Bell Laboratories and The Development of Electrical Recording
the Fletcher-Munson Curve from Bell Labs audio research
The Development of Electrical Recording at Bell Laboratories
Radio broadcasting in the United States began in November, 1920 when KDKA, Pittsburgh was granted its call letters and began broadcasting 1. This development represented a challenge to the phonograph, since free music and entertainment was available "off the air". Phonograph recording companies sought to respond by improving the quality of their recordings by adopting new technology.
A key innovator of this technology was Bell Laboratories. Bell Laboratories was the research division of the US telephone monopoly, American Telephone & Telegraph. Research had previously been done by Western Electric, the manufacturing arm of AT&T, and also by Bell System division until "...On January 1, 1925, AT&T established Bell Telephone Laboratories, Inc., in New York City as a joint venture with Western Electric. Gifford [President of AT&T] appointed Frank Jewett president of the new company in charge of 4,000 workers..." 2. Thereafter, a steady flow of ground-breaking technology was developed by Bell Labs during the next 60 years.
Bell and Western Electric research had been working since at least 1914 on technologies for long-distance telephone transmission to span the US. This research also provided the components of what later became the Westrex electrical recording system.
The first complonent was the condenser microphone, developed in 1916 by the brilliant electrical engineer Edward Christopher "E.C." Wente (Ph.D. from Yale in 1918). The condenser microphone became practical with the development of the vacuum tube (valve), used to amplify the low-level signal output of the condenser microphone 4.
Prior to the condenser microphone, since the earliest days of Bell's telephone, the carbon microphone had been used as a transmitter (and continues to be used today). The carbon microphone is basically a generator of voltage, and requires no amplification, at least for telephone transmission over distances up to several hundred miles (or kilometers). However, it does not produce an electrical signal sufficient for transcontinental telephone transmission. The condenser microphone connected to an amplifier was found to produce strong, transcontinental signals.
The condenser microphone design is innovative. It includes two parallel plates, each having an electrical charge, The front plate moves with the audio signal, while the rear plate is fixed. Sound waves moving the front plate cause a slight change in the capacitance between the plates, which allows a change in current flow in the electrical circuit connected to the microphone transmitter. The amplifier circuit then amplifies this low-level signal to a useable level 12.
This condenser microphone, although first developed for long-line telephone transmission, when the US was beginning transcontinental telephone service, had remarkable acoustic performance. In its initial versions with thicker steel plates stretched to give stiffness (so as to raise its resonant frequency above the audio range), it had nearly flat frequency reproduction up to about 6,000 Hz. With improved, thinner plates of aluminium, this reproduction was expanded up to 15,000 Hz, at a time when the acoustic recording process did not record much above 2,400 Hz 9.
Edward C. Wente of Bell Laboratories holding his Condenser Microphone
Prior to the Bell Laboratories research, a number of attempts had been made to develop an electrical recording process that involved microphones and electrical disk cutting. This was to replace the acoustic or mechanical recording technique used since Edison, which involved no electricity or amplification. An early example of electrical recording was the work of Lionel Guest and Horace O. Merriman in Britain. They had developed and demonstrated an electrical recording system using carbon microphones connected to a moving coil recording head 15. Their system produced the earliest issued electrical recording which has come down to us. This was the 1920 recording of the burial of the Unknown Soldier in Westminster Abbey, London on November 11, 1920 14. Guest and Merriman placed 4 carbon microphones inside Westminster Abbey 15, with their recording apparatus outside. The resulting recording was issued by Columbia Graphophone Co. However, the sound quality of the Guest and Merriman system was not satisfactory. British recording companies, including Columbia Graphophone, after evaluation decided not to adopt the Guest and Merriman system. They subsequently went to the U.S., but with similar results, and their system was not further developed.
The sound quality of the 1920 Westminster Abbey recording is sufficiently poor that it is difficult to discern if only instruments are playing, or if there is also singing. The result is significantly poorer than an acoustic recording of the period, except perhaps that it would be difficult for the acoustic horn to capture such a large group. You may hear and judge the sound quality of this 1920 recording for yourself by clicking on the link below. This reproduces the music Abide with Me from this 1920 Westminster Abbey service.
In about 1920, two teams at Bell Laboratories were created to pursue electrical recording. Joseph P. Maxfield and Henry C. Harrison, two of the leading developers of modern electrical components at Bell Labs, were charged with developing a phonograph recording system. E. C. Wente, Donald MacKenzie and Irving Crandall (who had originally hired Wente in 1916) 19 were assigned to develop a sound system for films to be shown in cinema theaters5. Both groups were overseen by Dr. Harvey Fletcher , who previously joined Bell research in 1916.
Dr. Harvey C. Fletcher was a leading research physicist born into a Mormon family in Utah on September 11, 1884. Fletcher studied and worked at the University of Chicago with Nobel Prize winner Robert A. Millikan. In 1911, Fletcher received his Ph.D. in Physics from the University of Chicago summa cum laude. After creating and heading the Physics Department at Brigham Young University in Utah, Fletcher in 1916 became Director of Research at Bell Laboratories, where he oversaw three decades of research and improvement in sound, hearing, transmission, and reproduction.
At Bell Laboratories, Fletcher and his predecessor Irving B. Crandall oversaw research in this area by Joseph P. Maxfield (1887-1977), Henry C. Harrison (1887-1971), Edward L. Norton (1898-1983), K. P. Secord, Donald MacKenzie, Rogers H. Galt (1889- ) 16, and Harold Black (1898-1983). Harold Black had invented the negative feedback amplifier in 1927. Fletcher also supervised Arthur C. Keller (1901-1983), who worked extensively on recording improvements, including binaural and stereo recording (as described below).
Harvey Fletcher, K. P. Secord, and Rogers H. Galt at the Bell Laboratories in New York City
Microphone design was a key part of the Bell Laboratories development of electrical recording technology. Carbon microphones, using loose carbon granules were employed from the earliest days of telephone instruments until today. The carbon microphone alters the transmission of electrical current from acoustic vibration. It requires no subsequent electrical amplification, although the circuit to which it is connected needs to have a voltage/current for transmission. In the 1920s, Bell Labs developed improved designs of carbon microphones, including the so-called "double button" carbon microphone separated the carbon powder from the microphone diaphragm, reducing noise and improving clarity.
The famous Western Electric double-button carbon microphone in the Western Electric 1B microphone housing circa 1925.
The double-button carbon microphone not only reduced noise and also greatly reduced harmonic distortion. These improved carbon microphones and later the condenser microphones described above were manufactured by Western Electric (the manufacturing arm of the Bell system). The Western Electric double-button carbon microphone in the 1B and 1C housing was widely used in radio broadcasting. In broadcasting, the carbon microphone was connected directly to the engineering control panel and then to the radio transmitter. A radio studio from late 1924 with this microphone arrangement can be seen in in the photograph below.
Art Gillham (with glasses) standing to left of Will Rogers in November 4, 1924 during the Eveready Hour, broadcast by WEAF New York. AT&T sold WEAF in 1926 as part of a settlement with GE/RCA to exit broadcasting, and GE and RCA committed to transmit all their network broadcasts via AT&T long distance lines 18. (note the Western Electric carbon microphone in the 1B housing on a stand)
Harrison and Maxfield made a number of developments which, combined together, allowed the creation of an electrical disk sound recording system. Already described is the condenser microphone. Another of these key developments was the matched-impedance amplification system, having such a condenser microphone, linked to a tube (valve) amplifier, with the amplified output voltage driving a moving magnet (called also a "moving armature") cutting head to scribe the sound into a wax master disk.
Bell Laboratories had early licensed the patents covering Lee de Forest's Audion tube. Bell Laboratories discovered a number of wrong solutions which de Forest had introduced - such as fillin his Audion glass tube with ionized gases. Bell Labs determined that the gas was not as effective as a vacuum, and also improved the internal construction. Bell Laboratories engineer Harold Arnold had by late 1914 developed an improved triode vacuum tube, replacing gas by a vacuum and redesigning the electrodes and filaments. This producing an amplifier having low distortion and good linear amplification 20. This newly perfected, efficient electrical amplifier Bell Laboratories delivered to the Bell System to make trans-continental telephone communication practical for the first time.
This amplifier was also an important part of the matched-impedance electrical recording system which was developed by Bell Labs in the early 1920s. Performance was encouraging. Initially, it had a a recording bandwidth from 50 Hertz to about 6,000 Hertz, beyond which its high frequency sensitivity declined11.
Frequency response of the electrical amplification and
condenser microphone of the Westrex system 11
This frequency reproduction range was noticeably superior to the highly variable acoustic process, which (with the best equipment and with the best recording engineers) could reproduce from about 250 Hz to about 2,400 Hz, with rapid reproduction fall-off thereafter. This wider bandwidth added another octave of sound reproduction, along with reduced harmonic distortion and a generally more realistic sound image, including recording of higher harmonics of the musicians. The dynamic range of the sound recording and the background noise were both also somewhat improved.
At the same time, Edward C. Wente and his team had developed the "light valve", a light-emitting vacuum tube that converted audio voltage into variable levels of light to expose movie film. This light output produced a variable density sound pattern on the film. Wente was granted a patent on this technology in 1923, and it was later the basis for the Western Electric "sound-on-film" process.
Wente's early condenser microphone designs continued to be developed into continuously improved microphone transmitter designs. The Western Electric Model 361 seems to have been available by late 1924 and the Model 394 transmitter by late 1926 (although these dates are my estimates based on contemporary literature). According to tests, the Model 361 had had several mid-range resonant frequencies, particularly "...a +7dB resonance at 2.9kHz..." 22. The later Model 394 microphone transmitter sought to improve this and other characteristics. The Model 394 transmitter was widely used in a number of Western Electric microphones for broadcasting, movie and disk recording in different versions from about 1927 6 and into the 1930s.
The improvement of the condenser microphone and its use for 1925-1930 electrical recordings, was key in the quality of the Bell Labs system. It should be added that there is disagreement among experts as to whether the initial recordings made with the Western Electric electrical recording system were made with carbon microphones or with condenser microphones. However, it seems clear that by later 19269 or early 1927, condenser microphones were definitely used in the Western Electric system. This was important not only because of the condenser microphone's expanded frequency range and lower harmonic distortion, but also because even well-engineered carbon microphones produced a low level hissing or sizzling background sound from the vibration of the carbon granules.
However, there were some inconveniences with early condenser microphones. The condenser microphone transmitter required vacuum tube (valve) circuitry to be located near the microphone, since the microphone transmitter required a polarizing voltage to be applied to its condenser plates, and it also required amplification. The output of the microphone transmitter was weak. It also had a high impedance. If this output were sent down a long electrical wire to the recording electronics, it would be degraded by noise. Therefore, in condenser microphones used for recording from in this period, the amplifier was placed in a box below the transmitter, or in another nearby housing. In later Western Electric models, the transmitter was sometimes attached to a tubular housing containing the amplifier circuits and the circuits to provide voltage to the plates. Such a base was initially large, as can be seen in the photograph below: the model 47A, containing the 394 transmitter at the top, and the electronics in the tubular base.
Western Electric 47A Condenser Microphone with 394 transmitter at top and a single vacuum tube (valve) amplifier in base
The initial condenser microphones using the Model 361 transmitter were replaced by improved condenser microphones such as the Western Electric 47A having the Model 394 transmitter beginning in late 1926 and into 1927-1928 in both movie sound recording and phonograph disk recording 6.
Electrical recoding systems previous to and contemporaneous to the Bell Laboratories development may be fairly described as being mostly developed by empirical or 'trial-and-error' development. In contrast, Bell Laboratories engineers, being fundamentally oriented to physics and electrical engineering theory based their design on these scientific concepts. This reflected the advanced electrical engineering eduction of these scientists; some of the few working in such research in that era. In their development of an electrical phonograph disk recording system, the Maxfield and Harrison team based their system on the theory of a many-sectioned electrical filter 5. Bell Laboratories had previously applied filters to the improvement of long-distance telephone transmission. The "theory of electrical filters" was well-developed and mathematically well-characterized. This allowed them to apply these known mathematical solutions to the recording and reproduction problems they were attacking.
Using this approach, a mechanical equivalent of each electrical component of such a filter was identified. For example, an electric filter terminated in a resistor. Bell Labs engineers terminated their electrical recording system in a rubber column in the disc cutting arm - functionally equivalent to a resistor. This approach was used both for the electrical recording system and for sound reproduction systems such as loud speakers. In the case of the loud speaker, the air in an enclosed cabinet could be considered to function like a capacitor in an electrical filter. Since the mathematical solutions in the theory of electric filters had already been solved, this accelerated a scientifically-based development of the Bell electrical recording and reproduction systems.
In the new Bell Labs recording system, the microphone was connected to the matched-impedance amplifier (described above), whose electrical output was connected to and controlled the electrical disk cutting mechanism.
Henry C. Harrison and Norman H. Holland examine a phonograph tone arm
For the disk cutting mechanism, the Bell engineers developed a "moving magnet", - also called a "moving armature" electromagnetic device. In this system, the electrical output of the matched-impedance amplifier was fed to an electro-magnet, causing the cutting stylus to move according to the changes in the magnetic field of the electro-magnet. In this way, the music, in the form of the varying electrical output of the amplifier, going to the cutting stylus via the moving magnet would inscribe the music wave form into the wax master. So, in this electrical system, the microphone and amplifier, moving the stylus, replaced the old acoustic horn mechanically vibrating a sound box which would move the stylus. Since the microphone and amplifier had a broader frequency range and less harmonic distortion and less acoustic resonances, the recording cut into the wax was dramatically better, as could immediately be heard in the first electrical disks released by the Victor Talking Machine Company. (See 1925 - The First Electrical Recording to read about the world's first electrical recording of an orchestra.)
There were a number of limitations of this new electrical system which the early recording engineer needed to overcome. If the electrical signal for the music exceeded certain limits, it could cause the electromagnet-stylus assembly to move too far and become stuck. To solve this, a small spring was added to the cutting head to control stylus excursion.
Another difficulty was in the conversion by the Bell engineers of an electrical component into an equivalent mechanical component. As described above, the Bell Labs design was in part based on the "theory of electrical filters". Electrical filter theory required that the final section of such a filter terminate in a resistor. The mechanical equivalent of the electrical resistor that the Bell engineers developed was a 9 inch long rubber line or rod. This rubber line was added at the end of the system to provide mechanical resistance or damping 5,10. The rubber rod was placed up the inside of the length of the cutting arm. This is the source of the description of the system by the engineers as being the "rubber line recorder".
Cutting Arm of the "Rubber Line" Recorder with the rubber damping rod inside 11
Without the damping resistance of the rubber line, the movement of the recording stylus at high volume levels, particularly of low frequencies, would have been excessive. The damping of the rubber line was one of the essential developments that would make this Bell system linear in its frequency response 8.
However, this rubber tube also introduced difficulties. First, it required careful design and repeated testing of the stiffness and compliance of the rubber column. The coupling of components, the mass of the components, and characteristics of the rubber each had to be varied and tested empirically to discover a combination that gave the disk cutter an essentially flat frequency response. In fact, in the lab, the Bell system was able to achieve a remarkably flat response from 250 Hz to about 15,000 Hz 9, although the total system in use commercially had an upper frequency response to about 6,000 Hz in the early years (progressively expanded each year thereafter). It was also found later, from experience, that the recording engineer would need to carefully pack and adjust the rubber line in the recording arm, prior to each recording session to assure good linearity and response in the recorded sound. In this way, the early Westrex electrical system, like the earlier acoustic recording system. depended on the craft and resourcefulness of the recording engineer to gain the best recording results from the system.
The microphone, plus the amplifier, plus the electromechanical disk cutting mechanism combined together was the new Western Electric electrical recording system, commercialized as the "Westrex" system.
Western Electric engineer George Groves cutting an electrical master.
Groves was assigned to Vitaphone, and later relocated to Hollywood
where he became one of the great sound engineers, winning 2 Oscars 9
By early 1924, the Western Electric system was ready for demonstration.
Western Electric's next task was to interest the phonograph recording companies to license and adopt the Westrex system, to replace the acoustic recording processes used for the last 50 years. The fascinating and often surprising story of this is quest to implement electrical recording of the phonograph disk can be seen by clicking on the link: Licensing the Westrex Electrical Recording System to Victor and Columbia.
Note: Of course, the Western Electric electrical recording system, while being the most successful early electrical phonograph recording system was not the only, or even the first electrical system. Allan Sutton in his fascinating and meticulously researched book: "Recording the Twenties. The Evolution of the American Recording Industry, 1920-29" 13, published in 2009 by Allan Sutton's excellent Mainspring Press describes the history of the development of the various technologies leading to electrical recording. Well worth a purchase and multiple readings.
1 Fisher, Marc. Something in the Air: Radio, Rock, and the Revolution That Shaped a Generation. page xiv. Random House Adult Trade Publishing Group. ISBN 978-0-375-50907-0
2 Adams, Stephen B. and Butler, Orville R. Manufacturing the Future: A History of Western Electric. page 115 Cambridge University Press. Cambridge, UK. 1999 ISBN 0-521-65118-2
3 Lurie, Maxine N. and Mappen, Marc. Encyclopedia of New Jersey. page 68. Rutgers University Press, NJ 2004 ISBN 08-13533-252
4 Fagen, M.D., ed. A History of Engineering and Science in the Bell System: The Early Years (1875-1925). New York: Bell Telephone Laboratories, 1975.
5 Frayne, John G. History of Disk Recording Journal of the Audio Engineering Society, Vol. 33 no 4. page 263 -266. April, 1985
6 Klapholz, Jesse. The History and Development of Microphones. Sound and Communications. September, 1986
7 Measuring Worth website for estimating current values measuringworth.com.
8 page 103. Burns, R. W. The Life and Times of A D Blumlein Institution of Engineering and Technology. Herts, UK 2000. ISBN 0-8529677-3-X
9 page 94 Burns, R. W. op. cit.
10 page 93 Burns, R. W. op. cit.
11 Maxfield, Joseph P. and Henry C. Harrison. Methods of High Quality Recording and Reproducing of Music and Speech Based on Telephone Research. Bell System Technical Journal 5, July, 1926
12 page 4, 5 Eargle, John. The Microphone Book. (Second Edition) Focal Press Burlington, MA 2004 ISBN-13 978-0-240-51961-6
13 Sutton, Allan. Recording the 'Twenties. The Evolution of the American Recording Industry, 1920-29. Mainspring Press. Denver, Colorado 2008. ISBN 978-0-9772735-4-6.
14 page 56. Chanan, Michael. Repeated Takes - A Short History of Recording and its Effects on Music. Verso Books. 1995. ISBN 1-85984-012-4
15 page 364. Hoffmann, Frank W. and Ferstler, Howard. Encyclopedia of Recorded Sound, Second Edition. Taylor & Francis, Inc. July 2004 ISBN-13 9-78041593835-8
16 Thanks to Christine Rankovic, Ph. D. for this information on Rogers Harrison Galt.
17 Jones, W. C. Condenser and Carbon Microphones: Their Construction and Use. Journal of the Society of Motion Picture Engineers. : January, 1931.
18 pages 110, 111. Adams, Stephen B. and Butler, Orville R. Manufacturing the Future: A History of Western Electric. Cambridge University Press. Cambridge, UK. 1999 ISBN 0-521-65118-2
19 page 92-100. Thompson, Emily. The Soundscape of Modernity: Architectural Acoustics and the Culture of Listening in America, 1900-1933. MIT Press. Cambridge, Massachusetts. 1999 ISBN-13: 9780262701068
20 page 334-348. Maxfield, J. P. and Harrison, H. C. Methods of High Quality Recording and Reproducing of Music and Speech Based on Telephone Research. Transaction of the American Institute of Electrical Engineers. February 1926.
21 Frederick, H. A. The Development of the Microphone. Journal of the Acoustical Society of America. July, 1931. New York, New York.
22 pages 116-127. Copeland, Peter. Manual of Analogue Sound Restoration Techniques. British Library Sound Archive. London, UK. February, 2001.
23 Moore, Jerrold Northrop. pages 239-251 Sound Revolutions, A Biography of Fred Gaisberg, Founding Father of Commercial Sound Recordings. Sanctuary Publishing Ltd. London. 1999. ISBN 1-86074-235-1.
If you have any comments or questions about this Leopold Stokowski site, please e-mail me (Larry Huffman) at e-mail address: firstname.lastname@example.org
Full Navigation Menu of www.stokowski.org site (click any button below):
Rosters of Musicians of some Great Orchestras:
Leopold Stokowski - Philadelphia Orchestra Acoustic Recordings 1917-1924:
Leopold Stokowski - Philadelphia Orchestra Electrical Recordings 1925-1940:
Leopold Stokowski Recording Discographies and Listing of Concerts:
Other Information about Leopold Stokowski:
Leopold Stokowski and Development of Recording: