Donald's Encyclopedia of Popular Music

A B C D E F G H I J K L M N O P Q R S T U V W X Y Z

RECORDED SOUND, history of

Sound was recorded in Paris, France, in 1857 by Édouard-Léon Scott de Martinville, often known as plain Leon Scott, a typesetter who invented the phonautograph ('sound-writer'), using a paper cylinder blackened with smoke. There was then no way to play back the recording. On 27 March 2008 the New York Times reported that the Lawrence Berkeley National Laboratory in Berkeley California, using optical imaging and a virtual stylus, managed to play back one of Martindale's phonautograms, made on 9 April 1860, of someone singing the folk song 'Au Clair de la Lune'. This would be the oldest sound recording that we have. Scott went to his grave convinced that Thomas Alva Edison had stolen his thunder.

Edison (b 11 February 1847, Milan OH; d 18 October 1931, West Orange NJ) invented the phonograph in 1877: wrapping tinfoil around a cylinder, he spoke 'Mary had a little lamb' into a diaphragm which caused a stylus to cut a groove of varying depth on the tinfoil (the 'hill and dale' method of recording); a separate stylus and diaphragm was used for playback. Edison ignored his invention for a few years while he worked on electric light; he had improved the diaphragm for the telephone of Alexander Graham Bell (b 3 March 1847, Edinburgh, Scotland; d 2 August 1922, Baddeck, Nova Scotia), then Bell's assistants Charles Tainter and Chichester Bell built a machine in Bell's lab which recorded in wax, called the graphophone. It was thought the machines would be useful for dictation, nobody thinking of entertainment at first, but a graphophone company became the world's biggest record label before 1900 (see entries for Columbia and RCA).

Meanwhile, Emile Berliner (b 20 May 1851, Hannover, Germany; d 2 August 1929, Washington DC) had first exhibited his gramophone in 1888, using a laterally modulating groove spiral cut on a flat record; it immediately began to overtake Edison's cylinder. At that point Edison's phonograph put out a very weak sound, requiring users to use stethoscopic earpieces to listen; the volume from Berliner's gramophone record was much louder, and the flat record was much more practical for mass production and for storage at home. Incessant lawsuits over patents destroyed the trademark 'Gramophone' around 1900 in the USA, and Edison was a national hero there, so Americans continued to use the word 'phonograph' to describe either kind of record player, but the term 'gramophone' won out in Europe.

Much research was done into acoustic recording, which required horns of decreasing diameter for recording or increasing diameter for playback, and a stylus/diaphragm assembly (or soundbox) at each end. Some remarkable results were obtained, especially with the human voice (a Caruso record was the first million-seller; see HMV) but the recording of large ensembles and difficult instruments such as the piano (with all its overtones) was problematic. Early record speeds varied between 50 to 100 rpm., so that later reissues by famous opera singers were sometimes transferred at the wrong speed, but the need to synchronize electric motors with AC current meant 78.26 rpm in the USA, with current at 60Hz, but 77.92 rpm. in the UK, with house current at 50Hz, and there is a story that Victor in the USA stuck to recording at 76 rpm so that its records would have more 'zing' when played back, to the chagrin of its European outlets. The U.S. speed, 78.26, became standard.

Sentimental and novelty songs were huge hits: Billy Murray (b 25 May 1877, Philadelphia; d 17 August 1954) had a huge number of hits 1903-26 including 'Meet Me In St Louis, Louis' and 'Come Take A Trip In My Air-Ship' in 1904-5 snd 'Take Me Out To The Ball Game' in 1908; he also recorded definitive versions of hit songs by George M. Cohan. Ada Jones (b 1 June 1873, Lancashire, UK; d 22 May 1922) was the most popular female vocalist of the pre-1920 era; she had solo hits and hit duets with Murray ('Come, Josephine, In My Flying Machine' 1911, backed by a quartet). Vocal groups were also big, the Peerless and Haydn Quartets both having hits with 'Sweet Adeline' in 1904. Jones also did duets with Len Spencer (b 12 February 1867, Washington DC; d 15 December 1914), whose solo hits included 'All Coons Look Alike To Me' (not as racist as it sounds: see Ragtime). All these were among the top 100 best-selling artists of 1890-1954 (so says Pop Memories; but see Charts): Bing Crosby was no. 1, but Murray was no. 5, Jones 33. The first blues record was made in 1920 by Mamie Smith, the first country music in 1922-3 by old-time fiddlers, and the first jazz in 1917 by the Original Dixieland Jazz Band, but more importantly by King Oliver in 1923. Edison made flat discs with hill-and-dale grooves until 1913; some gramophones had changeable soundboxes for the two types of flat disc, but Edison's records failed: his hill-and-dale groove was better, and Berliner had probably specified a lateral groove only trying to avoid infringing Edison's patent; Edison was also the first to experiment with electrical recording and to produce harder record surfaces and diamond styli, but Berliner's gramophone record won in the marketplace, because it was easier to manufacture and took up less space in the home, making record collections possible. Edison made cylinders until 1929 for owners of his phonograph.

Meanwhile, Guglielmo Marconi (b 25 April 1874, Bologna, Italy; d 20 July 1937, Rome) patented wireless telegraphy in 1900; Lee De Forest (b 26 August 1873, Council Bluffs IA; d 30 June 1961, Hollywood) invented the triode in 1906, which could amplify a weak signal: vacuum tubes made broadcasting possible (called valves in UK, because that's what they are, allowing alternating current in one direction but not the other). When music was broadcast, musicians spoke of 'De Forest's prime evil', but he also invented the soundtrack method used in talking films, creating a lot of work for musicians. Attempts to develop electrical methods of recording using telephone and broadcasting technology began c.1915, interrupted by WWI; in 1924 the Western Electric method was successful, adopted the next year by virtually all record labels: a microphone converted the sound to an electric signal which drove the recording stylus, and the frequency range available from the process nearly tripled. Louis Armstrong and his Hot Five recorded this way, while Crosby used the microphone in a style that seemed personal to each individual listener, and modern pop music began.

Many home record players were still acoustic, but yielding better sound with electrical recordings. Early electric playback pickups (cartridges) were magnetic and very heavy, wire wrapped around an armature converting the groove's squiggle to an electrical signal; lighter ceramic pickups became standard, the piezo-electric effect causing certain materials to become electrically polarized under pressure; ceramic cartridges were used for decades in cheap record players (the principle also used in cigarette lighters). Records made of powdered slate mixed with shellac wore out quickly under heavy tracking pressures; most 'needles' were made of steel and had to be changed often. Cactus or bamboo styli were kinder to discs but gave a weaker sound and had to be sharpened after every play. Before WWII smaller, lighter magnetic cartridges were made by the Germans, who also developed tape recording. Valdemar Poulson (b 23 November 1869, Copenhagen; d 1942) made the first magnetic recordings on wire in 1898; steel tape was also used but by the mid-1930s the first magnetic tape of plastic coated with iron oxide was made in Germany: pre-war tapes of complete opera performances were later transferred to LP. The first 'binaural' recordings were made in 1933 by EMI, using separate tracks on the record, a two-headed tone-arm playing both at once; but methods of cutting stereo grooves had been patented in England in 1931, invented by EMI's Alan Blumlein. Stereo records were made in 1932 but the process was not then commercially viable because of the noisy materials of which records were made. Meanwhile some record companies had cut two 78 masters of each recording in case one went wrong, and HMV and Victor for some reason used separate microphones as early as 1929; in 1984 collectors discovered that two versions of the same performance when combined yielded stereo: the album Stereo Reflections In Ellington included stereo medleys by the Duke Ellington band from 1932. Edison had made cylinders with the grooves closer together (more easily done with hill-and-dale grooves); then the Ellington medleys were examples of a Victor long-playing record which failed because it was essentially a 78 slowed down to 33 1/3 rpm: the doubled playing time did not compensate for higher surface noise, and during the Depression few people were buying new record players anyway.

After WWII CBS Labs led by Dr Peter Goldmark perfected the long-playing microgroove disc made of PVC (polyvinylchloride) (or anyway Goldmark got the credit, but see his entry). RCA introduced the microgroove 45 to compete, and the 'battle of the speeds' began; 78s became obsolete (though Beatle 78s were made in India as late as 1966). Wartime research in electronics meant great advances in sound quality; record companies switched to tape for master recordings, also making editing of recordings possible for the first time. It had long been evident that there was better sound on the record than home record players were capable of reproducing; records sounded better over the radio because broadcasters could afford studio-quality playback gear. After WWII veterans trained in electronics began to build their own gear, and 'high fidelity' was born. Good sound was measured three ways: dynamic range (soft to loud), frequency range (about 20 to 20,000 cycles per second defined as hi-fi) and definition (ability to distinguish among sounds). But there were no guidelines in consumer goods, and a lot of department-store junk was labelled as hi-fi; the big American manufacturers refused to upgrade their wares and were eventually put out of business by the Japanese, while the British had provided better quality in a smaller market. Modern magnetic cartridges were essential to good sound, but the signal, though higher in quality, was weaker than that from a ceramic pickup, requiring a pre-amp stage in the playback system.

Most record players had not included much in the way of wide-range tone controls because there was no point; solid bass had been available only from speakers in very large tuned cabinets for which most people simply didn't have room. Then the American Edgar Villchur invented the acoustic suspension loudspeaker in the early 1950s, the speakers sealed in an airtight cabinet, the cushion of air behind them making low-frequency reproduction possible in a 'bookshelf' speaker system. Recording characteristics were standardized in that decade: low frequencies are attenuated in manufacturing gramophone records because the grooves cannot store them, and are restored in playback, called compensation or equalization; each record company did this differently until the RIAA (Record Industry Association of America) standard made things easier for designers of quality amplifiers. But hi-fi remained an expensive luxury until the solid-state era and the rise of the Japanese electronics industry provided quality record-playing equipment that most people could afford (once better sound became common, however, the British and the Americans still made better loudspeakers than anybody else).

Crystals had converted alternating current to direct current in early radios, but De Forest's vacuum tube was better at the time; then it was discovered that germanium crystals containing certain impurities were better still: transistors were invented at Bell Labs in 1948, lasting longer than tubes and not generating as much heat; they won a Nobel Prize for their inventors in 1956 and by then all-transistor radios were available in the USA and transistors were being rapidly improved. Pre-recorded four-track stereo reel-to-reel tapes were issued commercially by EMI in the mid-'50s; attempts to develop stereo records included combining lateral and hill-and-dale motion in the same groove, but this would have meant mono records could not be played by the new pickups. The 45/45 system patented decades earlier in England was adopted, engraving separate stereo information on each wall of a standard groove, and stereo records were introduced in 1958, successful even though they required a new cartridge, the equivalent of two amplifiers and another speaker. Mono recording had reached a very high standard; engineers now had to learn how to do it all over again for stereo: it was necessary to record three tracks and mix them to two channels in order to avoid a 'hole-in-the-middle' effect. The public at large did not understand that a good mono recording was better than a poor stereo one; the Count Basie LP The Atomic Mr Basie, made in October 1957 in NYC, was an exceptionally fine mono record, but the wretched experimental stereo edition was the only one available for many years. Ersatz 'electronic' stereo reissues of older records became common; a stereo illusion could be created by spreading the frequency range across the stereo channels, but most record companies foolishly did it by adding distortion in the form of echo and out-of-phase elements.

['Stereo' recordings had been inadvertently made with two different takes of the same performance using different microphones, which could later be combined to make stereo. As we have seen, this happened to Duke Ellington in 1932, but again at Newport in 1956 (see Ellington's entry), and to Toscanini's 1952 broadcast of Verdi's Requiem, available from Pristine Audio in France.]

The Philips four-track tape cassette was introduced in 1963; the Lear eight-track continuous-loop cartridge '65 failed. Recording technology was improved by research into lower signal-to-noise ratios for cassettes, a milestone being the Dolby process; first used on master tapes, then in cassette decks, this was a new kind of compensation, raising the volume of quiet high-frequency passages during recording, then lowering it during playback, reducing tape hiss. Sales of cassettes passed those of LPs in 1983 in the USA, in '87 in the UK. Quadrophonic sound was introduced in the early 1970s, the idea being that speakers placed behind the listener would reproduce the ambience of the room in which the recording was made; it failed dismally. By the '70s recording technology was over-used, with too many microphones and too much mixing of master tapes often resulting in grossly unrealistic recordings in both classical and popular genres, but the records themselves were improved by DMM (direct metal mastering), eliminating steps in the manufacturing process and making possible well over 30 minutes on an LP side.

Everything changed again when digital recording began and the compact disc was introduced by Sony and Philips in 1982, rendering Dolby and other mastering processes redundant. For various reasons, the analogue recording process includes a certain minimum of distortion; in digital recording the signal is broken into billions of tiny bits which represent only the music: playback by means of a laser beam bypasses most sources of distortion and mercilessly reveals gimmickry in recording technique.

Reissue of vintage material in the past was often badly done, but standards had risen; RCA's LPV reissue series of vintage pop and jazz in the '60s was a model of its kind, and John R. T. Davies in the UK had overseen superb reissues including classic jazz, lately on the Hep label, dubbing clean copies of original 78s on to master tape. Ward Marston, working in his home studio in Philadelphia, performed miracles with vintage classical 78s, and there were other engineers who became well known. Digital reissues of analogue masters could be spectacularly fine, the new technology allowing extremely precise filtering; the revelation of direct-cut 78s can be startlingly good. English-born Australian engineer Robert Parker was one of the first in the field, issuing compilations of early jazz on CD, dubbing original 78s directly on to digital tape; much of his work was astonishing ('Someday Sweetheart' '26 by King Oliver; 'Stompy Jones' and 'Live And Love Tonight' '34 by Ellington; 'San' by Paul Whiteman) but he added ersatz stereo and phony ambience, often resulting in too much echo. Dennis Ferrante and Edward Rich remastered Glenn Miller's 'American Patrol' for Woody Allen's Radio Days soundtrack '87 (on RCA's Novus label), and the rhythm guitar, the reed section's baritone sax, etc could be clearly heard for the first time. But RCA's own initial transfers of vintage Ellington, Miller etc were botched; they learned to go back to the original metal parts, make a new 78 and dub that directly on to the digital master tape, so that Miller's 'Tuxedo Junction' not only sounded good but was available for the first time since the 78 era without the 'wow' dubbed many years earlier from an off-centre 78. (Bluebird reissues were generally very good, but the 1940-46 Ellington material had to be done a third time for Ellington's centenary.)

As for new recordings, the public thought it was buying better sound on CDs, but the CD as it was introduced was not an improvement on the LP: it should have been designed with a higher sampling rate to capture more ambience (the 'space' around the sound, determined by the room in which the recording was made) and engineers had to learn how to record all over again, as with stereo. CDs like Salif Keïta's beautiful Soro (made c.1987 on a 48-track digital machine in Paris) sounded ice-cold, as if they were made on the moon, and were improved by dubbing them on to a cheap cassette, introducing distortion to compensate for the complete lack of atmosphere in the recording. There were also bad transfers of analogue tape recordings: Sony developed a '32-bit mapping system' to compensate for the shortcomings of the CD and made very good transfers of classical reissues to sell at mid-price, but did shoddy work in jazz and pop; when they got around to decent transfers of Miles Davis's Kind Of Blue and Bob Dylan's Blonde On Blonde, they issued them on gold-coloured CDs and charged a premium price for the quality they should have offered in the first place. EMI made three CD transfers of Rafael Frühbeck de Burgos's marvellous recording of Orff's Carmina Burana, from c.1966; the first was dreadful and the second not much better: a decent one finally came along in the 21st century. But standards in general have improved greatly. Lester Koenig's classic Sonny Rollins album Way Out West on Contemporary, made in 1957, was on CD as Way Out West Plus: 70 minutes of prime Rollins sounding better than ever. Recordings for Blue Note and Prestige made since the '50s by Rudy Van Gelder have come up fresh as paint on CD (especially the RVG series on Blue Note) partly because they were made without gimmicks in the first place.

In the mid-'90s some CDs with as much as 80 minutes' playing time caused tracking problems on older players. The industry had been suspicious of CD at first, remembering the quad fiasco; the first CD factory outside Japan (Sony) and Hamburg (Philips, on the site of the first European gramophone-record factory) was built in the UK by Nimbus, an independent classical company tired of quality-control problems with LPs; they improved the laser mastering process, cut pressing time by more than half and won a Queen's Award for Technology; they stopped making LPs '85, and ten years later even England's Linn, who despised the CD, admitted that at its best it could rival the best analogue sound.

The new market was difficult for small independents who could not afford to issue CDs, but the new medium was a spectacular success: the world-wide production shortfall ended and production costs came down; retail prices did not come down enough, but CDs are small and cute, perfect impulse purchases, and collectors could sneak them into the house so that parents or spouses didn't notice how many they were buying. As with the transfer from 78s to LPs, the savings in space make possible larger collections at home. USA record companies shipped 8m CDs '83, 53m '86, 984m '91 and 1,956m '95: by '95 virtually no vinyl records were being pressed except by specialists. New products included DAT (digital audio tape) '87, which failed in the larger marketplace but became a useful professional tool; the CD had been a shot in the arm for an ailing record industry, and in the early '90s the multinationals tried to do it again with CDC (compatible digital cassettes) and the MD (mini-disc, a small recordable computer disc); these were predictable failures. (But cassette sales fell for the first time '93: from 1,524m '91 they had slipped to 1,363m '95 and were only being kept up by huge numbers in undeveloped countries; the mini-disc was doing well in Japan.) CD-ROM storage facilities for information and entertainment played back on computers and home entertainment centres were supposed to be the next big growth area, but suffered competition from the Internet; improved methods of compressing information, recordable CDs and the ability to download music from the Internet was causing new problems: the industry was paranoid about copyright control and seemed to have no idea what it was up against.

Thanks to the success of the CD, record catalogues were bigger than ever, more recordings available by Duke Ellington or Arturo Toscanini than there ever were in their lifetimes, but this meant that record shops began to become obsolete, partly because only the biggest ones could do justice to the selection. Hundreds of record stores a year were closing each year in the USA as music fans bought more and more CDs through the mail. Early in the 21st century it was possible to buy recordings downline at home, direct from a retailer's computer; but kids were downloading, in many cases illegally, off the Internet from each other, using the compressed MP3 format, which loses a great deal of sound quality, and the march of better sound was temporarily going backwards.

Meanwhile all this technology had an unfortunate effect on popular music. There was already too much gimmickry in recording when during the 1980s anything became possible; 'songs' were written and programmed using computers. In the '90s Glenn Gregory of Heaven 17 said, 'Technology did suddenly leap ahead and the only parameters you had were in your imagination. Unfortunately, you were likely to disappear up your own arse looking for your imagination.' Synth wizard Jean-Michel Jarre went back to analogue machines, saying that 'When you are listening to the music of the '80s today, it is very compact, very efficient and really dated.' Keith Richards had already observed, 'What's interesting is the indefinable ... You meter everything that's going down on tape, and the lights are flashing, and you've got all these readings, but what you're looking to get on the record, there ain't a meter for. It's that feeling, that groove ... There's no meter in the world that can measure that.' (Quotes from Neil McCormick in the Daily Telegraph.) When Jelly Roll Morton's piano solos were recorded acoustically '23-4 the engineers were capturing that groove, and all the technology in the world can't make it happen; today record companies like CIMP (see Cadence) and Clarity (see Chico Freeman) are making superbly natural-sounding recordings, while Concord, Linn, Nessa, Nimbus, Spotlite, Zephyr and many other small labels never did let the technology get in the way (it is significant that these are mostly jazz labels). Sony/Philips, Pioneer and Samsung were all developing super-CDs starting with 20-bit sampling instead of 16, compatible with existing CD players but sounding better on new machines; it remained to be seen whether most fans could hear the difference.

ADDENDA
January 2010

Nielsen SoundScan reported that in 2009 domestic album sales in the USA, including downloads, were 373.9 million units, down 13% from 2008, falling for the eighth time in nine years. The biggest selling album of the year was Taylor Swift's Fearless, selling 3.2 million copies, compared to N Sync's No Strings Attached, which sold nearly ten million in 2000, the year that CD sales peaked, when Americans bought over 785 million albums. Music fans bought 1.16 billion individual songs from services like Apple's iTunes Store, an increase of 8.3% over 2008, but that actually represented a slowdown, because in 2008 sales of digital songs had increased 27% over 2007.

Meanwhile record shops have disappeared: mom-and-pop shops are long gone, and even Tower, HMV and Virgin have disappeared from the USA, despite the fact that there are still a great many music fans who would love to be able to browse in a record shop (a 'record' is a 'recording', whether an Edison cylinder or an aluminum beermat.) The industry had learned to rely too much on the million sellers, forgetting about the 100 releases that sell 10,000 copies each.

Yet there is hope. Only the industry has collapsed, not the music. in 2009, Jim Fusilli in the Wall Street Journal wrote about the music festival Lollapalooza, quoting some of the participants: 'The music business is upside down,' said singer-songwriter Robert Earl Keen. 'You don't tour to support your record. You put out a record to support a tour.' But that's the way it was in the golden age: Paul Whiteman and Duke Ellington started out making a living performing, with records as promotion. Another singer-songwriter, Andrew Bird, said, 'A record is kind of a moment in time. Maybe you nailed it, maybe you didn't. It's far more interesting to do it differently every night.' Dan Deacon said 'Digital media is so devalued. Real value is in live shows.' At Lollapalooza, Fusilli wrote, Deacon brought 30 or so musicians on stage 'to play his kind of joyful, experimental electronica.' In fact, with the collapse of the record industry, 130 years after Edison invented the phonograph, maybe the music itself has again become the important thing.