TY - JOUR TI - The Art and Science of Acoustic Recording: Re-enacting Arthur Nikisch and the Berlin Philharmonic Orchestra’s landmark 1913 recording of Beethoven’s Fifth Symphony AU -Aleks Kolkowski AU -Duncan Miller AU -Amy Blier-Carruthers PY - 2015 VL - Special Issue: Communications IS - Spring 2015 KW - acoustic recording KW - re-enactment KW - Royal College of Music KW - tacit knowledge KW - wax discs AB - The Art and Science of Acoustic Recording was a collaborative project between the Royal College of Music and the Science Museum that saw an historic orchestral recording from 1913 re-enacted by musicians, researchers and sound engineers at the Royal College of Music (RCM) in 2014. The original recording was an early attempt to capture the sound of a large orchestra without re-scoring or substituting instruments and represents a step towards phonographic realism. Using replicated recording technology, media and techniques of the period, the re-enactment recorded two movements of Beethoven’s Fifth Symphony on to wax discs – the first orchestral acoustic recordings made since 1925. The aims were primarily to investigate the processes and practices of acoustic sound recording, developed largely through tacit knowledge, and to derive insights into the musicians’ experience of recording acoustically. Furthermore, the project sought to discover what the acoustic recordings of the past do – and don’t – communicate to listeners today. Archival sources, historic apparatus and early photographic evidence served as groundwork for the re-enactment and guided its methodology, while the construction of replicas, wax manufacture and sound engineering were carried out by an expert in the field of acoustic recording. The wax recordings were digitised and some processed to produce disc copies playable on gramophone, thus replicating the entire course of recording, processing, duplication and reproduction. It is suggested that the project has contributed to a deeper understanding of early recordings and has provided a basis for further reconstructions of historical recording sessions. N1 - The first recording of an entire symphonic work was of Beethoven’s Fifth Symphony by Friedrich Kark conducting the Odeon Streich-Orchester in Berlin in 1910, and issued by the Odeon label as four double-sided discs. The Berliner Philharmoniker was founded in 1882 and by 1913 had established an international reputation as a leading orchestra. Arthur Nikisch had been their principal conductor since 1895. His importance as a conductor is discussed by Elliot Galkin (1988, p 639). An overview of the Berliner Philharmoniker's early history is available at: http://www.berliner-philharmoniker.de/en/history/arthur-nikisch/ (accessed on 2 February 2015). N1 - The period of acoustic recording spans from Thomas Edison’s invention of the tinfoil phonograph in 1877 until 1925 and the swift change-over to electrical recording, pioneered by Joseph Maxwell and Henry Harrison at the Western Electric Laboratories. For a concise history of both the acoustic and early electrical eras, see Copeland (1991, pp 7–18). N1 - They were, for example, described as a ‘triumph’ and ‘beautifully interpreted’ in The Talking Machine World, 15 January 1914, p 47. N1 - Roland Gelatt has gone as far as describing pre-First World War orchestral recordings as being ‘laughable travesties of music’ to the modern ear (Gelatt, 1977, p 184). N1 - The shift from the romanticism of acoustic recordings towards greater realism (and eventually Hi-Fi) in phonography was the subject of debate in articles that appeared during the 1920s and 30s in the Gramophone. Editor and founder Compton Mackenzie stood for accuracy of reproduction (Mackenzie, p 2), while for contributor Cedric Wallis, the older (acoustic) recordings, ‘by their very unrealism, have the quality which goes to make a great painting transcend its subject matter’ (Wallis, 1936, p 10). N1 - An advertisement in the German record brochure Grammophon, January 1914 (EMI Archive), lists the orchestral instrumentation on the records: Kleine Flöte, 2 grosse Flöten, 2 oboen, 2 Klarinetten, 2 Fagotte, 2 Hörne, 2 Trompeten, 3 Posaunen, Pauken, – Stricherkörper [body of strings] (EMI Archive). N1 - Duncan Miller and Paul Morris have been most active in the field of wax cylinder manufacture and recording: http://www.paulmorrismusic.co.uk/WaxCylinders.asp (accessed on 11 February 2015). In the research field, George Brock-Nannestad has been particularly active in the reconstruction of acoustic recordings and his published articles are referenced here. Mr Brock-Nannestad also attended the re-enactment sessions at the RCM in November 2014, as an independent onlooker and adviser to the project. N1 - Disc Catalogue nos.: HMV 040786/87 (Andante Pts. 1&2); 040784/85 (Allegro Pts. 1&2); 040788/89 (Scherzo; Finale Pt. 1 & Pt. 2); 040790/91 (Finale Pts. 3&4). The Deutsche Grammophon Aktiengesellschaft was at this time partly owned by the Gramophone Company in England and was also linked to the Victor Talking Machine Company in the USA. N1 - His Master’s Voice New Records, p 1; Grammophon, February 1914, p 3 (EMI Archive) N1 - A Gramophone & Typewriter Co. ‘Senior Monarch’ (1905–11) fitted with an ‘Exhibition’ soundbox was a top-of-the-range gramophone that would have been up to the task of playing orchestral records (Oakley and Proudfoot, 2011, p 39). N1 - His Master’s Voice New Records, January 1914, p 2 (EMI Archive) N1 - Nikisch’s contract with the Gramophone Company is held in the Arthur Nikisch: Artist’s File, at the EMI Archive. N1 - As a resident in Leipzig, Germany, an enemy country, Nikisch’s contract became void and he received no further royalties from the Gramophone Company after 1914 (legal papers, Arthur Nikisch: Artist’s File, EMI Archive). N1 - Letter from Leo B Cohn to the Recording Department, The Gramophone Company Ltd., Hayes, 20 June 1911 (EMI Archive). N1 - Such overload or distortion was termed ‘blasting’. The experience and technique of singing into a recording horn are wonderfully described by the American soprano Anna Case in Harvith and Harvith, (1987, pp 43–44). N1 - Detailed notes and sketches of musicians’ placement in relation to recording horns is given by recording expert William Sinkler Darby to his colleagues in Copy of Letter to Scheuplein, Hampe I & Hampe II (1907) and in reports by W M Gaisberg, Gramophone Company recording manager, in correspondence with Theo Birnbaum, its managing director (1907, EMI Archive). N1 - Such re-instrumentation is described in Batten, p 35, and in internal documents of the Gramophone Company: The Victor Recording Laboratory, No.2, The Victor Orchestra (copy of letter to Scheuplein, Hampe I and Hampe II, 1907, EMI Archive). The practice of re-scoring orchestral music with reference to a good selection of acoustic recordings is found in Melville-Mason (1977, pp 95–100). N1 - Augustus Stroh, Improvements for Violins and Other Stringed Instruments, Patent No. 9418, Great Britain: 4 May 1899. A detailed history of the Stroh violin in the early recording studios is given by Rabinovici (2010, p 47). N1 - The Victor Talking Machine Company replaced conventional violins and violas with Stroh instruments in 1905. Source: Instrumentation for Orchestra, correspondence between C G Child, Head of Artists’ Department at Victor and T Birnbaum, 1905 (EMI Archive, reprinted in Brock-Nannestad, 1997, p 28). Studio photographs from the later acoustic period also show Stroh violins and violas in use, if not exclusively, then to reinforce the upper string sections, see Batten (1936, pp 19 and 66). N1 - A small Berlin factory premises converted into a recording studio by the Deutsche Grammophon AG http://www.emil-berliner-studios.com/en/chronik2.html (accessed on 2 February 2015). N1 - These recordings were the first ever made by the Berliner Philharmoniker, source: http://www.berliner-philharmoniker.de/en/history/arthur-nikisch/#49 (accessed on 2 February 2015). N1 - It is not known if the BPO had another engagement on 14September 1913, or if the musicians’ contracts necessitated a day off, but a test pressing could conceivably have been made on this day. Although it was common practice to play back freshly recorded wax discs after recording for test purposes (rendering them useless as master discs for duplication because of wear from the playback stylus), the mechanical-acoustic playback of a soft wax material may not have reproduced much of the high frequency sound on the disc and a test pressing would have given a far better indication of the overall quality of the recording. N1 - Alfred Hertz (1872–1942), although an accomplished conductor, did not have the same level of fame and adulation accorded to Nikisch. N1 - William Sinkler Darby worked with Fred Gaisberg to make the very first recordings for the Gramophone Company, see: Northrop Moore (1999,p 45). N1 - See footnote 6 N1 - EMI Archives N1 - The exact number of stringed instruments is not possible to determine from the photograph. N1 - Documents in the EMI Archive dating from 1905 and 1907 give many details about the techniques of acoustic recording and are referred to throughout this article. N1 - For images of acoustic recording sessions, see Batten (1936, pp 19, 35, 66, 82 and 131). The website of the Centre for the History and Analysis of Recorded Music (CHARM) also has a few images: http://www.charm.rhul.ac.uk/history/p20_4_1.htmlhttp://www.charm.rhul.ac.uk/history/p20_4_1.htmlfor (accessed on 2 February 2015). N1 - Conductor Stanley Chapple has given a graphic account of the cramped conditions endured by the musicians in the acoustic recording studios along with the use of mirrors, re-instrumentation of the orchestra and other studio practices (Chapple, 1928, pp 290–92). N1 - In Instrumentation for Orchestra Child states that first and second violins and viola are replaced by Stroh instruments (Brock-Nannestad, 1997, p 28). This is backed-up by contemporary studio photographs. The three Stroh instruments used in this re-enactment are from the collection of Aleks Kolkowski and have been restored to a professional playing standard. See: http://www.exhibitionroad.com/9384/stroh-violin-exhibition-debuts-at-royal-college-of-music (accessed on 3 February 2015) N1 - An example is a photograph from 1925 at the Velvet Face Studio, London. The French horns play backwards with their bells towards the recording horn and a large mirror on the wall helps musicians otherwise unable to view the conductor. In Batten (1936 p 66) N1 - The discs were loaned by collector and writer Dominic Combe and were played on his HMV Senior Monarch gramophone with wooden horn. N1 - For example, thelimitation in frequency range from gramophone reproduction – an advantage as most of the high frequency noise on the recording is therefore filtered out. N1 - References to instrumental balancing through the making and playing back of ‘test waxes’ are found in Chapple, 1928, p 293; Harvith and Harvith, 1987, p 43; and Melville-Mason, 1977, p 97. Once played, the test records were rendered unfit for processing and duplication and would have be shaved for reuse. N1 - See section two N1 - It had been intended that the students would be coached in the performance style of the period – replicating elements such as the approach to tempo, rubato, portamento, and other performance traits – which can be heard on the Nikisch recording. However, there was not enough time in the project to allow for the musicians to get into the complexities of playing in an unfamiliar style. Hopefully it will be possible to attempt this in future experiments. N1 - See section three N1 - Concert pitch in Germany during the early twentieth century followed the French diapason normal of A=435Hz, while Britain adopted the ‘Philharmonic Society’ tuning standard of A=439Hz. The international standardisation of concert pitch as A=440Hz was established in 1939. N1 - See section two N1 - Recording experts made their own soundboxes and horns, which were guarded as trade secrets from competitors (Copeland, 2008, pp 254 and 262). N1 - Also known as the recorder sound box. In electrical disc recording, the recorder became known as the cutterhead. N1 - Nitrocellulose began to replace wax as a recording material for blank discs in the mid-1930s, although wax continued to be used by recording studios, particularly outside of the USA, until well into the 1940s. Robert Morrison recounts the inception of lacquer discs in the USA (2005, p 23). N1 - 'Example wax recipe: Carnauba 36%; Bees wax 12%; Aluminium Oleate 27%; Sodium Stearate 25%. All the ingredients are melted together in the order given and maintained at 120° C for half an hour with constant stirring….'(Bryson, 1935, p 104) N1 - Wax Making Document, Columbia Graphophone Company, 1931 (EMI Archive) N1 - When a recording stylus is placed on the surface of the rotating wax blank it cuts or effectively ploughs a shallow v-shaped groove on the surface of the wax. The intention is that the sides of the groove are perfectly smooth, the interaction of the cutting edge and the wax will however produce microscopic fractures which means that when the wax or finished record is played back, a level of inherent hiss is heard. As this particular source of noise is caused at the point of the cutter it is often termed 'cutter noise'. One odd advantage of the acoustic process is that cutter noise, if excessive, can be heard via the recording horn and a poor cutter or blank can be detected at the test stage. N1 - Signal-to-noise ratio is a measure that compares the strength or level of a desired signal relative to the level of background noise; in this case the sound intended to be recorded relative to the undesirable cutter noise. N1 - The wax used in the RCM re-enactment was tested off-site, using both electrical and acoustic cutters, and gave an acceptable cut with reasonable signal-to-noise ratio. It was not as good as the sample of the original wax but better than any other material produced in the experimental stage. The wax also cut better at an elevated temperature (28° C) during these tests and a warming cabinet was therefore constructed for the recording sessions to facilitate this. N1 - It should be noted that because of the time needed to prepare discs there was a limit to the number of experimental waxes that could be made. Upwards of forty experimental disc wax batches were tested prior to this wax being settled on. N1 - The inherence of noise to the acoustic recording process, due to the wax material and the action of the cutting stylus that ploughs through it, is discussed by Brock-Nannestad (1998, p 3). The metal negatives produced through the electroplating of wax masters were used to stamp new disc copies or ‘pressings’. N1 - In a traversing turntable system the recorder assembly is set in a fixed, stationary position and the rotating turntable is made to move laterally underneath the recorder, allowing for a spiral groove to be cut from the edge of the record towards the centre spindle. N1 - In systems where the recorder is made to move across the face of the record, it is guided from the edge of the record to its centre by a lead screw. As the recorder has the horn connected to it, they would both be carried from right to left thus altering the direction of the horn. N1 - The turntable rotation is belt-driven via a speed controlled DC motor. The turntable traverse is driven by a separate speed controlled geared DC motor giving a range of spiral groove spacings of between 60 to 200 grooves per inch. N1 - The term ‘pitch’ in this context denotes the number of grooves per inch. A record cut with ‘low pitch’ means that it has widely spaced grooves but therefore a shorter recording time on the disc; a higher pitch would have more grooves per inch and a longer recording time, but with the increased risk of grooves intercutting one another when loud or bass sounds are recorded (White, 1991, p 351). Recording lathes of the acoustic era had fixed pitches which were determined by the density of the thread (or number of threads) on the lead screw – nominally between 90–110 grooves per inch. N1 - A drawing of such a coupling with precise diameter measurements exists in the EMI Archive: Coupling for Orchestra Accompaniment (1907). N1 - A famous example is the Enrico Caruso and Nellie Melba duet O Soave Fanciulla, (Victor C-4326, 1907). The overall level is noticeably lower compared with other records of the period (with thanks to Sean Davies for this reference). N1 - A floor plan of a late-period orchestral acoustic recording session utilising two recording horns is shown in Chapple (1928, p 291). However, his instrumentation shows a small group of strings with a bassoon doubling the cello part and double basses substituted altogether by a tuba and contra bassoon. It would have therefore been unauthentic to follow this schema in our re-enactment of the 1913 BPO recording. N1 - The recording horns made for the RCM sessions are based on historic horns as measured from the EMI Archive. N1 - Seymour (1917, p 66) describes a suitable room for general-purpose recording (including for orchestra) as being up to twenty feet in length, up to twelve feet in height and width and with ‘first-class acoustical properties’ (these are not specified). N1 - ‘The recording horn should be larger than that used for vocals, and should be at least 36 or 42 inches in length, with a flare-opening of, say 18 or 20 inches. The exact dimensions are of little consequence in band work. The best shape is that of the double cone. It may be made of charcoal iron, stout sheet brass, or blocked tin, each of which metals contribute to the production of a good sonorous tone.’ (Seymour, 1917, p 64) N1 - The two horns of the same dimensions were designed to be used as a pair for experiments in multiple horn recording. However, time dictated that the recordings were made exclusively with a single horn, therefore only the larger straight and flared horns were tested during the re-enactment. N1 - To celebrate the centenary of the oldest company in the EMI group (HMV) being established in London in 1897, EMI created a project to produce an acoustic recording using the recording machine from the archive and pressing the resulting 78rpm record. The recording of Roberto Alagna (tenor) accompanied by Antonio Pappano (piano) was made at Abbey Road studios in 1997. Some 3,500copies of the record were issued as gifts to employees/associates of the company worldwide; it was not offered for public sale. N1 - Detailed drawings of type ‘R’ recorders are found in correspondence and documents from the Gramophone Company and Victor Talking Machine Company by William Sinkler Darby, Fred and William Gaisberg, 1907–8 (EMI Archive). N1 - Ibid N1 - The type ‘R’ recorder was used in the 1924 recording of George V (EMI Archive). N1 - A so-called ‘colouring’ of sound. N1 - The wax discs are shaved on one turntable (the shaving machine) and recorded on another (the lathe). The slightest amount of dust or variation in the way the blanks sit on the turntables will cause variation in the order of one or two thousandths of an inch. Add to this the fact that the blanks are warmed and not recorded on for several hours, then some minor distortion is possible. Normally a recorder support mechanism is sufficient to allow for this movement. However, the soft nature of the wax means that the cutter may preferentially cut deeper into the wax rather than raise the mass of the recorder and so create more cutter noise on the upward movement of the blank and a consequence, variable level of cutter noise generally on a once-per-revolution cycle, is referred to as 'sweep noise'. N1 - The trunnion, in acoustic disc recording technology, is a device that allows for free vertical movement of the recorder body whilst maintaining an airtight connection with the horn. N1 - Playing back the wax discs, even once with a lightweight pick-up, would cause wear resulting in a deterioration of sound quality, making the discs less suitable for processing. N1 - Fibre or bamboo gramophone needles are preferable N1 - See footnote 35 N1 - A comparison of two recordings of ‘Waltz of the Flowers’ from Tchaikowski’s Nutcracker Suite by the Coldstream Guards, conducted by Lt Dr J Makenzie-Rogan (HMV C240, 1912), and the Royal Albert Hall Orchestra, conducted by Landon Ronald (HMV D126, 1922), show the later orchestral recording to be much lower in level than the version for military band. N1 - Improvements that might have helped to increase the level on the recordings would have been: (i) To use a glass rather than mica diaphragm, as these are often more sensitive; (ii) To use a smaller room with a livelier acoustic (the RCM Recital Hall was far from dead, but wider than needed and with an acoustic tiled ceiling); (iii) To have purpose-built staging to create separate levels for the instruments so that they are all much closer to the recording horn and play over the top of each other. Getting the musicians even closer to the horn and getting them to play even louder would have undoubtedly raised the recording level by a considerable amount. N1 - Nine of the 32 musicians in the orchestra were interviewed before, during and after the process, and interviews were also undertaken with the recording engineer Duncan Miller, the conductor Robin O’Neill, and George Brock-Nannestad, independent observer and adviser to the project. N1 - Vivid accounts of early-twentieth century musicians’ experiences of recording are recounted by Harvith and Harvith (1987), Day, (2000), Philip (2004) and Katz (2004). N1 - As with any experimental project, there were difficulties to overcome. This unusual situation – the time constraints, the necessary added focus on the technical aspects and documenting, the lack of time to communicate adequately with everyone at each step in the process – can be an ostracising experience. However, it can be argued that it was an authentic situation, which mirrored how earlier recorded musicians felt: cramped studios, alien technology, men in white coats giving instructions, engineers pulling and pushing musicians to and fro in front of the horn, the wax and its fragile constitution sometimes seemingly as important as the humans in the room. Nevertheless, the musicians’ sense of well-being in the studio is something that was taken very seriously in this experiment. Whether in 1913 or now, communication and inclusion is vital in the studio – and is something which needs to be worked on and carefully fostered. Despite the challenges, the musicians showed great patience and professionalism throughout this very experimental project. N1 - These photographs illustrate the intense close-quarters conditions of the recording session, contrasting them with the comparatively spacious and comfortable conditions of the rehearsals, conditions to which musicians are more usually accustomed (with the exception of instances in which a large orchestra must fit on to a small stage or an opera house or theatre pit). N1 - P19 (Interviewees are referred to by their participant number, therefore Participant 1 is shortened to P1, and so on.) N1 - P1 N1 - P2, P21, P23 N1 - P6, P23 N1 - As a secondary point it can be considered that people often hear an old recording and judge it as old-fashioned and therefore not relevant. This re-enactment can help to question what is meant when people say that early recordings sound ‘old-fashioned’. Is it the artefacts of the technology – the hiss and crackle – that make the recordings sound old, or is it the performance style (the rubato, portamento, lack of vibrato in the strings), or a combination of both factors? This current experiment might allow investigation of this question in a way that has not yet been attempted. Daniel Leech-Wilkinson has written about some of the reasons why we might react to the old-fashioned sound of earlier recorded performances (Leech-Wilkinson, 2010 and 2006). N1 - This acoustic recording was undertaken by Patricia Hammond (singer), Aleks Kolkowski (Stroh violin), Neil Brand (piano) and Duncan Miller (sound engineer) as part of a documentary presented by Neil Brand. The Sound of Song, BBC 4, 16 January 2015. http://www.bbc.co.uk/programmes/b04yk00k– accessed on 3 February 2015. N1 - There is a story that is told about Toscanini feeling that Nikisch’s recordings did not live up to the power of his live performances, but the original source of this quote is not cited in the literature. Philip, R, ‘Historical Recordings of Orchestras’, in Lawson, Colin (ed), 2003, The Cambridge Companion to the Orchestra (Cambridge: Cambridge University Press), 205; and Tolansky, J, 2003, ‘International Case Studies’, 128 N1 - It was originally intended to make a modern digital recording of the orchestra’s twenty-first-century performance in normal orchestral positioning to act as a control example, but unfortunately there was no time left at the end of the project to do this. A good comparison can be made, however, by using takes from the first and second days of recording. N1 - Field-note: Robin O’Neill, conductor, 5 November 2014 N1 - Field-note: Robin O’Neill, conductor, 5 November 2014 N1 - P1, P6, P21 N1 - P23 N1 - P19 N1 - P21 N1 - P2, P23 N1 - P23 N1 - P19 N1 - P23 N1 - P23 N1 - P21 N1 - P2, P6, P21, P23 N1 - P23 N1 - P6, P19 N1 - P1, P19 N1 - P6, P23 N1 - P6 N1 - P1 N1 - P19 N1 - P21 N1 - P19 N1 - P6 N1 - Personal communication: Interview with Duncan Miller, sound engineer, The Art and Science of Acoustic Recording project (Royal College of Music), over the telephone, Thursday 22 January 2015. N1 - Personal comment: Interview with Robin O’Neill, conductor, Acoustic Recording Sessions, Royal College of Music, Thursday 6 November 2014. N1 - P23 N1 - P1, P2, P19 N1 - P23 N1 - P19 N1 - One question that arises is can this experience be modelled without painstakingly making acoustic recordings of everyone, which, even if desirable, would be prohibitively expensive? Perhaps it might be possible to distil some key principles that can be used educationally. If students were given the rules or sets of parameters that have been discussed here – about what acoustic recordings do and don’t capture, and about what elements of their playing musicians had to change in order to come across well on an acoustic recording – they could then apply these to other recordings, helping them to listen closely and carefully in order to hear through to the performances behind early recordings. N1 - P6 N1 - MUZE Classical Catalogue 2008, Gramophone Classical New Release Information Service (Muze Europe Ltd, 2008) PB - The Science Museum Group SN - 2054-5770 LA - eng DO - 10.15180/150302 UR - https://journal.sciencemuseum.ac.uk/article/acoustic-recording/ T2 - Science Museum Group Journal