Sound recording involves the electronic detection of sounds and their preservation in analog or digital coding in a storage medium--usually a disc, tape, or film. In reproduction, or "playback," the encoded information is recovered from the storage medium, amplified, and fed to loudspeakers or headphones that re-create a facsimile of the original sound. In modern recording, sound-pressure waves in the air are detected by one or more \Lmicrophone\ls that translate the sound waves into electrical signals. Signals from many microphones may be combined in a "mixer" to produce a finished recording. In contemporary music, electrical signals corresponding to desired sounds may be generated directly by a SYNTHESIZER or \Tcomputer\t (see \Tcomputer music\t) and combined with microphone signals. The synthesized signals are transformed into sound during playback. Sound reproduction equipment that is intended for a special purpose, such as a public-address system, may be deliberately limited in volume level or frequency range. Equipment intended to reproduce music, however, is rated according to the accuracy, or fidelity, of its reproduction. Thus "high fidelity" systems are those which faithfully reproduce the original sound of music, authentically re-creating its full frequency range, dynamic range, and timbre (tone quality). Sound recordings and reproduction systems are characterized by the number of "channels" of sound that are provided: one (monophonic), two (stereophonic), or more (for "surround" or "quadraphonic" sound). RECORDING CLASSIFICATIONS Storage Media Sound recordings are classified into four general categories according to the technology employed to store the audio signal. Mechanical Recording. This technology is the basis of all \Tphonograph\t records. The audio signal is represented by an undulating groove in the surface of a cylinder or disc. For playback, the record spins on a turntable while a lightweight "stylus" traces the pattern of wiggles in the groove. Magnetic Recording. The basis of all \Ttape recording\t (both audio and video), magnetic techniques are also used for the sound tracks of some 70-mm motion-picture films. In all cases, a plastic tape or film carries a thin coating of magnetic material, usually iron oxide, on which a varying magnetic pattern is imposed during recording. Optical Film. This technology is used for the sound tracks of all but a few motion-picture films (see \Tcinematography\t). During recording a "light-valve" varies the amount of light passing through a magnetic field to reach the sensitized film. The strength of the magnetic field changes with the sound. In playback the developed film transmits a varying amount of light to a photocell, thus re-creating the electrical signal. Optical Disc. This is the basis of the \Tcompact disc\t (CD) and the optical \Tvideodisc\t. The signal is represented by a pattern of microscopic pits along a reflective spiral track in the disc. In playback, the pattern is "read" by a small laser and photocell. Signal Format Sound recordings are also classified according to the form of the signal that is recorded in the storage medium. Analog. All sound is characterized by a pattern of rapidly varying air pressure. In analog recording, that pattern is imposed directly upon the storage medium, as the undulating groove in a phonograph record, the varying magnetic pattern in a recorded tape, or the varying light pattern of a film sound track. The principal drawback of analog recording is that imperfections in the storage medium (for example, the particles of dust in a record groove) become part of the audio signal during playback. Frequency Modulation (FM). Used for recording the sound and picture in videodiscs and "HiFi" video cassettes, FM is also used in \Tradio and television broadcasting\t. The sound-wave pattern is represented by variations in the frequency of a "carrier" signal whose average frequency is above 1 megahertz (MHz). This approach requires complex circuitry, but it avoids the limitations of direct analog recording. Digital. Digital encoding is the fundamental data-processing method for most present-day computers and for a range of sampling and recording techniques in other fields (see \Tdigital technology\t). In sound recording, the sound is represented indirectly by a "binary" (two-state) code in which the recorded signal alternates between "on" and "off" states. Of several possible coding schemes, the most commonly used is "pulse code modulation" (PCM). Error-correction codes are included in the recording, allowing near-perfect recreation of the original audio signal during playback. In principle, each of these three recording methods (analog, FM, and digital) may be used to record either audio or video signals and may be employed with any of the four storage technologies, yielding numerous possible combinations. For example, mechanical disc sound storage with stylus playback has been used for analog recording (the familiar phonograph record), for FM recording (the CED videodisc system that was briefly marketed in the United States by \TRCA\t), and for digital recording (the Teldec Mini-Disk system, proposed as an alternative to the compact disc). THE PHONOGRAPH RECORD The phonograph record was the first successful medium for capturing, preserving, and reproducing sound. It remained the most popular recording medium for nearly a century. Early History Practical methods of recording sound waves on the surface of a disc or cylinder and playing them back were described in 1877 by Charles Cros in France and by Thomas \TEdison\t in the United States. Edison turned his idea into a practical working model, and he is credited as the real inventor of the phonograph. Whereas Edison's 1878 patent describes both cylinder and disc records, his original prototype and his later commercial production focused on the cylinder format. In Edison's phonograph a thin sheet of tinfoil was wrapped tightly around a pregrooved metal cylinder, which was turned slowly by hand. As Edison spoke into the mouth of a conical horn, the concentrated sound waves at the narrow apex of the horn caused a thin diaphragm to vibrate, and a blunt needle affixed to the diaphragm inscribed these vibrations as indentations in the foil. The same apparatus played back the recording, using a larger horn to amplify the weak vibrations picked up by the needle. Since the sound quality was barely good enough to reproduce intelligible speech, Edison foresaw its use principally as a dictating machine. During the 1880s, as the result of experiments sponsored by telephone inventor Alexander Graham \Tbell\t, the fidelity of Edison's phonograph was improved by replacing the foil with a layer of wax in which the vibrating stylus could engrave grooves of varying depth. If desired, the wax could be shaved smooth to prepare the cylinder for a new recording. Edison added an electric motor to turn the cylinder at uniform speed, and he conceived a method of producing several copies of a recording by metal-plating the cylinder to create a mold. Within a few years, production of cylinders and players had become a substantial commercial business. For a nickel per play, coin-operated players in "phonograph parlors" provided thousands of eager listeners their first experience of recorded sound. In 1888, however, a German immigrant named Emile \TBerliner\t demonstrated a working "gramophone" in which sound waves were engraved in a spiral groove on the surface of a flat disc, with the stylus vibrating from side-to-side rather than up-and-down. During the following decade Berliner improved the initially poor sound of the disc record and devised a method of mass-producing recorded discs at low cost from metal molds. He also began to manufacture disc players driven by spring-wound motors. His Victor Talking Machine Company, formed in 1901, quickly attracted both performers and the public to disc format. Cylinders provided slightly better sound, and Edison continued to produce them until the 1920s, but by 1910 the disc had become the dominant format, mainly because thousands of discs could be molded from just one "master." The ease of recording a song on a disc in a single performance attracted the world's leading singers--Enrico Caruso, Fyodor Chaliapin, John McCormack, and many others. Mass production allowed discs to be priced lower than cylinders. The acoustic phonograph rapidly gained worldwide popularity, though it suffered from two basic limitations. First, since the engraving stylus was moved only by sound vibration, it recorded loud sounds best, favoring opera singers and brass instruments over the subtle tones of violin and harp for example. Second, in playback, the acoustic output of the system depended on the mechanical power generated by the groove walls pushing the stylus back and forth. It reproduced the human voice remarkably well but could not generate the full volume or tonal range of orchestral sound. These limitations were only partly ameliorated by the use of large, elaborately contoured horns to focus the sound in recording and to amplify it in playback. Inspired by the 1912 invention of the "Audion" vacuum-tube \Tamplifier\t by Lee de Forest, however, by 1919 efforts had begun to develop electrically amplified phonographs. The Electrical Era Amplified recording and playback systems were developed at Bell Telephone Laboratories, the technology was licensed to other companies, and the first electrical recordings--using the microphone, developed from Alexander Graham Bell's telephone transmitter--were issued in 1925. The main advantages of electrical recording were practical. Soft sounds could be recorded, since the microphone's output was amplified to drive an electromagnetic cutting stylus. The microphone was connected by a flexible wire, so performers no longer needed to cluster around the mouth of a horn. In playback a less massive stylus assembly could be used, incurring less record wear, since its force no longer generated the sound directly; and with sufficient amplification the reproduced sound could fill a large room such as a motion-picture theater. Record manufacturing declined sharply in the 1930s because the advent of nationwide radio broadcasting promised consumers an endless supply of entertainment at no cost. In 1928 the Radio Corporation of America (\TRCA\t) purchased the Victor Talking Machine Company, forming \TRCA\t Victor; after two prosperous years, however, \TRCA\t converted most of the Victor factories to radio production. For over a decade, nearly all progress in sound reproduction was financed by the motion-picture industry and by laboratories that were not dependent on income from record sales. Wide-range condenser microphones were produced by Western Electric, while the "dynamic" (electromagnetic) loudspeaker was patented by General Electric in 1928. Most of the fundamentals of modern \Tloudspeaker\t design, including the correct use of a large "woofer" and small "tweeter" to reproduce low and high frequencies, respectively, were established in Hollywood about 1935. Experimental stereo recordings were first produced in 1931 at Bell Telephone Laboratories. Production of phonographs was halted by World War II, and the Japanese invasion of Southeast Asia cut off the supply of shellac, the base material that had been used for disc records since 1900. Forced to select a substitute, record manufacturers discovered the virtues of vinyl, a plastic resin derived from petroleum. It proved to be a fortuitous choice, for vinyl's freedom from surface noise was crucial to the postwar development of the long-playing record. The LP, Hi-Fi, and Stereo The playing time of the 12-inch (30.5 cm) record remained below 5 minutes from 1900 until 1948, when a team of CBS engineers headed by Peter \TGoldmark\t and William Bachman unveiled the microgroove "LP" (long-play) record. A playing time of 20 minutes per side was achieved by slowing the turntable speed from 78 to 33 revolutions per minute (rpm) and by employing narrow, closely spaced grooves. The malleability and low noise of vinyl, which is a congealed liquid rather than a brittle solid, facilitated the use of 0.003-in (.0076 cm) grooves in which sound waves are represented by microscopically small wiggles. Experimental 33-rpm records had been made during the 1930s, but they wore out quickly under the heavy stylus pressures of the time. The development of the new LP record included the design of low-mass tonearms employing a tracking force of less than 0.5 ounce (14 grams). In later development this force was reduced to 5 grams and, by the 1980s to 1 gram. Acceptance of the LP format was slow at first, for it required the purchase of a new record player, but the LP had the compelling advantage of accommodating an entire symphony or Broadway musical on one disc. In 1949, \TRCA\t introduced a second micro-groove format, a 7-inch (17.8 cm), 45-rpm disc. This became the standard format for "singles," while the 12-inch LP held song albums and longer compositions. For the next decade, most phonographs operated at all three speeds: 78, 45, and 33 rpm. The introduction of the LP disc was accompanied by two dramatic changes in the character of sound recordings. Full Frequency-Range Recording Postwar records faithfully reproduced the entire range of frequencies audible to the human ear, from 30 to 15,000 cycles per second. This was accomplished by "equalization," using amplifier circuits to strengthen high frequencies in recording and low frequencies in playback, so that all frequencies were reproduced equally well. The combination of full-range recording with the quiet vinyl disc launched the "high fidelity" era in which virtually any sound could be reproduced accurately in the living room. Instead of purchasing a radio-phonograph console, hobbyists in search of the best sound began experimenting with separate components (turntable, amplifier, loudspeaker, and so forth). Mastering On Tape Prior to 1950, all recordings were cut directly on a master disc. If a wrong note was played, the performance had to be repeated while the master disc was recut from the beginning. With the introduction of the magnetic tape recorder, the art of recording gained new freedom. Tapes are easily spliced, so mistakes can be edited out. The various parts of a performance can be recorded separately and combined to produce a finished tape. Tape machines are portable, so recordings can be made anywhere: in a European cathedral, at a New Orleans jazz festival, in the home of a folk singer. To make LP records, the master tape is played in a disc-cutting studio. The tape's signal is amplified to drive electromagnetic coils attached to a sharp-edged cutting stylus, which cuts a groove in a disc coated with acetate or soft copper. The resulting master disc is electroplated, forming a "metal master" that contains a perfect negative impression of the original disc, with raised ridges instead of grooves. This plating can be used to stamp our finished records. For quantity production the negative is replated to produce a positive mold called a "mother," which is repeatedly plated to produce negative stampers. Stampers representing the two sides of a record are mounted in the upper and lower jaws of a press, a "biscuit" of vinyl is placed between them, and the press is closed for about 20 seconds. After the molded disc cools it is removed, excess vinyl around the edges is trimmed off, the center-hole is punched, and the record is packaged. The high-fidelity movement revived interest in \Tstereophonic sound\t, the use of two recorded channels to enhance the realism of reproduced sound by providing the directional and spatial impressions that one is able to hear in live performances. Experimental stereo records employing separate grooves for the two channels had been demonstrated during the early 1950s, but this approach was mechanically complex and cut playing time in half. In 1957 the industry adopted the present system of "compatible" single-groove stereo, in which the left and right channels of sound are represented by diagonal waves cut into the two sides of a wall. Side-to-side vibration of the stylus corresponds to monophonic sound, while up-and-down motion produces the spatial impression of stereo. In stereo playback each ear hears sounds from both loudspeakers. To compensate for this blending of the two channels, recording engineers often use exaggerated microphone spacing (much wider than the width of the head), or microphones having a more sharply directional pickup pattern than the ears. When stereo recordings are heard through headphones the result is a wraparound sound that often seems to envelop the listener. A more realistic impression is provided by the relatively rare "binaural" recordings, which are made with microphones installed on a dummy head to produce a separation between the stereo channels that will be equal to the average distance between the ears. Binaural recordings are intended specifically for playback through headphones. To provide a more convincing aural impression through loudspeakers, the industry produced quadraphonic (four-channel) recordings and equipment during the early 1970s. But the records were marketed prematurely (before recording and playback techniques were fully developed) and in three incompatible disc formats, diluting their appeal. In the late 1970s recording engineers began to use digital tape recorders, improving the clarity of the resulting records. During the 1980s, as listeners turned to the cassette tape and the all-digital compact disc, LP sales declined sharply. MAGNETIC TAPE RECORDING In recent decades magnetic recording has become the most popular of all recording technologies, mainly because of the ease with which magnetic signals can be recorded, edited, copied, erased, or rerecorded. Magnetic recording is based on electro-magnetism. When an electric current flows in a coil of wire, it generates a magnetic field. Conversely, when a magnetic field moves near a wire, it generates an electric current in the wire. These physical phenomena are the basis of recording (conversion of an electrical signal to a magnetic pattern) and playback (conversion of the magnetic pattern back to an electrical signal). Early Development Whereas the principles of magnetic recording were described by Oberlin Smith in 1888, the first practical magnetic recorder was patented in 1898 by the Danish inventor Valdemar \TPoulsen\t. The device, called the Telegraphone, recorded on a long, thin steel wire. Wire recorders were manufactured for sale in the early part of the century, but their sound quality was inferior to the cylinder and disc records of the era. During the 1920s and 1930s some experimental recorders employed a 1/2-inch-wide (1.27-cm) steel tape instead of a wire, but a reel of this tape was heavy and costly. In 1928 a German patent was issued for a lightweight paper tape coated with iron powder, which proved superior to the continuous strip of solid metal. AEG Telefunken developed a tape recorder called the Magnetophon, and BASF, a German chemical firm, worked on the tape. BASF substituted cellulose acetate film for the breakable paper tape and developed a safe process for making tiny particles of iron oxide. (Safety was a problem because finely ground iron powder is flammable.) The Magnetophon was introduced in 1936, but its sound quality was mediocre--adequate for speech but not for music. In 1939, while experimenting with Magnetophons at the German radio network, engineer Walter Weber discovered that adding an ultrasonic tone to the signal during recording yielded much-better-sounding tapes. This "AC bias" was the key to high-fidelity magnetic recording. (Ironically, the use of ultrasonic AC bias to reduce distortion and noise in wire recordings had already been described in an obscure 1927 patent by W.L. Carlson and B.W. Carpenter of the U.S. Navy.) When the U.S. Army Signal Corps entered Germany at the end of World War II, its engineers were astonished at the excellence of AC-bias Magnetophon recordings. After the war, Magnetophon-like recorders were produced by the Ampex Corp. in California, at the request of popular singer Bing Crosby and the ABC radio network. (Because of the inadequate fidelity of disc recordings, musical programs were broadcast "live" to the eastern half of the U. S. and had to be performed again three hours later for the West coast. Tape allowed Crosby's show to be prerecorded and edited for broadcast without sacrificing sound quality.) The Ampex machines, recording on 1/4-inch-wide (0.64-cm) tape moving at 30 inches (76.2 cm) per second, were soon adopted throughout the recording and broadcast fields. With the help of better tapes and equalization circuits--which adjust amplification for accurate sound at all frequencies--tape speeds have repeatedly been halved (to 15, 7-1/2, 3-3/4, and finally 1-7/8 inches (36.1, 19.05, 9.525, and 4.76 cm) per second for cassettes), while the development of smaller magnetic heads has allowed several parallel tracks to be recorded on each tape. A four-track (bidirectional stereo) format has been the standard for nonprofessional recording since 1955. Meanwhile, studio recorders gained flexibility by raising the number of independently recordable tracks, from four tracks on 1/4-inch tape in the late 1950s up to 24 or 32 tracks on 2-inch-wide (5.08-cm) tape in the 1980s. Modern Tape Formats In a tape recorder the tape is spooled off a supply reel, threaded past the recording-playback heads, and rewound on a takeup reel. During the 1950s and 1960s many attempts were made to enclose reels and tape in a self-threading package to eliminate the inconvenience of tape handling. By far the most successful design was the "compact cassette," a small plastic box containing reels and tape, which was introduced by Philips in 1964. Since the tape in the cassette was only 0.15 inch (0.38 cm) wide and moved at only 1-7/8 inches per second, a pocket-size cassette could accommodate an hour-long recording. By 1970, with refinements in electronics, and in tapes and recording heads, the cassette was able to deliver high-fidelity stereo sound. Cassette tapes made music portable. In-dash stereo cassette players were developed for automobiles during the 1970s, followed by a broad range of battery-powered go-anywhere tape players in the 1980s--from tiny "Walkman" headphone units to loud "boom-box" portables. By 1983 the compact cassette was the most popular medium for recorded music, and it was widely used for other applications as well: in telephone answering machines, for example, and as a program-storage medium for low-cost home computers. An even smaller format, the "micro-cassette," running at a tape speed of 15/16 inches (22.38 cm) per second, has largely replaced the standard cassette for recording dictation. In recent years magnetic recording formats have diversified rapidly. Today's recorders may be classified in at least four ways: (1) by signal handling: audio-only or video with sound; (2) by physical form: open-reel or cassette--most open-reel systems are intended for professional use, while consumer formats employ a cassette with an automatic tape-threading mechanism; (3) by recording method: direct analog, frequency modulation (FM), or digital pulse code modulation, or PCM; (4) by head-tape relationship: stationary-head or helical-scan. In stationary-head machines the tape moves in a straight line past the heads, and signals are recorded on parallel tracks along the length of the tape. In helical-scan recorders the magnetic heads are mounted in the rim of a rapidly spinning drum that is tilted at an angle so that, as the tape moves slowly past, the heads record diagonal tracks across the tape. Fixed-head design is used in analog recorders, while helical-scan is found in video and digital (DAT) recorders. \Tfrequency modulation\t (FM), developed in the 1930s for broadcasting, was adapted for recording video signals and scientific data during the 1950s. In videodiscs, developed in the 1970s, picture and sound are both recorded as FM signals. "HiFi" video cassette recorders (VCRs) use videotape in helical-scan format to record picture and sound as FM signals (see \Tvideo recording\t) (While HiFi VCRs are mainly used to record television pictures with sound, they can also be used to record sound only, achieving a level of fidelity that is second only to a digital tape recorder. In ordinary VCRs, however, only the picture is recorded as an FM signal in helical-scan format, while the sound is recorded directly in analog form by a stationary head. Magnetic Tape Although recording tapes have evolved in quality during the past half-century, their fundamental composition remains the same: a mixture of iron oxide particles and lacquer coated on a thin plastic film. The manufacture of tape begins with the base film, or backing. While cellulose acetate is still used occasionally, most high-quality tapes employ polyester film such as Mylar, selected because it is flexible, strong, and unaffected by humidity. Large rolls of clear film, typically two feet (0.6 m) wide and less than 0.001 inch (0.0025 cm) thick, are coated continuously and then slit into individual tapes. The most widely used magnetic material is gamma-ferric oxide, a form of finely powdered rust. "High-bias" audio tapes and most videotapes use chromium dioxide or iron-oxide powder treated with cobalt. In high-performance "metal-particle" tapes, powdered iron is combined with other metals instead of oxygen. In all modern tapes the magnetic particles are needleshaped, typically 5 millionths of an inch thick by 20 or 30 millionths of an inch long (0.000127 mm by 0.000503 or 0.000762 mm). Modern tapes use plastic resins for the "binder" that glues the magnetic powder to the backing. The resin is mixed with a solvent such as alcohol to produce a liquid of paintlike consistency. Then the magnetic particles are stirred in and mixed until every particle is completely coated. (If particles clump together, metal-to-metal, their magnetic fields may cancel out.) A thin coating of the mixture, less than 0.0005 inch (0.0127 mm) thick, is spread on the clear film. While the coating is still in semiliquid form, it passes under a powerful magnet that rotates the needlelike particles to orient them lengthwise along the tape. Then the tape passes through drying ovens that evaporate the solvent, leaving the powder-binder mixture glued securely to the backing. The dry tape is pressed between polished metal rollers to produce a smooth surface, and then slit into individual tapes of the desired width. Recording A tape recorder consists of two systems: (1) a "tape transport" mechanism that moves the tape past the heads at a uniform speed; and (2) record-play electronics that prepare the signal for recording and then amplify it in playback. The actual recording or playback is done by a "head," a small electromagnet mounted in a shielded case. For recording, signal current generates a magnetic field in the head that is imposed on the magnetic particles in the tape. For playback, the magnetic fields in the moving tape generate tiny electric currents in the head. At each instant the head is in magnetic contact with a very small area of tape (in an audio cassette, that area is about 0.0001 inch wide by 0.02 inch high 5 or 0.00254 mm by 0.508 mm) containing several thousand particles. (Some recorders have separate heads for recording and playback.) Each microscopic particle of iron oxide is an individual bar magnet oriented lengthwise on the tape, with a north pole and a south pole. In an unrecorded tape, approximately half of the particles are magnetized with their north pole forward, and the other half are south-forward. When exposed to a magnetic field strong enough to overcome their "coercivity" (resistance to change), the particles adopt the imposed field direction, reversing poles if necessary. Thus the process of recording is simply one of flipping each particle's magnetic orientation one way or the other. Once set, the particles retain their magnetic orientation until exposed to another strong field. Since a digital signal consists only of ones and zeros, a digital recording flips the magnetic pattern alternately full-north or full-south. For analog recording, the audio signal is combined with a strong "AC bias" signal that alternates from north to south and back approximately 100,000 times per second. The result is that the audio waveform is faithfully represented by the percentage of north-forward particles at each location along the recorded track. To erase a recording, an "erase head" exposes the tape version of the bias signal, whose rapid polarity reversals leave approximately half of the particles magnetized in each direction. Copying Recordings When an LP or CD is pressed to make copies, an entire hour of music is duplicated in each 30-second pressing cycle. To make prerecorded tapes, the master tape is run and copied from beginning to end, a process that occurs simultaneously on dozens of recorders. Usually both the master and the copies are run 64 times faster than normal, so an hour-long tape is copied in less than a minute. One of the principal attractions of a tape recorder is the ease with which it can copy existing recordings or tape live radio broadcasts. In many stereo systems, signals from records, CDs, radio, or other tapes can be fed to the tape deck for recording. A tape may be copied either by connecting two recorders together or by using a "dubbing deck," a recorder with two cassette compartments and internal connections for copying. This activity, like the photocopying of books and magazine articles, may be unethical if it violates the copyright of the person who created the original recording. Illicit copying has been so widespread, according to the record industry, that they have proposed legislation either for copy-prevention circuits to be built into the new digital tape recorders or for special taxes to be imposed on tape sales and used for royalty payments. DIGITAL AUDIO In analog recording systems, a representation of the sound wave is stored directly in the recording medium. In digital recording what is stored is a description of the sound wave, expressed as a series of "binary" (two-state) numbers that are recorded as a simple on-off signal. The methods used to encode a sound wave in numeric form and accurately reconstruct it in playback were developed during the 1950s and 1960s, notably in research at the Bell Telephone Laboratories. Digital recording became practical about 1970 with the arrival of integrated-circuit chips that allowed the circuitry of a room-size computer to be shrunk to desktop size. Although usable digital recorders were developed during the early 1970s in several nations, Sony and Japan Victor Corp. first made digital recording practical by manufacturing converters that store digital audio signals on videotape, using the same helical-scan recording technique that is used for taping television signals. By 1976 major record companies were making digital LPs from digital master tapes. Sony's PCM-F1 converter (1981), allowing digital recording on a consumer-grade VCR, made the process so inexpensive that even the smallest record companies could afford it. Digital audio tape (DAT) recorders for consumer use were introduced in 1987. Just as most digital audio recording is done on videotape, the compact disc (CD) digital record is based on technology created for the optical videodisc. The "Laservision" system, which uses a low-power laser to read a video signal encoded as a series of microscopic pits in a reflective disc, was developed by Philips in Holland during the mid-1970s, and was later adapted for digital audio as the compact disc recording. While American companies created the ingredients that made digital records possible (lasers, plastics, computer circuits, pulse-code-modulation encoding, and integrated circuit chips), the CD system was developed by Philips in Europe and Sony in Japan. An industry committee adopted the CD format as an international standard, and CDs first appeared on the market in 1983. The CD is 12 cm (4.7 in) in diameter and contains up to 74 minutes of music in two-channel stereo, or up to 37 minutes of four-channel sound. The success of the CD has spawned several alternative formats: the mini-CD, a 3-inch (7.6-cm) CD limited to 20 minutes of music; CD-V, or CD-Video, containing 20 minutes of music plus 5 minutes of video; CD-ROM, CD-read only memory, containing 500 megabytes of computer-coded data or programs; and CD-I, CD-Interactive, containing a mix of audio, video, and computer programming for entertainment or educational use. Sampling and Quantization While there are many methods of representing a signal in digital code, the approach used for high-fidelity audio is called linear pulse code modulation, or PCM. It is a two-stage process using both sound sampling and binary quantization. At regular intervals (44,000 times per second), a "sample and hold" circuit momentarily freezes the audio waveform and holds its voltage steady, while a quantizing circuit selects the binary code that most closely represents the sampled voltage. In a 16-bit PCM system the quantizer has 65,536 possible signal values to choose from, each represented by a unique sequence of 16 ones and zeros. With 88,000 16-bit conversions per second (44,000 in each channel), a total of 1.4 million code bits are generated during each second of music, or five billion bits per hour. Error Correction Much of the circuitry in a digital tape recorder or CD player is devoted to detecting and correcting any bit-reading errors that might be caused by microscopic tape flaws, disc-pressing defects, dust, scratches, or fingerprints. Error correction is based on "parity" testing. When the recording is made, an extra bit is added at the end of every digital code, indicating whether the number of "ones" in the code is odd or even. In playback this parity count is repeated to detect whether any bits have changed. By cross-checking parity tests involving various combinations of the bits in each code, it is possible to identify exactly which bits are wrong, and to correct them, reconstructing the original code exactly. This high-speed arithmetic is simple work for the microprocessor that is contained in every digital recorder and CD player. The data samples are "interleaved" on the tape or disc in a scrambled sequence, so that samples that originally were contiguous in time are not placed adjacent to each other on the disc. Correct order is restored during playback, by briefly storing the digital data in a "random access memory" and reading it back in a different order. During this de-interleaving, any large block of false data caused by a scratch or pressing flaw will be split into small groups of bad data between good samples, making it easier for the parity-checking system to identify and correct the lost data bits. The Compact Disc The digital code is engraved on the disc in the form of microscopic pits along a track that begins near the center of the disc and spirals outward toward the edge. The pits are approximately 0.5 micrometers (um) wide, with a uniform 1.6-um spacing that causes light to be diffracted in a colorful rainbow spectrum (1 um is one-millionth of a meter, or 1/25,000 of an inch). Mass-production of CDs, like LPs, begins with a master disc. It may be coated with copper, in which the pits are embossed by a piezoelectric stylus, or with light-sensitive "photo-resist." In the latter case a laser is switched on and off by the digital code, and a chemical bath dissolves away the laser-exposed areas to form the pits. Nickel-plating and molding operations yield a negative metal stamper that is used to impress the spiral pattern of pits on the final plastic disc. The pitted surface is coated with a molecule-thick layer of reflective aluminum, overcoated with a protective coating of lacquer, and covered by the printed label. For playback, a laser is focused through the transparent rear surface of the disc onto the aluminized pits, and an optical sensor detects the alternately strong and weak reflection. Small scratches and dust particles on the plastic surface have little effect, since they are out of focus. The disc plays from the center out, rotating at about 500 rpm initially and gradually slowing to about 200 rpm as the spiral track approaches the rim, so that the pits pass under the laser beam at a constant linear speed of 1.25 meters ( 4 ft per second. After deinterleaving and error correction, the digital code is fed to digital-to-analog converters and output filters that re-create the original audio waveform in each channel. Digital Audio Tape Digital audio tape (DAT) machines made for at-home recording and in-car playback employ the same 16-bit PCM code as the CD, with similarly elaborate interleaving and error correction. The digital code is recorded on tape in helical-scan format, like that of a VCR, using a smaller version of the mechanism in an 8-mm VCR. MODERN STEREO SYSTEMS After World War II the near-simultaneous arrival of full-frequency-range disc recording, magnetic tape recording, and the long-playing record stimulated popular interest in "high fidelity" sound and a rapid expansion of the record industry. Since then, interest in better and more convenient sound reproduction has been stimulated by a constant parade of innovations--stereo tapes (1954); the compact acoustic-suspension loudspeaker (1955); stereo LP records (1958); stereo FM radio (1962); the spread of multitrack recording during the 1960s, adding new variety and creativity to rock music; the arrival of transistors during the same period, making equipment smaller, cheaper, and more reliable; Dolby noise reduction (1967); stereo cassette tapes (1970); four-channel sound (early 1970s); Dolby Stereo sound for movies (1976); digital recording (late 1970s) go-everywhere "Walkman" portable headphone stereo (1980); the compact disc (1983); the HiFi stereo VCR (1984); stereo television sound (1985); the DAT (1987); and so on. Today's music listener has a very broad range of equipment choices. The highest sound quality, and the greatest operating flexibility, are obtained by connecting individual audio components to form a system. "Rack systems" (audio components preselected and installed in a rack by a manufacturer or dealer) offer less flexibility, "One-piece" stereos with everything housed in one chassis except the loudspeakers, sacrifice sound quality in favor of simplicity and low cost. A similarly broad range of cost and potential sound quality is available in stereo systems designed for installation in automobiles. In recent years, American listeners have spent more money each year on car hi-fi than on audio equipment for the home. Anatomy of a Stereo System The following audio components, in combination, make up a complete stereo system. The Record Player. This component consists of three subassemblies: a pickup cartridge whose stylus assembly traces the groove and translates its wiggles into an electrical signal (see \Tcartridge stereo\t); a tonearm that holds the cartridge; and a turntable whose motor spins the record at the required speed. The turntable plater may be part of the motor assembly ("direct" drive), or the motor's rotation may be transferred to the plater via a rubber belt (belt-drive), or a rubber wheel (idler drive). The Tuner. This component receives radio broadcasts, usually from both AM and FM stations. In a digital tuner the tuning frequency is set in discrete steps by digital calculating circuits; in an analog tuner, stations are tuned by adjusting a variable capacitor. Either type may have a "digital" (numeric) frequency display. True digital broadcasts, in which digital signals are transmitted to the home, employ UHF-TV or cable TV channels, and video-based receiving equipment. In stereo FM, the quality of the reception depends as much on the antenna's size, aiming, and location--preferably on the roof--as on the quality of the tuner. With an inadequate antenna or poor location, FM signals may be distorted by "multi-path" interference, caused by the reception of both direct and reflected signals, particularly in a moving car. Amplifiers. An integrated, or complete, amplifier consists of a preamplifier and a power amplifier. The power amplifier drives the loudspeakers. It has a power supply (transformer and filter capacitors), plus an output stage that behaves as an electronic gate controlled by the audio signal, opening and closing to regulate the flow of current from the power supply to the loudspeakers. The flow of electric current through each loudspeaker's voice coil causes it to vibrate back and forth, producing sound. The preamplifier has two functions: it amplifies and equalizes the very small signal produced by the stereo cartridge or the other inputs--tape cassette, tuner, and so forth; and it includes the operating controls--the input selector and volume and tone controls. The receiver is a tuner and an integrated amplifier, combined in a single chassis for convenience and economy. Loudspeakers. A loudspeaker system usually consists of two or more drivers (woofer, midrange, tweeter); a crossover circuit that channels low frequencies to the woofer, high frequencies to the tweeter, and so on; and a cabinet. Speakers are classified by the way the crossover divides up the frequency range (2-way, 3-way, and so forth) and also by the woofer's relationship to the cabinet (acoustic suspension, ported, corner horn, open-baffle). Headphones. Headphones are simply miniature loudspeakers that play directly into the ears. They provide dramatic stereo separation and vivid clarity, without the blending that occurs when sounds bounce around in a room, and they require so little power that they can be driven from battery-operated radios and tape players. Other Components. In addition to a record player and a tuner or receiver, a modern stereo system is likely to include a compact disc player and a tape recorder of some kind--usually, a cassette player but, for very serious high-fidelity enthusiasts, an open-reel player; or a hi-fi videocassette recorder used for taping hi-fi sound. Digital audio tape recorders are available, but they will remain fairly rare and expensive, at least until the question of tape copying is resolved. A stereo system may also include signal-processing accessories that manipulate and enhance the sound, such as a graphic equalizer (an elaborate tone-control device), a dynamic-range expander, a sub-woofer for more powerful bass, or a surround-sound unit with extra loudspeakers to reproduce the spaciousness of sound that can be heard in a concert hall. Rating Performance While educated ears are the best judge of any sound reproduction system, many of the desirable performance qualities of audio equipment can be measured in laboratory tests, which give ratings for distortion and noise (the equipment should add as little as possible to the sound), output power (the more the better), dynamic range (the larger the number the better), and frequency response (ideally, the equipment should reproduce the entire range of frequencies audible to the ear, from 30 Hz to 15,000 Hz, with only small deviations from accurate response). Component Placement Most stereo equipment can be installed wherever it proves most convenient, with two exceptions. A record player should be on a stable, vibration-free shelf. And for good stereo, the loudspeakers must be equally distant from the listener. If they are not, the sound from the more distant speaker will arrive slightly later, unbalancing the stereo image. Most loudspeakers should be elevated on stands or shelves so that their sound can travel in a straight-line path, unobstructed by furniture. Noise Reduction All recording processes add some low-level noise to the reproduced sound. LP records have high-frequency hiss, caused by minute groove irregularities, and low-frequency "rumble"--often due to vibration from the turntable motor and bearings. Tapes add hiss because of the imperfectly random distribution of magnetic particles in the binder. Digital recordings have low-level "quantizing" noise because of the finite number of "bits" used to measure the signal. Even purely electronic circuits add noise because of tiny irregularities in the flow of electrons. Much of the excellence of modern sound reproduction is due to processes that reduce background noise so that every detail of the recorded sound is clearly audible against a background of velvety silence. The first of these processes is "equalization," or EQ, used in every recording medium. The naturally weak high-frequency overtones in music are boosted by a standard amount during recording; then in playback the highs are cut back by the same amount, restoring the original tonal balance and reducing noise added by the medium. In most audio equipment, equalization is built in and needs no attention from the user. Cassette tapes, however, have two different EQ standards, one for "normal" tape and a different EQ for "high bias" and "metal" tapes. For best results, the "tape" switch on the recorder must be set to match the type of tape used, both in recording and in playback. To suppress background noise more thoroughly, noise-reduction circuits operate by compressing the dynamic range of the signal during recording, thereby boosting the level of the quietest sounds, especially at high frequencies. The process is reversed during playback, pushing the quiet sounds down to their original level and pushing noise down at the same time. Dolby Laboratories has created the most widely used noise-reduction systems for professional studio and cassette recording, motion picture sound tracks, and stereo television sound. DBX provides similar systems. SOUND RECORDING ON FILM As early as the 1890s, Thomas Edison produced a series of "talking kinetoscopes," motion pictures that were accompanied by sound from cylinder phonographs. Many other attempts were made to bring sound to the movies, but all faced the same obstacles: inadequate volume to fill a theater and uncertain synchronization between sound and picture. Both problems were ultimately solved by Lee De Forest. His invention of the Audion vacuum tube paved the way for electrically amplified sound reproduction, and during the 1920s he developed a "Phonofilm" system for recording sound directly on the film. By 1930, Phonofilm had been adopted by the major film studios, and, with refinements, it remains the basis of optical film sound to this day. The optical sound track is recorded in the narrow space between the edge of the picture frame and the sprocket holes. The exposure of the sound track is controlled by a light valve consisting of a pair of metal ribbons that flex when driven by the amplified audio signal, admitting varying amounts of light to the film. In variable-density recording, the light valve varies the exposure of the film. In variable-area recording (used for 35-mm Hollywood films), the light valve varies the width of the exposed sound track. In either case, when the developed film is projected, the sound track transmits varying amounts of light to a photo electric cell, generating an electrical signal that is amplified to drive a loudspeaker. Some films are produced with magnetic sound tracks for use in specially equipped theaters. The 70-mm Dolby Stereo format, for example, is a six-channel magnetic recording, in which iron-oxide stripes, located along each edge near the sprocket holes, are placed on the film after it has been exposed and developed. In a motion-picture theater, the loudspeakers are located behind the screen, which is perforated with many small holes to let the sound pass freely. Additional speakers may be installed along the walls for "surround" effects. THE RECORDING INDUSTRY History Since the early days of the Edison wax cylinder, the recording industry has been a complex mixture of art and highly profitable commerce, aided both by the universal appeal of music and by world wide standardization of every important recording format. Records and tapes of today's popular artists are heard around the world--a situation that is strikingly different from other media industries. (Because of language differences, book and magazine distribution tends to stop at national boundaries; the international sale of videocassettes is crippled by incompatible television formats.) The recording industry has had two immensely profitable "golden" eras. The first spanned the three decades from the arrival of the mass-producible disc record in 1901 until about 1930. The second has extended from the early 1950s to the present day. Before and during World War I, records were the most popular form of home entertainment. After a brief postwar slump, the arrival of electrical recording in 1925 stimulated a strong resurgence in record sales during the late 1920s. But the Depression of the 1930s and the growth of radio--which offered entertainment at no cost--reduced record sales by 93 percent in just four years. The industry was saved by the repeal of Prohibition in 1933: records were needed to restock \Ljukebox\les in reopened taverns and lounges all over the country. Coin-operated record players spread rapidly to diners, drugstores--anywhere that people gathered. After World War II, record companies worried that the growth of television would kill record sales, as radio did after 1930. In fact, TV spawned the industry's second golden era. When TV took over the drama programs, comedies, and soap operas that had been radio's mainstay, radio stations were left with nothing to broadcast. To stay on the air, they turned to playing records full-time, providing unlimited exposure for the record industry's product. With this saturation campaign of free advertising on the radio, plus the allure of long-play records and high-fidelity sound, record sales boomed. As radio stations competed for audiences by broadcasting top-40 hits and new records, they stimulated the explosive growth of rock music. The arrival of the small, portable transistor radio completed this process by giving teenagers the freedom to listen privately, away from the living room. Annual U. S. record and tape sales grew from $200 million in the early 1950s to over $7 billion (1989), spurred by the success of compact discs, which were rapidly overtaking cassettes as the most popular recorded medium. Recording Music There are two basic approaches to making present-day recordings. In documentary-style recording, the goal is to capture a musical performance and produce a recording that sounds substantially like a concert. If the performers are few in number, they visit a recording studio. If the ensemble is large, like a symphony orchestra, recording equipment is brought to the concert hall. From 2 to 20 microphones are mounted on stands or hung on strings. As the performers rehearse, the microphone signals are mixed as needed to obtain a well-balanced sound; then the music is performed and is recorded on 2, 4, or 8 tracks. If mistakes occur, portions of the music are repeated; later the tape will be spliced to replace faulty sections with the correct versions. When a satisfactory recording of the entire performance has been obtained, the edited tape is copied and used to produce LPs, cassettes, and CDs. For many rock groups a recording studio is an experimental laboratory where sounds are invented and a performance is assembled, one track at a time. A recording may begin as a vocal line or a rhythmic pattern on one track of a 24-track tape. Additional tracks (piano accompaniment, keyboard synthesizer, drums, guitar flourishes) are recorded individually. Next, various tracks may be enhanced by using signal processors to add "effects"--changes in tone quality, distortion for crispness and "bite," echo, doubling, electronic reverberation, and so forth. Finally all of the tracks are mixed and remixed until a satisfactory result is achieved. The aim is to achieve a record that stands on its own merits as an artistic achievement, independent of whether it can be duplicated in concert. Since there is so much experimenting involved, many rock groups have their own small recording studios where much of the preliminary work is done. A new kind of recording technology, derived from research in the creation and recording of \Telectronic\t music, is beginning to make the studio recording process much more flexible. Called "tapeless" recording, it stores digitally-coded sound in the random access memory (RAM) of a specially-equipped computer, and on the computer's hard disc for permanent storage. Using RAM gives the recording technician almost instantaneous access to what has just been recorded, a feature that can be immensely valuable in dubbing additional sounds or patching imperfect notes. Recordings can be edited without physically handling the recording medium, as tape editors must do. Sounds can be altered, cloned and repeated; tempos changed without altering pitch; music and sound effects synchronized electronically. The technology has had particularly wide use in motion picture sound recording, where electronic signals create and coordinate visual, musical, and sound effects to a precision that was only crudely attainable in the era before computers. A recording studio consists basically of two rooms: a large room that has been carefully soundproofed so that sounds can be recorded accurately with the assurance that no unwanted noise will intrude; and a small "monitoring" room containing the tape recorders, mixers, effects processors, and high-quality loudspeakers that can be used to evaluate the recording. Peter W. Mitchell Bibliography: Alkin, Glyn, Sound Recording and Reproduction (1987); Berger, Ivan, and Fantel, Hans, The New Sound of Stereo (1985); Eargle, John, Sound Recording (1980) and Handbook of Sound System Design (1989): Gelatt, Roland, The Fabulous Phonography, 1877-1977 (1977); Maestas, Bob, and Goldfield, Paul, Sampling Basics (1989); Nardantonio, D.N., Sound Studio Production Terchniques (1989); Pohlmann, Kenneth, Principles of Digital Audio (1985); Read, Oliver, and Welch, Walter, From Tin Foil to Stereo: Evolution of the Phonograph, 2d ed. (1976); Thom, Randy, Audiocraft (1989); Woram, John, The Recording Studio Handbook (1983).