First sound recording First stereo sound recording

First sound recording

1. The technology

2. The human subject

First sound recording

Jean-Marie Constant Duhamel (1797–1872) discovered that a pen attached to a tuning fork would record its vibrations on a piece of paper, when the fork was struck. Thomas Young (1773–1829), in 1806, applied Duhamel's apparatus to a rotating cylinder coated with wax. Neither, of course, constituted the recording of sound in air, and neither was capable of reproduction. [Morton 2004]


Phonautogram 29, 'Timbre du Cornet à piston', 1857

'Phonautogram' is the term used by Édouard-Léon Scott de Martinville (1817–1879) to describe his experimental sound recordings. The earliest such, from as early as 1853/4, still survive, but the sound content is too poor to reproduce intelligibly (although a very brief example is played in Patrick Feaster's presentation to the 2011 ARSC conference, available on YouTube, identified by Feaster as notes played on a guitar).

In 2010 Patrick Feaster, of First Sounds, produced a comprehensive annotated discography of all Scott phonautograms currently known to exist. In this he wrote cautiously about an 1857 cornet recording—phonautogram 29, 'Timbre du Cornet à piston'—saying 


. . . early investigation suggests that item 29 documents several eight-note sequences played on a cornet à piston between each of which Scott stopped the rotation of the cylinder. A repeated eight-note sequence is likely to be an ascending scale, and 'correcting' the notes on that assumption does yield a plausible-sounding result.

The recording was played at the annual meeting of the Association for Recorded Sound Collections in May 2009, but only became available on the internet in June 2014, as part of a nine-part presentation by David Giovannoni, also of First Sounds, entitled Humanity's First Recording of Its Own Voice. The cornet sequences appear in 'Act 8–The Imperfect Mirror', and sound more than plausible. Giovannoni says confidently that this phonautogram is "the earliest airborne sound reproduced to date".

In the final part, 'Act 9–Did the French Invent Everything?', Giovannoni gives due credit to the Frenchman Charles Cros (Émile-Hortensius-Charles Cros, 1842–1888), as the first person to describe a method of playing back recorded sound, having proposed photoengraving Scott's traces and using this in turn to drive a diaphragm.

The first sound recording, and the first of the human voice, that could be heard by contemporaries, was made by Thomas Edison (1847–1931) and his colleagues. Giovannoni says this team were the first to achieve the recording, storage and playback of airborne sounds. Although the chronology has frequently been presented in somewhat garbled form (often by Edison himself), the circumstances seem to have been established definitively in Randall Stross's 2007 biography of the inventor. Working late in the lab on 18 July 1877, comparing telephone diaphragms, Edison felt the vibrations as he spoke into one, and voiced his suggestion that, with a point on the diaphragm, it should be possible to make a recording while pulling something beneath it. With the assistance of John Kruesi and Charles Batchelor the experiment was rigged up there and then. Batchelor pulled a strip of wax paper through the device, while Edison spoke the standard phrase the lab used to test telephone diaphragms: "Mary had a little lamb." Playback was indistinct but audible, and by breakfast time the following morning they had achieved clear articulation from the waxed paper. The lab notebook for 18 July includes the brief entry:


Just tried experiment with a diaphragm having an embossing point & held against parafin paper moving rapidly the spkg vibrations are indented nicely & theres no doubt that I shall be able to store up & reproduce automatically at any future time the human voice perfectly

The entry written by Edison is also signed by Batchelor and James Adams, another assistant. It seems clear that Edison himself didn't immediately see the significance of what he had achieved, and it wasn't till 5 November 1877 that he set down on paper his conceptualisation of the tinfoil cylinder phonograph. The first actual working model was built by John Kruesi on 4 December 1877; the sentence "How do you get that?" is said to have been recorded and reproduced by that date.

The Smithsonian Institution's National Museum of American History holds three tinfoil recordings made by Thomas Edison. One is undated, but the other two date from 18 April 1878, when Edison demonstrated his recording machine to the National Academy of Sciences at a meeting held in the Smithsonian Institution Building, also known as the Smithsonian Castle. The exact content of the recordings isn't known, but a report in the following day's Washington Star gives some idea. Edison had his assistant, Charles Batchelor, demonstrate the machine: "Into the mouthpiece of the instrument Mr. Edison's assistant sung and shouted and whistled and crowed, and in a few minutes the same sounds floated out upon the air faint but distinct." Then,


Just as the assistant was remarking in a bass voice to the phonograph on the subject of Uncle Ned that "he had gone wh—," the tin foil gave out and the line remained unfinished. One of the audience, not a member of the Academy, however, suggested that the line be completed by adding the words "where tin foil fuses."

[Stross; Washington Star; Morton 2000; email communications from the Smithsonian Institution, April 2011]

The Smithsonian recordings could possibly be reconstructed in the same way as was recently successfully undertaken with a tinfoil recording, made on 22 June 1878 in St Louis, Missouri, held by the Schenectady Museum, using non-contact scanning, to preserve the encoded sound and reproduce it in a digital format. Having made a three-dimensional scan of the entire surface of the tinfoil, software translated that virtual surface back into sound. The 78-second recording starts with a 23-second cornet solo, and is followed by a man—believed to be Thomas Mason, a St Louis political columnist—reciting 'Mary Had a Little Lamb' and 'Old Mother Hubbard'; an unidentified voice recites 'Old Mother Hubbard'. It was played in public on 25 October 2012 for the first time since it was created. [presentation, Lawrence Berkeley National Laboratory, Times Union, FoxNews, Huffington Post]

The earliest surviving sound recording, reproducible both at the time and currently, may be the lead cylinder recording for an experimental talking clock, 1878, made by Frank Lambert (1851–1937), but the evidence for its early date remains contentious. [Wichert, Cramer and Koenigsberg] The device and its recording are currently on display at the National Watch and Clock Museum in Columbia, Pennsylvania.

[Claude Servais Mathias Pouillet (1791–1868) described, as early as 1850, a method for determining frequencies corresponding to the notes of the musical scale, involving recording the vibrations of a tuning fork directly. He published a sample wave form, which has been reconstructed by Patrick Feaster, and may be heard at First Sounds; however not only is no frequency stated for the waveform, leading Feaster to resample at eight possible frequencies, but there is no certainty that the sample waveform was recorded by Pouillet's own method: it could easily be a simple graphic.]



First disc recording

The German-born inventor Emile Berliner (1851–1929) began experimenting with sound recording in 1886, building on the work of his predecessors Édouard-Léon Scott de Martinville and Charles Cros. His debt to them was explicit—in his 1888 address to the Franklin Institute he not only uses Scott de Martinville's term 'phonautograph' for his recordings, but states that "to Mr. Charles Cros belongs the honor of having first suggested the idea of, and feasible plan for, mechanically reproducing speech once uttered."

Berliner's technique differed from Edison's not only in using a disc rather than a cylinder, but also in employing a lateral zigzag system, rather than the vertical 'hill and dale' system. In the system he patented (US Patent 371,786, applied for 4 May 1887, granted 8 November 1887) he initially recorded onto a glass disc coated in lampblack, then fixed this with varnish and had it photoengraved in metal. During the winter of 1887–8 he improved upon this with a chemical etching process direct to a zinc disc, which was more amenable to duplication.

By March 1888 he was making test records using the new process, and on 16 May that year he made a presentation to the Franklin Institute in Philadelphia, later published as 'The gramophone: etching the human voice'. As part of the presentation he played a number of recordings he had made in Washington within the previous two weeks. These were (from Berliner's Ms notes, so probably with some transcription errors by me):

  1. Yankee doodle, Baby mine, Nancy Lee &c., O du lieber Augustin

  2. Cornet—G. Samuels [?]

  3. Th Mauro—Tars Farewell

  4. Mrs Tone (Home Sweet home, Laurrie

  5. A wandring minstrel I, &c.

  6. Pity O Saviour, Decl. of Indep., counting.

The disc gramophone was first marketed, not in the USA, but in Germany. In 1889 a firm of toymakers in Waltershausen—Kämmerer & Reinhardt—obtained a licence from Berliner to manufacture miniature hand-cranked gramophones as novelty gifts. The associated 5 in discs, initially made from celluloid, later from hard rubber, were not of good quality. The gramophones were retailed all over Europe, and were sold in England for 2 guineas, including six records. Most of the selections were in German, though other languages were also recorded. The big sellers in England were The Lord's Prayer and Twinkle, Twinkle, Little Star. The voice recorded in both these last was probably that of Berliner himself. Copies of both recordings, dated to 1890, are held in the British Library's National Sound Archive. [See also Gelatt]

In February 2012 Patrick Feaster successfully recovered a Berliner gramophone recording of the inventor reciting Friedrich Schiller's 1797 ballad 'Der Handschuh' ('The Glove'), from a photograph of the disc published in February 1890 in Über Land und Meer, a German illustrated magazine. Feaster believes the recording to have been made on 11 November 1889, and as such it is the earliest gramophone disc recording that may be heard today. The audio, and a detailed account, is at Media Preservation.

Earlier experimental disc-based sound recordings had been made at the Volta Laboratory in Washington, DC, by Alexander Graham Bell (1847–1922), his cousin Chichester Bell (1848–1924), and Charles Sumner Tainter (1854–1940), roughly between 1880 and 1885. Recordings were sealed into tin boxes and deposited with the Smithsonian, where they remained, sealed and unplayed, until 1938. Six of these early recordings have now been recovered, and were played publicly for the first time on 13 December 2011; all may now be heard on YouTube. The earliest was deposited in October 1881. Inevitably these recordings were superseded by Berliner's process. [Smithsonian Newsdesk, Sound Reproduction R & D, Feaster, NMAH, Hear My Voice]


First electrical sound recording

Although electrical sound recording was to prove markedly superior to acoustic, this was not evident from the first electrical recording to be sold to the public. This was made at the burial service of the Unknown Warrior at Westminster Abbey on 11 November 1920, by Major the Hon. Lionel Guest and Captain H.O. Merriman, a Canadian electrical engineer. Just two records were made, one being 'Abide with Me', featuring choir, congregation, and the band of the Grenadier Guards. At least 1000 copies of the records were pressed by the Columbia company, for sale @ 7s. 6d. the pair, in aid of the Abbey Restoration Fund. In the absence of better microphones, telephone mouthpieces were used, and results were disappointing. This story was covered in the Illustrated London News 157.4261: 1030, 18 December 1920), which has detailed illustrations of the recording set-up.

In 2009 Pristine Audio issued a CD called Mengelberg at the Dawn of Electrical Recording [PASC 184], which includes four test sides of experimental electrical recordings made by Bell Laboratories, of a concert given by Willem Mengelberg and the New York Philharmonic Orchestra at Carnegie Hall on 2 April 1924. They were made from a cable feed intended for radio broadcast, and transferred from a private collection.

High-quality microphone-based electrical test recordings are known to have been made at Western Electric by Joseph P. Maxfield (1887–1977) and Henry C. Harrison (1887–1971) before Christmas Eve 1924, but it's not certain that any have survived (although Horning reports recent discoveries, as of 2013, of discs recorded with the Western Electric system from later 1924). According to Horning Maxfield had led a team at Bell Labs that had begun work on electrical recording in autumn 1919, and that a dozen others are known to have experimented with electrical recording between 1919 and 1925; among these was Sir Louis Sterling, of the Columbia Graphophone Company, in the UK; and Victor employees Albertis Hewitt and James W. Owen had been experimenting with electrical recording from 1916.

Using Western Electric's improved-fidelity recording, both Columbia and Victor began issuing electrical recordings from Spring 1925, both companies agreeing not to announce the fact until November that year, to allow dealers to liquidate stocks of acoustic recordings. Columbia's first electrical recording was made just one day before Victor's. From 25 to 27 February 1925 Art Gilham 'the Whispering Pianist' recorded seven electrical masters, six of which were released. The first electrical master was number 140125, 'You May Be Lonesome', released on Columbia 328-D. NB Horning says that Victor's first experimental electrical recording, of a piano solo by a Mr Watkins from the Bell Labs, was made on 9 February 1925 at the Victor lab in Camden, New Jersey.

However, the first recorded commercial electrical recording was not the first issued commercial electric recording, which was Victor 19626, both sides being selections from the University of Pennsylvania's Mask and Wig Club's production Joan of Arkansas, conducted by Nathaniel Shilkret. The A-side was recorded on 16 March 1925 and the B-side on 20 March 1925. The record was in the Philadelphia record shops in April. [Copeland, Butterworth, Gelatt]

Independently of either Bell or Western Electric, it now appears, however, that the electrical engineer Orlando Rivenius Marsh (1881–1938) had successfully developed a functioning, commercially viable process of electrical recording at his own Marsh Laboratories, Inc., and had released electrical recordings a full year before either Victor or Columbia, on his own Autograph label (although contemporaries reported considerable variation in quality from one recording to another). An Autograph recording of Jesse Crawford playing the Wurlitzer organ at Balaban & Katz's Chicago Theatre, recorded in the early Spring of 1924, is to be found on YouTube. But Marsh may have been recording electrically even earlier than that: a Billboard article published on 13 January 1923 includes a photograph of a record being made in the Marsh Laboratory using a sound-collecting device quite unlike any acoustic recording horn. The recording made on that occasion was of 'Jane, Dear', sung by August E. Bredemeier, accompanied by Alexander Kominsky on violin and a Mrs Phillips on piano. Patent-related complications restricted Marsh's ability to capitalise on his system. [Sutton; Wurtzler]



First full frequency range recording (ffrr)

ffrr was a spin-off from the development during the Second World War by Arthur Haddy (1906–1989), chief engineer of the Decca Record Company, of a high fidelity hydrophone capable of detecting and cataloguing individual German submarines by each one's signature engine noise; this enabled a greatly enhanced frequency range (high and low notes) to be captured on recordings. The frequency range of ffrr was 80–15,000 Hz, with a signal-to-noise ratio of 60dB.

On 8 June 1945 Decca announced that its ffrr system had been "in daily use for the past twelve months". The first recording had been of the 2nd movement of Tchaikovsky's Symphony No 5 in E Minor, Op. 64, made experimentally on 12 May 1944 at the Kingsway Hall in London by the National Symphony Orchestra, conducted by Sidney Beer; recorded as AR8430-32, it was later remade as 8489-91. The first issued recording was the same ensemble production of the full symphony, on 8 June 1944, issued as AR8486-95.

According to Roland Gelatt, "English Decca's early 'ffrr' issues were more commendable for quality of sound than quality of music . . .", but their release in May 1945 of a recording of Stravinsky's Petrushka—Ernest Ansermet conducting the London Philharmonic Orchestra—made a "dazzling impact", with The Gramophone pronouncing it "a new and very exciting page of gramophone history." [Decca label discography, Gelatt] This recording is now available in the Eloquence box set of Ansermet: The First Decca Recordings.


First wire recording

Franz Josef recording, 1900

The wire recorder was invented by the Danish engineer Valdemar Poulsen (1869–1942), who first patented his 'telegraphone' in Denmark on 1 December 1898 (no. 1260). [Patent]

By 1900 Poulsen and his collaborators were also experimenting with magnetic recording onto steel tape, and this was demonstrated at the World's Fair in Paris that year, where among others the voice of Emperor Franz Josef of Austria was recorded. This recording is preserved in the Danish Museum of Science and Technology, the earliest surviving magnetic recording. [Daniel, Mee, & Clark] The recording (or perhaps an extract) may be heard here.



First tape recording

Joseph A. O'Neill of New York received US Patent 1,653,467 on 20 December 1927 for his suggestion that sound could be recorded onto reels of paper strip coated with magnetisable metal particles, but the invention was not followed up in the United States. [Patent, Butterworth]

The Austrian-born inventor Fritz Pleumer (1881–1945) developed a process for striping powdered bronze onto cigarette paper, then realised that he could use the same process to coat paper recording tape simply by replacing the bronze with a magnetisable material. Using pulverised iron particles for the purpose, he created the first magnetic tape, and built the first tape recorder in the modern sense, receiving German patent DRP 500,900 on 31 January 1928. It used 16mm paper tape, recording 8mm in each direction; a 300m reel would make a 20 minute recording. Sound quality from his first recorder was poor, however, and suitable only for demonstration purposes.

Subsequently Pleumer contracted with the German electronics manufacturer AEG, and development continued in conjunction with a team led by Friedrich Matthias, of BASF (a subsidiary of I.G. Farben), until they were in a position to launch their Magnetophon K1 recorder and Magnetophon Type C tape at the Internationale Funkausstellung Berlin (the Berlin Radio Show) in August 1935. [Daniel, Mee, & Clark, Schoenherr, Milner] However, four earlier prototypes had been built between 1932 and 1934, and a recording of the last of these survives, recorded on 27th April 1935 in the Ludwigshaven Theater. This can be heard on YouTube.

On 19 November 1936 a recording was made of the third movement of Mozart's 39th symphony played by the London Philharmonic Orchestra conducted by Sir Thomas Beecham in Berlin, using an AEG K2 Magnetophon. The recording is claimed by BASF as 'the first tape recording in the world', and an extract can be heard on their website.



First 45 rpm single

RCA Victor began the development of a new system, involving a new record format (7 in microgroove vinylite, playing at 45 rpm), a new turntable, and a new changer mechanism, in 1939, but it wasn't marketed until after Columbia stole a march by introducing the 33⅓ rpm LP in 1948.

In introducing the first 45 rpm gramophones, RCA Victor knew the importance of having not just one 45 rpm record available, but a selection of different artists and styles from which buyers could choose. So in February 1949 they mixed a little of everything in with the very first batch of 45s shipped to record stores. They arrived in a custom envelope labelled: "This Is Your Preview of the New RCA Victor 45 R.P.M. RECORD LINE!" Inside were seven singles, each of which was made using a different colour plastic, each colour representing seven different musical styles.

The colours and the records in one set were: cerise for Blues and Rhythm ('That's All Right,' Big Boy Crudup, 50-0000); green for Country and Western ('Spanish Fandango,' Spade Cooley, 48-0027); sky blue for International ('A Klein Melamedl,' Saul Meisels, 51-0000); midnight blue for Popular Classics ('The French Marching Song,' Al Goodman and His Orchestra, 52-0006); and black for Popular ('Because,' Dick Leibert, 47-2857). The remaining two in the series were: red for Red Seal Classical and yellow for Children's Entertainment. Artists and titles for this pair of discs have yet to be established.

Other sets were also sent out. Although each had a selection from each colour group, the selections themselves varied. One had: green 48-0064 Rosalee Allen & Elton Brit; sky blue 51-0008 The Silver Bell Orch.; midnight blue 52-0024 Al Goodman & Orch.; black 47-2787 Tony Martin; yellow 47-0177 Turhan Bey with Henri Rene & Orch., and red 49-0318 The First Piano Quartet. The cerise is unidentified in this set.

As the music industry's very first 45s, the preview envelope suggested: "Use these seven records as samples between now and March 31st, 1949, and for use with the forthcoming window and counter displays." [Senior Moments, History of Rock]

Accompanying these sets was a marketing record for use in record shops, known as the Whirl-Away Demonstration Record. This extremely rare recording is now available on YouTube.

March 31st was the official launch day. That day RCA Victor released no less than 76 albums and 104 singles in the new format, including 30 albums and 65 singles in Red Seal; 25 albums and 11 singles in Pops; 3 albums and 12 singles in Western; 5 singles in International; 15 albums and 1 single in Children's; 5 singles in Race. At this stage all these were also available in 78 rpm format. [Katuna]



First LP

The prototype of the LP was the soundtrack disc used by the Vitaphone motion picture sound system, developed by Western Electric and introduced in 1926. The maximum playing time of each side of a conventional 12 in (30 cm) 78 rpm disc, slightly less than five minutes, was not sufficient. The disc had to play continuously for at least eleven minutes, long enough to accompany a full 1000 ft reel of 35 mm film projected at 24 frames per second. The disc diameter was increased to 16 in (40 cm) and the speed was reduced to 33⅓ revolutions per minute. Unlike their smaller LP descendants, they were made with the same large 'standard groove' used by 78s. The groove started at the inside of the recorded area and spiralled outward. Like 78s, early soundtrack discs were pressed in an abrasive shellac compound, and played with a single-use steel needle held in a massive electromagnetic pickup with a tracking weight of five ounces.

From 1928 syndicated radio programming was distributed on 78 rpm discs, but the desirability of a longer continuous playing time soon led to the adoption of the Vitaphone soundtrack disc format for this purpose. 16 in 33⅓ rpm discs playing about 15 minutes per side were used for most of these 'electrical transcriptions' beginning in about 1930. Initially, transcription discs were pressed only in shellac, but by 1932 pressings in RCA Victor's vinyl-based 'Victrolac' were appearing, which provided a much quieter playing surface.

RCA Victor introduced an early version of a long-playing record for home use in September 1931. These 'Program Transcription' discs, as Victor called them, played at 33⅓ rpm and used a somewhat finer and more closely spaced groove than typical 78s. They were to be played with a special 'Chromium Orange' chrome-plated steel needle. The 10 in discs, mostly used for popular and light classical music, were normally pressed in shellac, but the 12 in discs, mostly used for 'serious' classical music, were normally pressed in Victor's vinyl-based Victrolac compound. They could hold up to 15 minutes per side. Beethoven's Fifth Symphony, performed by the Philadelphia Symphony Orchestra under Leopold Stokowski, was the first 12 in recording issued. These Victor discs were not commercially successful, and development of the LP went on the back burner for the ensuing decade. One example, recorded on 30 July 1931, is this selection played by the Victor Symphony Orchestra.

CBS Laboratories head research scientist Peter Carl Goldmark (1906–1977) led Columbia's team to develop a gramophone record that would hold at least 20 minutes per side. Research began in 1941, was suspended during World War II, and then resumed in 1945. The first vinyl LP came off the press at the CBS laboratory on 27 February 1946. A photograph of one of the first three experimental LPs to be pressed appears at Jac Holzman's Dreams of Vinyl. On the assumption that all three were of the same recording, the first ever experimental pressing of a modern LP was of Rimsky-Korsakov's 'Scheherazade' Part 2, movements 3 & 4.

It took two further years for the process to be perfected. The LP as introduced allowed for 23 minutes per side, being recorded on nonbreakable vinylite microgroove (224–300 grooves per inch). Columbia Records unveiled the new LP at a press conference in the Waldorf Astoria on 18 June 1948 in two formats: 10 in (25 cm) and 12 in (30 cm) in diameter. A hundred recordings were released simultaneously to allow for a purchasing catalogue; the first catalogue number for a 10 in LP, CL 6001, was a reissue of Frank Sinatra's 78 rpm album set The Voice of Frank Sinatra; the first catalogue number for a 12 in LP, ML 4001, was the Mendelssohn Concerto in E Minor for Violin and Orchestra, Op. 64, played by Nathan Milstein with the Philharmonic-Symphony Orchestra of New York conducted by Bruno Walter. This may be heard here.

[See also Butterworth, Gelatt, and Morton]


First audio cassette

In 1962 Philips invented the compact audio cassette medium for audio storage, introducing it in Europe in August 1963 (at the Berlin Funkausstellung). The original player, the Philips EL3300, was designed for consumer recording onto blank tapes, and had no tab for the detection of pre-recorded tapes. [mkkiani on YouTube] The mass production of compact audio cassettes began in 1964 in Hanover, West Germany.

Although there were other magnetic tape cartridge systems, the Compact Cassette became dominant as a result of Philips's decision in the face of pressure from Sony to license the format free of charge. Philips also released the Norelco Carry-Corder 150 recorder/player in the US in November 1964. The Carrycorder 150 came with a demonstration tape—Philips EL 1903d/15 K—which at the time of writing would appear to be the earliest-known pre-recorded audio cassette (referred to then, however, as a 'cartridge').

Pre-recorded music cassettes (also known as Musicassettes) were launched in Europe in late 1965, though only 9000 were sold in the first year. [History of Recorded Music]


First audio CD

Both Philips and Sony worked on developing optical digital audio discs from the mid 1970s. Sony publicly demonstrated prototypes in 1976 and 1978, and Philips in 1979. Later in 1979 the two companies set up a joint task force of engineers, to agree on a common standard. The so-called 'Red Book' standard was published the following year, and was later (1987) adopted by the International Electrotechnical Commission (IEC) as an international standard.

The first test CD using the new standard was pressed in Langenhagen near Hannover, Germany, by the Polydor Pressing Operations plant. The disc contained a Deutsche Grammophon recording of Richard Strauss's Eine Alpensinfonie, played by the Berlin Philharmonic and conducted by Herbert von Karajan. The first public demonstration was on the BBC television program Tomorrow's World when the Bee Gees' album Living Eyes (1981) was played; the clip's to be found here. The first CD to be manufactured at the new factory was The Visitors (1981) by ABBA. The first album to be released on CD was Billy Joel's 52nd Street, reaching the market alongside Sony's CD player CDP-101 on 1 October 1982 in Japan. [The CD Turns 30]


First digital sound recording

Digital sound recording is based on Pulse Code Modulation, which was invented in Bell Labs in the 1930s. Researchers at Japan's NHK broadcast network developed a monophonic PCM audio recorder in 1967, and a 2-channel stereo version by 1969, using an industrial helical-scanning videotape recorder as its storage medium. Denon—part of Nippon Columbia—leased an NHK stereo PCM recorder, and in 11 January 1971 recorded The World of Stomu Yamash'ta 1 & 2 by Stomu Yamash'ta, the first commercial digital recording; (one track can be heard here). The first commercially-released digital recording was the LP Something, by Steve Marcus, Jiro Inagaki & Soul Media (Nippon Columbia NCB-7003), recorded in Tokyo two weeks later, and released in February 1972. [Fine; Wikipedia]

The first MP3 was made by Karlheinz Brandenburg (b. 1954) and his collaborators, who were the principal developers of the audio compression format. It's often stated that the song he chose for his experimentation was the original a capella version of Susanne Vega's 'Tom's Diner', which she had recorded on 1 April 1987. [Ewing] But this is an oversimplification: Sterne lists ten distinct audio samples used for MPEG high quality listening tests in July 1990, among which a 20 second sample from 'Tom's Diner' was included, but which also included 21 s of Tracy Chapman's 'Mountains o' Things', 16 s of glockenspiel, 20 s of a Haydn trumpet concerto, Ornette Coleman's 'In All Languages', 30 s of fireworks, 20 s of castanets, 19 s of male speech, Kofi Bentsi-Enchill on bass guitar, and a piece on bass synth; the actual glockenspiel (track 35) and castanet (track 27) test samples used are online as FLAC files; Sterne gives full details of the recordings. Strictly speaking it's anachronistic to refer to the tests as MP3s, as during testing the file extension was .bit, and the switch to .mp3 was first announced internally to the company that developed it (Fraunhofer) on 14 July 1995.


First recording of a radio broadcast

The earliest surviving recordings of a radio signal are segments of Morse code transmissions recorded off the air (from the Sayville, Long Island, Telefunken station) in late 1913 or 1914 by Charles Apgar, a New Jersey radio amateur. By fitting the electrical element of a headphone to a home-made electrical recording head attached to an ordinary Edison phonograph Apgar was able to record radio signals picked up by his receiver on wax cylinders, and he made a number of such transcriptions during 1913–1915.

Apgar's original wax cylinders are lost, but samples of his recordings survive, courtesy of an uncoated aluminium aircheck of Apgar's appearance on station WJZ in New York on 27 December 1934. Apgar was interviewed by NBC announcer George Hicks, and highlighted his description of his experiments by playing two of his cylinders into the microphone. 12 in aluminium copy discs of this program are held by the Antique Wireless Association, and a tape copy by the Library of Congress.

NB Technically these recordings are of point-to-point transmissions, not broadcasts. [Documenting Early Radio, Oldest Surviving Airchecks]

There is documentary evidence for numerous recordings of broadcasts made by technicians working for the Brunswick-Balke-Collender Company and the Victor Talking Machine Company in 1921–1923, but none are known to survive. [Documenting Early Radio] It's not 100% clear which is the earliest surviving recording of actual broadcasts, but it must be among the following:

  • The RadioGOLDINdex lists the earliest surviving recording as being one of 4 min 38 s of US President Warren Harding from 24 May 1922, in which the president speaks from Hoboken, New Jersey, at what sounds like a Memorial Day Ceremony. He says, "It must not happen again." The audio condition is said to be good to very good, and the recording apparently complete.

  • The RadioGOLDINdex lists two recordings, presumably from the BBC, from 24 May 1923, both recorded and released by HMV. The first is 3 min 20 s of King George V and Queen Mary, and is the 'Empire Day Message to Children'; the King tells the youth of the world to "Always play the game." The second is of the Band of the Coldstream Guards playing 'God Save the King' and 'Home Sweet Home'. It seems likely that these were recorded off-air, however.

  • Several excerpts from 1923–24 New York Philharmonic Symphony Orchestra broadcasts over station WEAF were recorded as experiments by Bell Laboratories, and numerous examples survive. These are brief segments, and not complete programs. Some were released on CD by the Philharmonic in a collection entitled Historic Broadcasts: 1923–1987, with a five minute segment from a December 1923 broadcast being the earliest included. There are no announcements on any of these music recordings. [Documenting Early Radio] Some of these surviving excerpts apparently date from several months before November 1923. [Oldest Surviving Airchecks]

  • The 10 November 1923 recording of the Armistice Day Speech by US President Woodrow Wilson, broadcast on WEAF New York, WCAP Washington, and WJAR Providence. The 7 min 56 s recording was made by Frank L. Capps, recording engineer. The process used is believed to have been electrical, but this has not been conclusively determined. The existing vinyl pressing of the recording was made by the Compo Company of Lachine, Quebec, around 1940, and is held by the Franklin D. Roosevelt Library. Both sides of the disc contain the same material, with slight variations. Audio quality isn't good, partly owing to Wilson's ill health. His voice is weak and distant as he discusses the significance of Armistice Day, and stresses the need for international cooperation in the future. The recording includes no announcements. [Documenting Early Radio, RadioGOLDINdex] This is the recording selected for the US National Recording Registry in 2005, as the earliest surviving sound recording of a regular radio broadcast. [National Archives press release]

  • The 14 March 1925 recording of a trans-Atlantic broadcast from 2LO London via 5 Daventry over WJZ. The first relay of an overseas signal survives in a series of test pressings of undocumented origin, formerly owned by Dr Albert Goldsmith of RCA, and now held by the University of Maryland's Library of American Broadcasting. Ten sides were recorded at 78rpm, most likely by placing a radio horn speaker next to a microphone, which would explain the hollow, metallic tone of the recording. Much of the 37-minute recording is unintelligible, due in part to the poor recording quality, but also due to the poor quality of the shortwave reception. There are frequent crashes of static punctuating a fairly constant roar of atmospheric noise. However, there are short passages of recognizable dance music from London, with 'Alabamy Bound' one of the selections heard most distinctly, along with brief phrases from the BBC announcer. More clearly, announcer Milton Cross of WJZ can be recognized toward the end of the sequence, breaking in to explain what is happening, and to deliver a station identification. Newspaper accounts from the broadcast's relay point in Belfast, Maine, indicate that the BBC material was heard very clearly by listeners there, who picked up the signal directly from the RCA relay station. This would indicate that perhaps much of the interference was encountered on WJZ's end of the relay circuit. Although this is said to be a difficult recording to understand, it's nonetheless "an invaluable window into the past. It preserves, as perhaps no other early recording does, the sound of a broadcast as it actually sounded to a listener in 1925. For that reason it stands as a true historical treasure." [Documenting Early Radio, Oldest Surviving Airchecks]


First recording of a telephone call

Poulsen's telegraphone could be used for the recording of phone calls, but it's not clear whether any such recordings survive. Louis Blattner (1881–1935) produced an answering machine—a version of his 'Blattnerphone'—at Elstree in 1929 [New York Times, Armour Tech News]. Again, it doesn't appear that any recordings survive.

The earliest surviving recorded phone call, however, appears to be the 1931 Pathé news film Longest Telephone call in the World, in which the conversation between the Lord Mayors of London and Sydney is heard as it happens, with film footage of both ends of the conversation.

YouTube has a totally mundane wire-recording of a call to the Time Bureau (a live 'speaking clock'), from 23 April 1950. This seems to be the earliest surviving wire-recorded phone call so far located.


First reproducing piano

In this context it's important to distinguish the reproducing piano from its predecessors, the pianola and the player piano. The distinguishing feature is that reproducing pianos are able to reproduce the rubato, dynamics and pedalling—in other words, the complete performance—of the pianists who recorded for them.

This is clearly a very different kind of recording from any other described on this website, since what is recorded mechanically is reproduced as the actual sound exactly as it was recorded, played directly on the piano itself. That's to say, while the recording itself is not of the sound, but of the playing performance of the pianist, it's the sound itself that is reproduced.

The reproducing piano was first produced by Edwin Welte and Karl Bokisch, of the firm of Michael Welte und Söhne in Freiburg-im-Breisgau, Germany. It was patented in Germany as Patent 162.708 (applied for 21 May 1904, granted 25 September 1905, and subsequently in the UK, Austria, France, and the USA), as for an 'Apparatus for Graduating the Striking of the Keys in Mechanical Piano Playing Apparatus'. The reproducing piano as manufactured is universally known as the Welte-Mignon. [Welte-Mignon – The World of Mechanical Music (in German, with no subtitles)]

In the recording process, all aspects of the performance were recorded automatically as ink or carbon traces on a paper roll. The piano roll for reproduction was then punched out by hand exactly as indicated on the graphic recording.

The earliest precisely dated Welte-Mignon roll recording was made on 13 January 1905, by Ferrucio Benvenuto Busoni, of J.S. Bach's Choral Prelude 'Nun freut euch liebe Christen' (Roll number 0439). Roll 0001, dated only to 1905, was Eugenie Adam-Benard's recording of Beethoven's Grande Sonate (pathétique) No. 8, Op. 13 c-Moll, I. Satz Grave—Allegro di molto e con brio.

The earliest born pianist ever to record on music roll was Carl Reinecke (1824–1910). Born three years before the death of Beethoven, Reinecke was a friend of Schumann, was helped in his career by Mendelssohn, knew Brahms as a young man, and probably heard Chopin play in Paris. The Pianola Institute has an MP3 of a Reinecke recording from 21 January 1905 (as well as a recording by Manuel de Falla from 1912); these examples show the exceptional quality of the Welte-Mignon. [Welte-Mignon, Pianola Institute]

There is an interesting example of a very early precursor, in the form of the engraved musical program of a 'Romance' by Claude Balbastre, as rendered by Père Marie Dominique Joseph Engramelle for Dom François Bédos de Celle's 1778 treatise L'art du facteur d'orgues [The Art of the Organ-Builder]. The program reflects actual live performance by Balbastre, in the presence of Engramelle, who monitored the timing of the recording precisely to the second. Engramelle's work is a unique document in the history of 18th century performance practice, and the only piece for a mechanical instrument both set up under the direction of the composer and specifically designed to exhibit its ability to reproduce the ornamentation, articulation, and rhythmic nuances of an ideal live performance. The engravings as published are reproduced in Feaster, and Feaster's audio rendition is included on the accompanying CD.

As a variation on the reproducing piano, companies including Welte and Hupfeld also manufactured an external reproducing mechanism described as a Vorsetzer, in which the recorded rolls drove the playing of the keys of the piano in front of which it was set up. [Reblitz: 416; a 1924 Welte example can be seen in action here.]

Welte was not the only company that subsequently produced reproducing pianos and piano rolls: others included Aeolian and Wurlitzer. There were also reproducing organs, and according to Arthur A. Reblitz "The rolls capture the hand playing of many fine organists of the day, with a high level of musical sophistication equivalent to that of reproducing piano rolls." [Reblitz: 274]

The present-day equivalent of the Welte-Mignon reproducing piano is the DisklavierTM, manufactured by the Yamaha Corporation of Japan, of which the earliest version, in 1982, was an upright reproducing instrument called 'Piano Player' [Yamaha]. It featured a record-and-playback system, floppy disk storage of performance data, and the ability to playback multi-track performance files that included instrumental tracks, the sound of which was reproduced by a tone generator. From 1992 the next generation of Disklaviers included support for General MIDI. The typical Disklavier is a real acoustic piano fitted out with electronic sensors for recording and electromechanical solenoids for playback. Sensors record the movements of the keys, hammers, and pedals during a performance, and the system saves the performance data as a Standard MIDI File. On playback, the solenoids move the keys and pedals, exactly reproducing the original performance. The current seventh-generation model, the Disklavier ENSPIRE, introduced in January 2016, can record direct to USB, includes internet streaming, and has V-sync technology, which allows users to create video recordings that synch to recorded piano performances. The starting price (SMRP) for the ENSPIRE is in excess of US$80,000, with a range up to US$145,000 for the DS7X ENPRO, and concert grand pianos beyond that, unpriced. [ENSPIRE]



Full references for printed works

W.E. Butterworth (1977) Hi-Fi. From Edison's Phonograph to Quadraphonic Sound

Peter Copeland (1991) Sound Recordings. London: The British Library

Patrick Feaster (2012) Pictures of Sound. One Thousand Years of Educed Audio: 980–1980. Atlanta, Georgia: Dust to Digital. Book with included audio CD.

Harry M. Geduld (1975) The Birth of the Talkies. From Edison to Jolson. Bloomington & London: Indiana University Press

Roland Gelatt (1977) The Fabulous Phonograph 1877–1977. 2nd rev. edn. London: Cassell

Susan Schmidt Horning (2013) Chasing Sound. Technology, Culture & the Art of Studio Recording from Edison to the LP. Baltimore: Johns Hopkins UP

Peter Martland (2013) Recording History: The British Record Industry, 1888–1931. London: Scarecrow Press

Greg Milner (2009) Perfecting Sound Forever. The Story of Recorded Music. London: Granta

David L. Morton jr (2000) Off the Record. The Technology and Culture of Sound Recording in America

David L. Morton jr (2004) Sound Recording. The Life Story of a Technology

Arthur A. Reblitz (2001) The Golden Age of Automatic Musical Instruments: Remarkable Music Machines and Their Stories. Woodsville, NH: Mechanical Music Press

Jonathan Scott (2023) Into the Groove. The Story of Sound from Tin Foil to Vinyl. London: Bloomsbury Sigma

Jonathan Sterne (2012) MP3. The Meaning of a Format. Durham, NC, and London: Duke University Press

Randall Stross (2007) The Wizard of Menlo Park. How Thomas Alva Edison Invented the Modern World. New York: Three Rivers Press

Paul Tritton (1991) The Lost Voice of Queen Victoria. The Search for the First Royal Recording. London: Academy

Walter L. Welch & Leah Brodbeck Stenzel Burt (1994) From Tinfoil to Stereo. The Acoustic Years of the Recording Industry 1877–1929. Gainsville, FL: University Press of Florida


© 2009–2024 Benjamin S. Beck

Return to Technology contents page


If you know of any earlier examples, please contact me.



Return to Firsts contents page

Return to website home page


This page was last revised on 2024-07-05.