First stereo sound recording

1. The technology

2. The human subject

First stereo sound recording

In 2020 Patrick Feaster published in his blog an account of 'Berthold Laufer's Chinese Stereo Recordings of 1901', convincingly demonstrating the existence of the first known stereo sound recordings of the acoustic era. Recording simultaneously on two phonograph cylinders, one for vocals and one for instruments, when fully synchronised, as Feaster has achieved, the stereo effect is unmistakeable. There are actually some 500 cylinders held in two collections, of which the earliest are 25 pairs of cylinders recorded in Shanghai in the week ending 27 September 1901. These include 遊龍戲鳳 = Yu lung hsi fêng (Wandering Dragon Plays with Phoenix). Feaster argues that the pairs of recordings "were originally created to function as two parts of a coherent whole and to document a meaningful performance only when taken together," and that it could be held that, although "the recordings weren't meant to sustain a stereophonic illusion in playback," they were not accidental but intentional stereo, Laufer's purpose having clearly been to capture two different perspectives.

More than a quarter of a century later Duke Ellington and his Orchestra made some accidental stereo recordings ('Mood Indigo, Hot and Bothered, Creole Love Call' and 'East St Louis Toodle-o, Lot o' Fingers, Black and Tan Fantasy'—it's not clear which was recorded first), on 3 February 1932 for RCA-Victor. It was standard practice at that time to record using more than one microphone and disc cutter, allowing for backups in case something happened to the original. Although the records are fairly rare, a collector (Brad Kay) had both versions and noticed that, while they appeared to be the same performance, the sound mix was different on each (although the microphones were close together). When the two recordings were synchronized, it became acceptable stereo. [Fox] The resulting recordings are commercially available on disc 6 of the 24 CD set The Duke Ellington Centennial Edition.

Later Brad Kay discovered even earlier accidental stereo recordings, made with a dual lathe/mike setup. All these were of classical music, and the earliest such now identified is a 1929 recording of an excerpt from Massenet's Le Cid ballet music, conducted by Eugene Goossens, at the Queen's Hall in London. EMI, however, don't accept Kay's claim. [Kay, ARSC stereo controversy] The Goossens recording is on YouTube. Patrick Feaster finds more acceptable the Pristine Classical release Accidental Stereo, 1929-1933 (PASC 422), in which Mark Obert-Thorn and Andrew Rose have used Capstan software to mitigate wow and flutter, and in which the earliest recording is the introductory part of Stravinsky's The Rite of Spring, recorded 24 September 1929 in the Academy of Music, Philadelphia, by Leopold Stokowski and the Philadelphia Orchestra, originally issued as Victor 7227 and 7228 in album M-74.

Several stereophonic test recordings by Bell Laboratories, using two microphones connected to two styluses cutting two separate grooves on the same wax disc, were made with Stokowski and the Philadelphia Orchestra at Philadelphia's Academy of Music in March 1932. The first, made on 12 March 1932, of Scriabin's Poème du feu Op. 60, is the earliest surviving stereo recording that was heard as such at the time. [Copeland, p24; Stokowski] The two tracks recorded on this date are as follows: A, B.

The earliest commercially sold stereo recordings were made by the audio engineer and inventor Emory Cook (1913–2002), and released on his own record label. As his pressing technique involved two separate tracks on the same side of the record, he also had to market a 'binaural phonograph adapter' as pickup cartridge. Cook released around 50 binaural recordings, of which the earliest could have been his LP Rail Dynamics—engine sounds recorded on a Cook-modified Magnecord PT-6 tape recorder during the autumn of 1950 near Peekskill station, New York, first released as Cook Records COOK01070. [sleeve notes, Cook Recordings Inventory, 60th anniversary of the modern stereo record, Milner, Morton, Bruil] Rail Dynamics (now COOK01270) is still available from Smithsonian Folkways, as download or custom CD. However it's not clear that Rail Dynamics wasn't originally actually mono. That said, it's also clear that nearly the whole of Cook's catalogue was also available in binaural form. [Conly] If the 1950 recording of Rail Dynamics was not stereo (or binaural, in Cook's terminology), perhaps recording number 1000, Titus Moody's DDDs of Binaural /Introduction to Binaural, Cook's 1952 binaural sampler, may have been the earliest. [Cook Recordings Inventory, Discogs]

Quadraphonic sound (4.0 stereo): Although multi-channel sound had been experimented with both for audio and in movies since at least the 1930s, it wasn't until 1969 that technology and the audio industry were ready to commercialize it. At the Audio Engineering Society meeting in 1969 Vanguard records demonstrated discrete four-channel sound on a four-channel reel to reel deck. The first home quad recordings on open reel tape were released in 1970. ['Quad' Bob] According to CD-4guy, a seller on eBay, Vanguard's Quad demo tape numbered VSS-1 was the "First tape issued by Vanguard, perhaps the first tape that started Quad rolling, way back in 1969!" This discrete four-channel quadraphonic factory pre-recorded 7.5 ips open reel tape was a compilation of the following selections: Handel: Jephtha (Sinfonia); Berlioz: Requiem, Op. 5, Tuba Mirum; Mahler: Symphony No. 3, Fifth Movement; Mahler: Symphony No. 3, Sixth Movement; Joan Baez, 'Love is Just a Four Letter Word';  Buffy Sainte-Marie, 'Poppies'; and Jean-Jacques Perrey, 'Gypsy in Rio'.

Discrete reproduction is the only real quadraphonic system. As its name suggests, with discrete formats the original four audio channels are passed through a four-channel transmission medium, presented to a four-channel reproduction system, and fed to four speakers. This is defined as a 4-4-4 system.

Surround sound: The first documented use of surround sound was in 1940, for the Disney studio's animated film Fantasia. Its multichannel audio application was called Fantasound, and comprised three audio channels and speakers. The sound was diffused throughout the cinema, initially by an engineer using some 54 loudspeakers. The surround sound was achieved using the sum and the difference of the phase of the sound. In the 1950s Karlheinz Stockhausen experimented with and produced ground-breaking electronic compositions such as Gesang der Jünglinge (1955–1956, originally in five-channel sound) and Kontakte (1958–1960), the latter using fully discrete and rotating quadraphonic sounds generated with industrial electronic equipment in Herbert Eimert's studio at the Westdeutscher Rundfunk (WDR).

Ambisonic sound: Ambisonic sound attempts to recreate the true three-dimensionality of the recording space along all three spatial axes (most surround sound systems don't include height channels).

On 20 September 2002, Dolby premiered a master of the movie We Were Soldiers which featured a Sonic Whole Overhead Sound soundtrack, developed by Randall Wallace. This mix included, in two of its reels, a new ceiling-mounted height channel. It appears, though, that We Were Soldiers was a one-off experiment, and the top channel was never used again. [Benjamin Wright, in Théberge, Devine, and Everett]

The now-archived Aurophonie website lists six ambisonic systems, though one (Imax) is not fully ambisonic, as it has a single mono height channel. Three are proprietary ambisonic systems of different degrees of complexity, one is Dolby Digital Plus, the fifth being Hamasaki 22.2. Currently the most advanced 3D multichannel system is Dolby Atmos, which has a native 9.1 layout fed to up to 64 loudspeakers, including height channels. [Benjamin Wright, in Théberge, Devine, and Everett]

Ambisonics: Ambisonics proper is a full-sphere surround sound technique, covering sound sources above and below the listener, in addition to the horizontal plane. Unlike some other multichannel surround formats, its transmission channels do not carry speaker signals. Instead, they contain a speaker-independent representation of a sound field called B-format, which is then decoded to the listener's speaker setup. This extra step allows the producer to think in terms of source directions rather than loudspeaker positions, and offers the listener a considerable degree of flexibility as to the layout and number of speakers used for playback.

Ambisonics was developed in the UK in the 1970s under the auspices of the British National Research Development Corporation. The significant paper on the subject (then described as 'periphony') was by Michael A. Gerzon (1945–1996): 'Periphony: With-Height Sound Reproduction'. Journal of the Audio Engineering Society, 1973, 21(1):2–10.

It has proved an effective way to present spatial audio in Virtual Reality applications such as YouTube 360 Video, as the B-Format scene can be rotated to match the orientation of the user's head, and then be decoded as binaural stereo.

Binaural sound: The recording of binaural sound replicates the human listening experience directly, with microphones placed at left and right ears of a dummy head. Though giving the most accurate directionality, its drawback is that optimum playback requires the use of headphones. Possibly the first true binaural recording (as opposed to earlier recordings which were described by this term, but actually used what is now described as stereophony) was The Binaural Demonstration Record put out by Stereo Review in 1970. In the studio music recordings on Side Two (ZD-BN-1B) it used a home-made sculpted head known as the Blue Max; while on Side One (ZD-BN-1A), a selection of live field recordings, small capacitor microphones were taped directly onto the recording technician's head. The Side One recordings include the voices of men, women and children in a variety of urban contexts, along with a selection of sounds selected to display the binaural effect to advantage. [album sleeve and insert, 1970]

From 2016 a more advanced form of binaural sound has been developed, for use with VR systems, in which the listener's head can move about independently of the sound source, which remains static. [BBC Click, 2016-04-22]

Wave-field synthesis: Wave field synthesis (WFS) is a spatial audio rendering technique, characterised by the creation of virtual acoustic environments. It produces artificial wave fronts synthesised by a large number of individually driven speakers. Such wave fronts seem to originate from a virtual or notional source. Unlike traditional stereo, the localization of virtual sources in WFS does not depend on or change with the listener's position.

Early development of WFS began in 1988 at the University of Delft. Further work was carried out in the context of the CARROUSO project, promoted by the European Union (January 2001 to June 2003). In Europe, ten institutes were included in this research. The IOSONO WFS system was developed by the Fraunhofer Institute for Digital Media Technology (IDMT), at the technical University of Ilmenau. IOSONO technology includes an UPMIX! engine allowing up to 64 audio sources to be converted into object-based components, which can then be redistributed into 3D space and rendered into any speaker configuration (from 5 to 500 speakers), using a Spatial Audio Processor ^®. IOSONO has its headquarters at the KinderMedienZentrum in Erfurt, Germany, where it has its main 3D sound showroom. [IOSONO]

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