States at the request of the Submarine Signal Company in Boston. It was a short-lived affiliation, and Guest, Merriman, and his wife went on to rent an apartment in Queens, where they set up an experimental electrical recording studio. Columbia made a series of experimental electrical recordings during November 1921, possibly using Guest’s and Merriman’s equipment. These tests, beginning with a session by Gladys Rice on November 3, 1921, are documented in the Columbia files, 11 but they were quickly suspended, and no issued records resulted.
Having made some technical strides, Merriman recalled that in 1923 he and Guest were invited to make simultaneous recordings during regular commercial Columbia sessions, placing their microphone alongside the recording horn. The resulting electrical recordings clearly exhibited greater frequency response and higher fidelity than their acoustic counterparts, in Merriman’s estimation. But in the meantime, Columbia had passed into the hands of receivers who had no interest in developing electrical recording, and Guest and Merriman abandoned their work. Back in England, the Gramophone Company appointed Brenchley E. G. Mittell to investigate electrical recording processes in November 1923, with no discernible results.
Orlando Marsh and the First American
Electrical Disc Recordings
In the United States, Orlando Marsh had been developing an electrical recording system since approximately 1914. A 1931 advertising flyer declared, “Seventeen years ago, Marsh instituted the first electrical recording lab in the world.” At that time, Marsh is known to have been employed by George K. Spoor’s Essanay movie studio in Chicago. It seems likely that Marsh was responsible for the Spoor Sound-Scriber, a cylinder-record system designed to be synchronized with motion pictures. In 1977 researcher Tim Fabrizio discovered the device, along with a cracked celluloid cylinder, in the vault of the International Museum of Photography and restored it to working order.
Once repaired and played, the Spoor cylinder turned out to be a promotional skit for the process, on which a speaker declares that the recording “is accomplished by special telephonic apparatus. That is all I can say about the system.” Although it is impossible to say definitively whether the recording was electrical, Fabrizio noted a “thin, hollow, even garbled character…unlike any acoustical or home recording I had ever heard. Yet, there seemed an odd sensitivity to peripheral noise.”
The earliest confirmed Marsh disc recording, made in the yard of Chicago’s Essanay movie studio (John R. T. Davies, via Malcolm Shaw)
The earliest confirmed Marsh disc recording (matrix #2, a test pressing of which was discovered by the late John R. T. Davies), was of the George Spoor and the Wood Brothers Quartet singing “Bells of Shandon.” According to its handwritten label, the recording was made “in the open air 12 ft distance in the yard of the Essanay Co.” The recording probably dates to to the autumn of 1921. Marsh continued to record at Essanay through late 1922, then consolidated his office and studio in Chicago’s Kimball Building.
(Above) Orlando Marsh recording in the Chicago Theatre, 1924; note the old-fashioned phonograph horn being used to focus sound on the microphone. (Below) Orlando Marsh in his laboratory, date unknown.
By then, Marsh was producing electrically recorded masters for his own Autograph label, as well as for several short-run custom labels that included Messiah Sacred Records, Crown Records, Greek Record Company, and Ideal Sacred Records. Although these were the earliest electrical recordings to reach the American market (albeit primitive ones) — beating Columbia and Victor by three years — the labels carried no notation to that effect. The claim would not appear on Autograph labels until 1925. At that point, Marsh declared himself “The Originator of Electrical Recording,” but he never patented his process.
Early Experimentation at the Major Companies
Among the market leaders, Thomas Edison had experimented sporadically with telephonic recording, to no avail. After World War I, he had even attempted to make recordings using surplus military radio equipment. Recalling those experiments, he stated, “I found when I tried [radio] for recording there was too much mutilation of sounds, which is rather difficult to overcome.”
Frank L. Dyer, a longtime Edison associate, filed a patent application for an electromagnetic recording head in February 1921, but apparently nothing was done to develop it, and Thomas Edison remained emphatically opposed to the process. His company would be the last to convert to electrical recording, one of several factors that led to its demise in 1929.
For a newcomer like the Brunswick-Balke-Collender Company, however, electrical recording must have seemed promising. In December 1920, Percy L. Deutsch, Brunswick’s vice-president and grandson of company founder J. M. Brunswick, initiated formal research into electrical recording. Although circumstantial evidence suggests that Deutsch was aware of General Electric’s experimentation with electrical recording processes, the initial experiments were carried out independently at Brunswick’s Chicago headquarters. Deutsch entrusted much of this work to inventor Benjamin Franklin Meissner, who had earned a reputation as an expert in wireless torpedo-guidance systems during World War I.
The Talking Machine World for December 1921 reported that Meissner had “for some months been working in the Brunswick experimental laboratories here [in Chicago] on various methods for converting sound waves into electrical waves, and reconverting these back to sound waves on the phonograph record.” Meissner conducted experimental electrical sessions at the Brunswick studio during much of 1921. Unfortunately, paper documentation of these sessions has vanished along with Brunswick’s early recording ledgers. Test pressings are rumored to survive, but to date, none has been reliably reported.
In December, TMW also broke the news of Brunswick’s experiments with wireless remote disc mastering in Chicago. On November 22, an operatic performance was transmitted from the Auditorium Theatre to a Magnavox receiver in the Brunswick laboratory. There, TMW reported, “the electrical waves were switched from the Magnavox directly to the recording apparatus.” Despite an apparently promising start, no commercially issued records resulted from Meissner’s experiments, and Brunswick seems to have abandoned its electrical experiments in 1922.
As Meissner was winding down his work at Brunswick, Albertis Hewitt was undertaking similar experiments at Victor. Hewitt and James W. Owen, another Victor engineer, had been experimenting with microphones since 1916, when they patented an improved design for use in “the recording or reproduction of sound.” Hewitt went on to patent many other devices relating to electrical recording and reproduction over the next eighteen years, all of which were assigned to the Victor Talking Machine Company or the Radio Corporation of America. However, when Hewitt began experimentation in earnest at Victor in 1922, it was not with his own equipment, but with Pallophotophone equipment loaned to him by Charles Hoxie.
Hewitt’s experimental electrical installation was completed at Victor’s Camden studio on December 7, 1922, and the next day he conducted the first of many test sessions, beginning with staff pianist Myrtle Eaver. More tests were conducted over the next two weeks, involving Eaver and tenor William Robyn, with musical director Joseph Pasternack voicing his approval of the results. A final report on the Pallophotophone tests was drafted at the end of the month and apparently was buried, after which no more was heard of the device at Victor. Hewitt, however, continued to make some experimental recordings from radio broadcasts during 1922–1924 using an electrical recorder of his own design. In 1923 he undertook further microphone experiments for Victor.
In the end, nothing came of Hewitt’s research, and Victor continued to record acoustically. Probably unaware of Hewitt’s secret experiments, orchestra leader Paul Whiteman invested heavily in the electrical recording process of an unnamed English inventor in 1923, hoping to license it to Victor. For his efforts, Whiteman earned only a rebuff from company executives.
Frank Capps also experimented independently with electrical recording. On November 10, 1923, he recorded former president Woodrow Wilson’s Armistice Day speech, as broadcast on over radio station WEAF (New York). Capps — who allegedly was later involved in leaking news of Western’s Electric’s proposed Victor deal to Louis Sterling at Columbia’s English branch — sent his masters to be processed by the Compo Company in Canada, a venture headed by Emile Berliner’s son Herbert. It is tempting to speculate that Capps’ electrical masters were the impetus for Herbert Berliner’s own experiments, which resulted in the first Canadian electrical recordings.
Western Electric Courts the Recording Industry
While the phonograph companies were abandoning their in-house experiments, and Orlando Marsh was puttering with his homemade electrical equipment, Western Electric’s engineers were making steady progress toward a high-quality, commercially viable electrical recording system during 1922–1923. The team of Joseph P. Maxfield and Henry Harrison had recently taken over much of the project, signaling a definitive change in corporate attitude toward electrical recording methods.
Thus far, Western Electric’s engineers had worked under highly controlled conditions in laboratories that had little in common with concert halls or commercial recording studios. However, Maxfield was now determined to deal with the variables inherent in recording live performances in public venues. He had already experimented with remote electrical recording, establishing a wireless connection from New York’s Capitol Theatre to Western Electric’s experimental recording laboratory and broadcast station at 463 West Street in late 1922. By 1923, Western Electric was regularly making test recordings via the remote link from the Capitol Theatre. The company also made experimental recordings from radio broadcasts, including excerpts from the 1923–1924 New York Philharmonic Symphony Orchestra’s broadcasts over radio station WEAF.
The Capitol Theatre’s cavernous space presented an especially difficult challenge to the Western Electric team. After much experimentation in the theater, the engineers determined that the microphone placement needed to replicate what was heard by an average member of the audience was forty feet above floor level, and forty feet in front of the stage. The quality of these early electrical recordings varied tremendously, as surviving test pressings demonstrate. Several Western Electric experimental pressings have surfaced in recent years, the earliest of them a Capitol Theater performance dated July 20, 1923. Other surviving test pressings include public performances by the New York Philharmonic under Willem van Hoogstraten, made in December 1923, and some January 1924 recordings from WEAF radio broadcasts.
Maxfield emphasized the importance of the studio monitor, volume level indicator, and potentiometer in his process, establishing a degree of control unattainable with the acoustic process:
“Without the monitoring system, the fact that a record is unsatisfactory cannot be ascertained until the master record is made, plated, and reproduced…. In the case of “acoustical” recording from a symphony orchestra, the orchestra must play so that the fortissimo is suppressed and the pianissimo amplified in order to drive the stylus within proper bounds. With the present system, such an orchestra may play with natural force and effect, the current from the amplifier being kept within proper limits by manipulating the potentiometer as suggested by monitoring with loudspeaker and voltmeter.”
In October 1923, Maxfield filed a patent on an improved electrical recording system. Well aware of the failings of the earlier, cumbersome Wier and Craft–Colpitts processes, he greatly simplified the apparatus. At this juncture, Maxfield appears to have still been concerned primarily with the recording of live rather than studio performances, noting in his patent application, “The object of the present invention is to produce master phonograph records electrically without interfering with the public performance of the artist or artists.”
While Maxfield satisfactorily addressed the technical aspects of an electrical recording system, its suitability for commercial applications had so far gone largely unexplored. Little commercial demand could be anticipated for Maxfield’s live recordings, other than as a source of broadcast material. Consequently, Maxfield set out to refine his system for commercial studio use. In December 1923 he filed a patent application on a “studio for acoustic purposes,” stating,
“The object of the invention is to provide a studio in which sounds may be recorded or broadcasted with substantially all the natural effects that an auditor listening directly to the sounds would receive… More specifically, the invention provides a studio in which the walls are damped by a hanging curtain or applying other damping material to the walls, damping them to a degree such that the reverberation will be between .5 of a second and 1.0 second.… The curtains may be hung on horizontal poles or rods by any suitable fasteners which may be slideable on the rods, whereby the curtains may be adjusted to cover any desired surface to control the damping.… The ceiling as shown is not damped, but the floor is substantially covered with a heavy rug. Smaller rugs may be used on the floor and damping material may also be used on the ceiling if desired.”
Maxfield’s design for an electrical recording studio would be
adopted by Victor in 1925.
Henry Harrison made improvements to the electromagnetic cutter in early 1924. Charts included in his patent filing depict a fairly flat frequency response curve ranging from 35 to nearly 8,000 cycles per second. In contrast, the very best acoustic recordings could only offer a range of approximately 200 to 3,000 cycles per second, and few studios other than Edison’s performed even that well. However, much of the experimentation at Western Electric had been carried out using nonstandard disc formats designed to take full advantage of the new electromagnetic cutter, without regard for the needs of commercial producers. Oversized, vertically cut, and lacking the abrasive fillers required in commercial shellac pressings, these discs were superior from an engineering perspective, but they were totally incompatible with the millions of phonographs already in homes.
Anticipating resistance from an industry that was heavily invested in the standard ten- or twelve-inch lateral-cut shellac disc, the Western Electric engineers began to tailor their system to that format. The result was a recording curve designed to compress a modern, wide-range electrical recording into a groove configuration and disc format that were relics of the nineteenth century. With a reduced frequency range of approximately 100–5,000 cycles per second, the new Western Electric process still offered dramatic improvement over the best acoustic recordings, but fell far short of what could be achieved under laboratory conditions.
Columbia and Victor Go Electric
With a commercially viable system finally complete, Western Electric approached the Victor Talking Machine Company. In February 1924, Victor executives George W. Smith and Fenimore Johnson visited the Western Electric laboratories for a demonstration of the new electrically recorded discs. As they soon discovered, the process would not come cheaply. Western Electric demanded royalties on sales of all records made by their process, and further stipulated that Victor’s studios be rebuilt to Joseph Maxfield’s exact specifications.
Victor president Eldridge R. Johnson, coping with health problems and declining revenues from record sales, tabled the Western Electric proposal. The Victor Talking Machine Company had always developed its technology internally, but the Western Electric deal would require the active involvement of outsiders. In addition, the swift transition that adoption of the system would entail was at odds with Eldridge Johnson’s conservative approach to product development. Just four years earlier, he had declared to the press, “It will take twenty-five years more to perfect the talking machine.”
According to an oft-repeated story, Western Electric made its initial offer only to Victor. In the meantime, the tale continues, a bit of industrial spying was under way that would rob Victor of its potential edge. Under the supervision of Russell Hunting, Western Electric was pressing 16” test records at Pathé’s Brooklyn plant, which was the only U.S. plant equipped at that time to press the oversized discs. According to this tale, which appears with some variations in several early phonograph histories, Hunting leaked word of the process to his old business associate, Louis Sterling, at Columbia’s London headquarters. Purloined Western Electric tests are said to have arrived in London on December 24, 1924, with Sterling setting sail for the U.S. two days later, frantic to negotiate use of the Western Electric system for Columbia.
Unfortunately, this widely circulated account is seriously flawed in many respects, and it is contradicted by dated test pressings. Sterling did indeed sail to the United States in December 1924, but for the purpose of acquiring rights to the Western Electric system for English Columbia, under the same terms that Western Electric had already offered to both Columbia and Victor in the United States.
In fact, Western Electric had begun making test recordings for both of those companies many months before Sterling’s visit, as proven by a surprisingly large number of surviving test pressings. The earliest of these electrical tests to surface thus far, made for Columbia, shows a recording date of August 25, 1924, in the wax. Many other Columbia electrical tests exist that show dates throughout the late summer and autumn of 1924 in the wax.
Two Columbia-Western Electric tests, both from September 1924. By that time, electrical tests were being produced in sufficient quantity that a special label was introduced for them. (Courtesy of Kurt Nauck)
Columbia made some of its most notable performers available for these early Western Electric tests, including violinist George Enescu and soprano Florence Macbeth. Although files for the earliest tests have not been located, the excellent sound quality on surviving test pressings is clearly indicative of Western Electric’s work. The earliest surviving confirmation in Columbia’s files that Western Electric equipment was indeed in use is a notation for a session on November 10, 1924. Clearly, a Columbia–Western Electric alliance had been forged well before Sterling’s December dash to the States.
At the same time, Western Electric was also recording tests for Victor, despite Eldridge Johnson’s apparent lack of interest. Electrically recorded Victor test pressings, showing dates in the wax ranging from October 7 to December 17, 1924, survive in a private collection. Interestingly, neither Harry nor Raymond Sooy, Victor’s chief recording engineers, mentioned these sessions in their memoirs. Harry Sooy recalled having first been apprised of “three or four records submitted by the Western Electric Company” on January 3, 1925. It is therefore likely that these early test sessions, which are not documented in the surviving Victor files, were conducted in Western Electric’s studios rather than Victor’s.
Although Victor was clearly considering the Western Electric process during the autumn of 1924, it took news of the impending Columbia–Western Electric deal to force Eldridge Johnson’s hand. Threatened with obsolescence at the hands of his old rival, Eldridge Johnson finally assented to Western Electric’s terms, which included an advance payment of $50,000 in addition to the royalty clause that had caused earlier caused him to balk. The deal was a closely guarded secret — so much so, that no mention of it appears in the minutes of Victor’s managing committee.
On January 27, 1925, Western Electric dispatched Joseph Maxfield to Camden to lay out the wiring for Victor’s first electrically equipped studio in Building No. 15. The Western Electric equipment was shipped to Camden on February 2 and arrived the following day. The first electrical session to be held there — an experimental piano solo recording by one Mr. Watkins — occurred on February 9. Over the next several days, experimental sessions continued with Helen Clark, Elsie Baker, Olive Kline, and other Victor studio artists.
While the Western Electric installation was under way at Victor, Columbia was readying its own Western Electric system for commercial use. In January 1925 the company had Art Gillham, “The Whispering Pianist,” make a series of electrical recordings. Gillham was an excellent choice to demonstrate the new system’s capabilities. His subdued crooning style was poorly suited the old acoustic system, but it registered quite well with the microphone. The results were good enough that three Gillham selections, recorded on February 25, 1925, were accepted for release.
Just one day after Gillham’s electrical Columbia session, the Eight Famous Victor Artists (a traveling promotional troupe featuring Billy Murray and Henry Burr) were assembled at Camden to make comparison recordings of “A Miniature Concert” using Victor’s acoustic and Western Electric’s electrical equipment. Initially, the acoustic version was approved for release, but in April there was a change of heart on the part of Victor management. Instead, the electrical tests, covering two sides of a 12” disc, were approved for a July 1925 release on Victor 35753.
Let the conversion begin: The Victor Recording Book sheet for the electrical version of “A Miniature Concert,” which was recorded as an experiment but was then approved for release in place of the acoustic version.
The “Miniature Concert” comprised the earliest electrical recordings to be released by Victor. They were not, however, the first Victor release to use an electrically recorded master. That honor is held by the Mask and Wig Club Male Quartet’s rendition of “Joan of Arkansas,” recorded on March 16, 1925, and released on Victor 19626 a month before “A Miniature Concert.”
At the end of February, with finalization of its Western Electric deal virtually assured, Columbia became the first major record producer to convert to fill-time electrical recording. Acoustic recording sessions for full-priced releases were suspended on February 28 at Columbia’s New York studio (acoustic equipment remained in use for several more years, but only for recordings allocated to Harmony and other low-priced labels).
Two of Gillham’s February sides were listed in the May 1925 Talking Machine World advance list for a June release on Columbia 328-D, the earliest electrical recordings to be issued by Columbia. In the same listing were four sides recorded electrically during a March 31 public performance by the 850-voice Associated Glee Clubs of America — the first “live” electrical recordings to be issued in the United States.
There was nothing in the new listings or advertisements that might alert the public that Columbia was employing a new recording technology, nor would there be for another year. The only clue, other than an obvious change in sound quality, was a circled-W logo in the pressing, required as part of the licensing agreement with Western Electric. Even that small hint was absent on some of the earliest pressings.
Columbia’s adoption of the electrical process had the unforeseen effect of driving the company into the cheap-record market. Having invested heavily in improvements to its acoustic studios in 1924, Columbia decided to recoup its costs by launching a low-priced label that would continue to use acoustically recorded masters. The result was the introduction of Harmony, a 50¢ brand, in September 1925. Velvet Tone, a companion label using the same masters and couplings as Harmony, followed in the summer of 1926. Both labels continued to use acoustically recorded masters through 1929, although the occasional electrical master (probably recorded for the full-priced line but rejected) found its way into the series.
Victor lagged a bit behind Columbia in its conversion. The electrical sessions of February through mid-March 1925 were still considered trials, although they yielded some recordings that were approved for release. The Western Electric contract was finally signed on March 18, and at the end of that month, Victor retired the recording horns in the Camden studios. However, the company was in the process of acquiring a new studio location in New York at the time; thus, Victor’s first New York electrical sessions were delayed until July 31, 1925.
Both companies began releasing electrical recordings with some regularity beginning in the early summer of 1925. However, neither Victor nor Columbia publicly acknowledged the conversion during 1925–26, allowing themselves time to dispose of obsolete acoustic stock while building new catalogs from scratch. The closest Victor came to publicly acknowledging the new process was Eldridge Johnson’s misleading statement, in response Brunswick’s introduction of the electric Panatrope in August 1925, that the company would soon introduce a new system representing “the ultimate in sound reproduction.” 25 Johnson coyly refused to elaborate on the new recordings to a New York Times reporter, even though they had already been on the market for several months. But the change was immediately obvious to dealers and consumers alike, and by the end of 1925 Victor dealers were openly referring to the new process, even if the manufacturer was not.
Victor’s Canadian branch took the opposite approach, heralding the new electrics in July 1925 with a national advertising campaign, and initiating deep price cuts on its now-obsolete acoustically recorded discs. The impetus might have come from Herbert Berliner’s upstart Compo Company, which had begun marketing electrically recorded discs on its Apex label in Canada. “New Victor V.E. Process a Master Stroke in Recording,” the ads proclaimed. “All the new July releases out today are recorded by the new V.E. process.”
Victor’s Canadian affiliate was the first to publicly announce the company’s conversion to electrical recording, in July 1925. Its American counterpart waited until 1926, as did Columbia.
In October 1926, Western Electric loaned Joseph Maxfield to Victor to pursue further improvements in the electrical process. Appointed as Victor’s manager of research and development in September 1927, Maxfield was given free rein to remake the Victor studios to his specifications. Remote recording locations were added or upgraded, including the Philadelphia Academy of Music and New York’s Liederkranz Hall. Camden’s former Trinity Baptist Church, converted to a Victor studio during the acoustic era, was thoroughly overhauled, including replacement of the original organ. A New York Times reporter who toured the studio observed,
“Hidden from view is the arched roof to which boomed hasannas and hymns… a flat, sound-proof and false ceiling of burlap is better for recording. An organ is there, to be sure, but it is a special one recently installed, and now there is a microphone before it. … Downstairs, where prayer books had been stored … is some $150,000 equipment bearing trademarks of Western Electric, Electrical Research Products, and Victor Talking Machine.”
For a time, the church did double duty as a temporary Vitaphone sound stage, with the lower level used for filming. Films were shot as silents, and the actors then dubbed their parts onto synchronized discs in the main church recording studio.
New, unfamiliar equipment and studios required that studio engineers be retrained or even replaced. Nathaniel Shilkret, one of the few veteran Victor musical directors to make the transition successfully, recounted his company’s problems in adapting the new process:
“Almost everything that had been learned about orchestration and recording seemed useless. The musician’s favorite tricks in orchestration became obsolete; the recorders’ art of handling the recording horns had no more value…. No doubt you will be interested to know that our first successful recordings were with the symphonic orchestras, large choirs and whispering vocalists. Then came the Salon Orchestra which improved immensely over the old recordings, after most of us were convinced that this new way of recording an intimate style of orchestra would never do at all. The piano quality of the new recordings, while not perfect, is surely superior to the old recordings. The tenor voice gave us plenty of grief for a while. At first they sounded rather thick, like baritones. At times, hollow; but all voices finally were conquered. And to think that all this has happened in about one year and a half.”
Nathaniel Shilkret (front row, third from left) and orchestra in a
Maxfield-designed Victor studio.
Pressings were proving to be a weak link in the new system, with dealers complaining that the surfaces were noisy and prone to premature wear. The increased surface noise resulted from use of a coarser, more abrasive pressing material developed for the new electric discs, while the tendency toward premature wear resulted from the more heavily modulated groove.
One of Maxfield’s solutions to the latter problem was to slightly smooth the master recording by high-speed mechanical burnishing “at a pressure which is reasonably constant and of just sufficient magnitude to cause a very slight surface flow of the material without macerating it.” Charles O’Connell, a later Victor recording director, took a dim view of the practice, recalling that masters “Went flawless into these laboratories. They emerged pitted, peaked, and perverted. I say perverted because in some instances, in an effort to reduce the scratch that inexpert handling had brought to the records, a polishing stone was run through the grooves, eliminating some of the scratch and all of the high frequencies that give music color and brilliance.”
The general public, still playing its records on steel-needle acoustic machines with tracking forces measured in pounds rather than grams, would scarcely have noticed such technical flaws. Victor’s record sales rebounded in 1926, jumping to nearly 32 million copies from the previous year’s 25 million. The leap into electrical recording had come at a high cost to Victor’s shareholders, however. In July 1925, the company announced that it was suspending its quarterly dividends in view of “important improvements in the product [that] will require considerable outlay of funds.”
COMING IN PART 2: Herbert Berliner, General Electric, RCA,
and the Minor-Label Systems
“A New Invention” (re: Marsh’s Kimball Building studio). Billboard (January 13, 1923), p. 58.
“Advance Record Bulletins for June 1925.” Talking Machine World (May 15, 1925), p. 157.
“Advance Record Bulletins for July 1925.” Talking Machine World (June 15, 1925), p. 166.
Biel, Michael Jay. The Making and Use of Recordings in Broadcasting Before 1936. Dissertation, Northwestern University (1977), pp. 284–285.
Brooks, Tim. Columbia Master Record Book — Vol. 1. Westport, CT: Greenwood Press, 1999.
“Camden in Film Field.” New York Times (August 26, 1928), p. 98.
Craft, Edward B., and Colpitts, Edwin H. U.S. Patent #1,540,317 (filed November 25, 1919).
Dyer, Frank L. “Magnetic Recorder.” U.S. Patent #1,544,379 (filed February 16, 1921; issued June 30, 1925).
Egerton, Henry Clifford. “Phonographic Transmitter.” U.S. Patent #1,246,895 (filed November 25, 1914; issued November 20, 1917).
— . “Telephonic Recording and Reproducing Apparatus.” U.S. Patent #1,284,623 (filed February 1, 1918; issued November 12, 1918).
Fabrizio, T. C. “Before the Jazz Singer” (re: Spoor Sound-Scriber). Antique Phonograph Monthly (V:5, 1977), pp. 3–6.
— . “The Spoor Sound-Scriber and its Relation to the Sound Synchronization of Motion Pictures,” and “Transcription of the ‘Spoor’ Cylinder.” Antique Phonograph Monthly (V:6, 1977), pp. 5–8.
Giovannoni, David. E-mail to author re early Victor–Western Electric test recordings (September 15, 2007).
Guest, Lionel George William, and Merriman, Horace Owen. “Improved Means for Recording Sound.” British Patent Office: Patent Application #141,790 (filed January 18, 1919; issued April 19, 1920).
Guest, Lionel George William, and Merriman, Horace Owen. “Improved Means for Recording Sound.” British Patent Office: Patent Application #141,790 (filed January 18, 1919; issued April 19, 1920).
Harrison, Henry C. “Device for the Transmission of Vibratory Energy.” U.S. Patent #1,663,884 (filed May 5, 1924; issued May 27, 1928).
“Historic Gramophone Records — Major Guest and the Abbey Service.” London Observer (December 12, 1920), p. 17.
Hoxie, Charles A. “Production of Phonographic Records.” U.S. Patent #1,637,903 (filed December 28, 1921; issued August 2, 1927).
— . “Recording Apparatus.” U.S. Patent #1,456,595 (filed April 13, 1918; issued May 29, 1924), assigned to General Electric Company.
Marsh Laboratories, Inc. (advertising flyer, 1931).
Maxfield, Joseph P. “Phonograph System.” U.S. Patent #1,661,539 (filed October 2, 1923; issued March 6, 1928).
— . “Studio for Acoustic Purposes.” U.S. Patent #1,719,481 (filed December l5, 1923; issued July 2, 1929).
Merriman, H. O. “Sound Recording by Electricity, 1919–1924.” Talking Machine Review (June 1976), pp. 666–670, 680–681.
Nauck, Kurt. Vintage Record Auction #33 (containing a large group of early Columbia–Western Electric test pressings). Spring, TX: Nauck’s Vintage Records (April–May 2003), p. 11.
O’Connell, Charles C. The Other Side of the Record, p. 126. New York: Alfred A. Knopf (1947).
Owen, James W., and Albertis Hewitt. “Microphone.” U.S. Patent #1,509,818 (filed August 25, 1916; granted September 23, 1924), assigned to the Victor Talking Machine Company.
“Record Opera via Wireless.” Talking Machine World (December 15, 1921), p. 118.
“Sales by Class of Record and Total Sales of Records by Units, Years 1901 and 1941 Inclusive.” Exhibit in: U.S. Dist. Court, D.D. of N.Y., January 26, 1943.
Shilkret, Nathaniel. “Modern Electrical Methods of Recording.” Phonograph Monthly Review (June 1927), p. 382.
Sooy, Harry O. “Memoir of My Career at Victor Talking Machine Company.” Unpublished manuscript, n.d. David Sarnoff Library, Princeton, NJ.
Tennyson, James R. “Oh, Canada!” New Amberola Graphic (July 1987), p. 5.
“The Future Development of the Talking Machine.” Talking Machine World (July 15, 1920), p. 16.
Victor Talking Machine Company. Managing Committee Minutes, Vol. 1 (1924–1926).
— . Recording ledgers (Sony archives, New York); data courtesy of John R. Bolig.
“Victor Talking Machine Co. Omits Quarterly Dividend.” Talking Machine World (July 15, 1925), p. 110.
Watkins, Stanley. “Madame, Will You Talk?” Bell Laboratories Record, August 1946 (Vol. XXIV, No. VIII), p. 291.
Whiteman, Paul (David A. Stein, editor). Music for the Millions, p. 5–7. New York: Hermitage Press, 1948.
Wier, Henry B. “Recording of Music and Speech” (U.S. Patent application filed August 14, 1919). The patent was later divided into recording and playback sections, with the recording portion (#1,765,517) not being granted until June 24, 1930.
© 2019 by Allan R. Sutton. All rights are reserved.