Go back to article: 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
Historical context: orchestral records – a step towards phonographic realism
Arthur Nikisch and the BPO’s recording of Beethoven’s Fifth Symphony was published in Germany by the Deutsche Grammophon AG in February 1914 as a set of four, doubled-sided, twelve-inch discs. The Gramophone Company released it in Britain as a series of eight singled-sided twelve-inch discs, each movement on two discs being issued separately over a period of eight months, from January until August 1914, beginning with the Andante movement. The complete set of discs sold for a total of £2 in Britain at a time when ‘the average weekly wage was £1 6s. 8d. (or £1.33)’ (Day, 2000, p 7). These expensive editions were aimed at wealthy and discerning record collectors in order to satisfy a growing demand for orchestral recordings of classical music that were not truncated versions or arrangements for military band (Arnold, 1997, pp 15–19). Such a collector would also have owned a superior quality gramophone, capable of reproducing a twelve-inch orchestral record without distortion. The records were reissued in Britain as four double-sided twelve-inch discs in 1922.
© Aleks Kolkowski
Record Label of 1914 Pressing of Beethoven’s 5th Symphony
Beethoven’s Fifth Symphony – Allegro (Side One) Arthur Nikisch & the Berliner Philharmonic Orchestra
The greatest selling-point of these records was not the complete Beethoven symphony in C minor, but the ‘combination of artists equal to any in the world under a conductor of such world-wide reputation’. This was indeed a significant event in the history of recorded music: 'as the first conductor of high eminence to work before the recording horn, Nikisch bequeathed a distinguished endorsement of the phonograph as a medium for symphonic music' (Gelatt, 1977, p 183). By 1913, Nikisch already had an illustrious international career as a conductor, but his sole contribution to the record catalogues had been as piano accompanist to the mezzo-soprano Elena Gerhardt, making recordings of German Lieder in 1907 and 1911. On 18 June 1913, he signed an exclusive five-year contract with the Gramophone Company, with an agreement to conduct three recording sessions per annum. After the much-publicised Beethoven's Fifth and a clutch of overture recordings with the London Symphony Orchestra the following year in June 1914, Nikisch’s recording career with the Gramophone Company was dramatically cut short by the outbreak of the First World War.
Nicola Perscheid’s portrait of Arthur Nikisch, c. 1905
The Gramophone Company had sought to engage Nikisch to conduct an orchestral recording in 1911. A letter from Leo B Cohn, director of Deutsche Grammophon to his British colleagues, in which he discusses a proposed recording with Nikisch and Gerhardt, also reveals the Gramophone Company’s plans for a far more ambitious project:
You will understand that Nikisch has a very great name all over the world, especially in Germany, and I am sure, that it will help the sale of the records, if we can say that Nikisch has [plaid] the accompaniment. If you tell Nikisch what you want, I am sure that he will play the piano just the way we want it. I do not think, you will make any records with orchestral accompaniment because, Miss Gerhardt is singing 'Lieder' without exception.
It is a splendid idea, which I have considered long ago frequently, if we could make records of a large Symphony Orchestra with Nikisch as conductor. It is a big scheme and a great talking point. I think you can get Nikisch very cheap and anyhow, it is worth our while to find out, what Nikisch wants.
The letter shows the high regard in which Nikisch was held. The bonus of getting the great conductor 'very cheap' may well have been a deciding factor and one should not overlook the importance of financial concerns in the decision making and choices of repertoire and recording artists for what was, essentially, a large corporation with commercial interests. What is more telling, however, is the remark that if instructed, Nikisch 'will play the piano just the way we want it'. This would suggest the importance of a musician’s complicity in the process of acoustic recording; a studio ‘expert’ would dictate the manner in which musicians should play or sing, in order that they make a recording that is suitable for reproduction on a gramophone, never mind their artistic status. This had nothing to do with the musical taste or judgement of the studio bosses, but was a matter of necessity.
In order to achieve a proper balance on the recording (in this case, of voice and piano) and to maintain a constant audible level throughout, the playing style might have to be radically altered – for example, the piano is played louder, perhaps in a more precisely articulated or exaggerated way than when playing in a concert recital. The singer would also have to obey technical directions, directing her voice into the mouth of the horn, turning slightly or stepping back during loud passages so as not to overload the sensitive recording diaphragm and so ruin the recording.
An orchestra in the acoustic studio presented an additional set of problems. It was not possible simply to point a horn at a large group of musicians and expect them all to be registered on the recording; their positioning required very careful planning. Quieter instruments had to be placed close to the horn mouth, the louder ones further away or to the side; some had to be substituted because of lack of space in front of the horn mouth. This was especially true of stringed instruments, their open sound being difficult to capture by the acoustic process. Cellos were substituted or reinforced by a bassoon or bass clarinet, double basses by a tuba or contrabassoon.
A solution for recording the upper strings was found in the mechanically amplified ‘Stroh’ violins and violas – a radical re-design of the violin by Augustus Stroh using the sound reproducing technology of the phonograph. Characterised by its large aluminium horn, the Stroh violin is a highly directional instrument and could simply be pointed towards the recording horn. The first and second violins and violas would be reinforced or substituted entirely by two Stroh violins and a Stroh viola. This became standard practice by 1905 and continued until the end of the acoustic era in 1925.
© Robert Baumbach
Rosario Bourdon, the Victor Company musical director, conducting a studio orchestra c. 1920 - 1925
In this re-modelling of music for the gramophone, the orchestral rendition becomes an artifice. Recordings of the acoustic era are an impression, rather than a phonographic representation. But the recordings made by the Berlin Philharmonic Orchestra in 1913, I will argue, represent a step towards greater realism in sound recording and reproduction. Recorded music, formerly seen as a novelty without serious artistic value, was gaining critical acceptance and as recording and reproduction techniques improved, with them came the desire to hear faithful representations of the classical repertoire, as interpreted by great artists of the era.
Arthur Nikisch’s first orchestral recording was of Beethoven’s Egmont Overture, with the London Symphony Orchestra on 28 June 1913, at the Gramophone Company headquarters in Hayes. It is a standard orchestral recording of the period in which one clearly hears a tuba substituting the basses in the loud passages and the orchestra is audibly reduced. The Stroh violins are very prominent, resulting in the lack of a coherent ensemble string sound. This session may well have been a warming-up exercise, to initiate Nikisch into the trials of recording an orchestra in the studio, ahead of the ambitious Beethoven’s Fifth recording that was to take place five months later. In September 1913, in Berlin, an equally ambitious recording project took place in the same studio space where Nikisch would later record the Beethoven symphony. The BPO, conducted by Alfred Hertz, recorded excerpts from Wagner’s Parsifal on eight sides that were published as twelve-inch records. The exercise took four days to complete, from 12–16 September, with a day’s break in-between, possibly allowing for a test pressing to be made. In contrast, for the Fifth Symphony Nikisch would record the entire symphony with the same orchestra, also on eight sides, in just one day on 10 November 1913.
The extra investment in time, studio resources and, presumably, musicians’ fees on the extended Parsifal sessions would suggest that much effort and experimentation took place in the Berlin studio until satisfactory results were achieved. While these recordings of a large Wagnerian orchestra are an outstanding achievement in their own right, they may well have served as a testing ground in preparation for the session with Nikisch in November 1913. The feat of recording an entire symphony with Nikisch at the helm would have been a far greater prize in terms of prospective record sales and prestige. The recording expert for both the September and November sessions was Max Hampe, who had gained considerable experience of acoustic recording as an assistant to recording pioneer William Sinkler Darby.
Aside from listening to the recordings there is some additional evidence confirming that neither the September nor November BPO recordings were re-orchestrated (i.e. the musical score adapted and instruments substituted). Firstly, a German record brochure advertisement for Beethoven’s Fifth Symphony lists the full instrumentation as specified by the composer, and is therefore promoting it as an authentic orchestral recording. Secondly, there is a portrait photograph from September 1913 of Alfred Hertz and the BPO in the studio, taken at the recording of Parsifal. The picture shows a fifty-two-piece orchestra, unusually large for the studios of the time, with approximately sixteen violins and violas, at least two cellos and three double basses. The customary Stroh instruments are out of view, hidden in the back row of the string sections, a lone horn from a Stroh violin betraying their use and evidence that they would have reinforced the conventional violins and violas rather than dominate them. The orchestra for the recording of the Fifth Symphony would have been a smaller size than for Parsifal, but there is no reason to doubt the advertised authentic instrumentation. Max Hampe and his studio assistants had managed to capture the sounds of both a Wagnerian and a classical orchestra, without re-orchestration and for commercial release, perhaps for the first time in recording history.
© EMI Archive Trust
The Berlin Philharmonic Orchestra directed by Alfred Hertz, in the Berlin studio of Deutsche Grammophon, recording excerpts from Wagner’s Parsifal, September, 1913
© EMI Archive Trust
The close-up on the right clearly shows the horn of a Stroh violin (or viola) and the neck and scroll of the instrument
Historical basis for the staging of the recording re-enactment
The techniques of recording smaller ensembles and soloists have been fairly well documented in studio documents of the period. However, very little is known of how a large orchestra might have been recorded acoustically.
We are left to rely on contemporary photographic evidence to guide us in arranging the musicians in front of the recording horns. While some photographs show the musicians facing towards the camera rather than the recording horn, their positions remain basically the same as during recording, although players at the front might be placed far closer to the mouth of the horn for the recording than pictured. A common arrangement was to place the stringed instruments (and soloists) at the front, facing towards the horn mouth; woodwinds playing across the face of the horn, while the brass is situated at the very back. Timpani, if used, are placed outside the sector of the horn to prevent overloading the soundbox diaphragm during loud passages. In most cases, the brass and woodwinds would be seated on specially constructed platforms as this would enable the sound of each instrument to have a clear trajectory towards the mouth of the recording horn. Space was at a premium and musicians were packed very closely together to stay within the horn’s recording range. Music stands gave way to sheet music suspended by string from the ceiling using clothes pegs and some studios installed special rigging on the ceiling for this purpose. The conductor would stand elevated to the side of the recording horn; the room, even for an orchestral session, would be of a small size so as to contain the sound and reflect it back into the mouth of the horn. Because the musicians had to play louder than in a concert hall and very much closer together, the sound in these studios would have been intense.
Positioning of the orchestra for the re-enactment
With no other existing sources of information about the November 1913 session available, the photograph of Alfred Hertz and the BPO served as a starting point for the RCM recording re-enactment. Although they are posing for the camera, the general positions of the musicians seem correct for recording, except that the string players would be further forward, the first row being right in front of the recording horn (its mouth appears just to the right of the picture). Hertz is shown standing on a small platform in front of the horn, but this space would certainly have been occupied by the violins and the conductor’s real position is out of shot, towards the narrow end of the horn. For the re-enactment at the RCM, the same arrangement was followed except that the cellos and double basses were moved close to the front of the horn, the instruments facing just underneath the horn mouth.
© Aleks Kolkowski
The RCM Chamber Orchestra recording Beethoven’s Fifth Symphony acoustically on wax discs, No.1: cellos and basses
Two Stroh violins and a Stroh viola were used along with their conventional counterparts, as specified in studio memoranda. They were initially positioned at the very back of the string sections on a riser, as suggested by the Stroh horn in the 1913 photograph of the BPO, but because of the general faintness of the violins and violas on our test recordings, they were moved forward, taking a central position in among the other string players and closer to the mouth of the horn. The improvement in the overall balance of orchestral forces and the added presence of the violins and violas on the recordings was especially noticeable in the quieter passages of the Andante movement. The stage of the RCM’s Recital Hall was used for the brass and timpani; the two French horn players had to be seated backwards so their bells faced towards the recording horn, and a mirror was installed so that they could watch the conductor’s movements, as was standard practice in the acoustic recording studios.
© Aleks Kolkowski
RCM Chamber Orchestra, No.2: Stroh instruments in use (initial positions)
© Aleks Kolkowski
RCM Chamber Orchestra, No.3: woodwinds positioned to the side of the recording horn
© Aleks Kolkowski
RCM Chamber Orchestra, No.4: French horn player views conductor via iPad
Preparations and test recordings
The RCM sessions began by playing the original 78rpm disc pressings of the Beethoven symphony from 1914 on a horn gramophone of the period. As well as demonstrating mechanically reproduced sound, the playback served as a reminder that these records (and the ones we were about to make) were designed to be played on such technology with all its inherent limitations and advantages. The original pressings were also recorded as digital files via a separate turntable and these were used for comparison purposes, to gauge the level of the newly recorded wax discs against the original recordings and to compare the overall balance of instrumental sections.
© Aleks Kolkowski
HMV Senior Monarch Gramophone c. 1910, used to play back the original Nikisch/BPO 78rpm discs
Once the physical arrangement of musicians had been established in front of the recording horn and the music sections rehearsed in these unfamiliar positions, selected music passages were recorded on to the wax discs for test purposes. Recording tests were made on the first day until we were able to establish the final positions for the orchestra and an acceptable level of amplitude on the recording. The making of test records was very much a part of the process during the acoustic era also, as there was no means of monitoring the recording other than by listening through the horn with its narrow end placed in the ear. While this might give an indication of the sound reaching the recorder diaphragm, it cannot predict the quality of the recording and is impossible to do during the act of cutting a record. By making numerous test recordings, we were following a customary mode of procedure established by the early acoustic studios (Chapple, 1928, p 293; Harvith and Harvith, 1987, p 43; Melville-Mason, 1977, p 97). The second day was devoted to recording the two movements on four sides as well as making additional versions for a later electroplating and duplicating process.
© Aleks Kolkowski
A freshly recorded wax disc plays on the turntable used for transcription
A note on interpretation
Because the focus of the re-enactment was on the technical aspects and the playing experience, no attempt was made to follow Nikisch’s interpretation of the symphony. Instead, conductor Robin O’Neill led his own interpretation of the symphony. Only the first and second movements, the Allegro con brio and Andante con moto, were recorded for the RCM re-enactment. Likewise, while the BPO followed the contemporary German tuning system of A=435Hz, the RCM Orchestra was tuned to a modern-day A=440Hz. Nor were period instruments used or stringed instruments strung with gut strings. While the RCM re-enactment was therefore not an authentic reconstruction in terms of musical style, it did allow us to hear how a modern orchestra would sound when recorded using the acoustic process.
The RCM acoustic recordings followed exactly the same side-breaks as the original discs. Nikisch’s version of the first movement is taken at an unusually slow tempo, and omits the repeat of the exposition. Had the repeat been taken at this tempo, the first part of the Allegro would not have fitted on the record side. The first side of his Andante is also unusually long and at five minutes it barely fits on a side of a twelve-inch record at 78rpm. The centre label of this side is half an inch smaller in diameter than is standard and the music stops right at the end groove. In conductor Robin O’Neill’s interpretation of the Allegro, the exposition is repeated, except in a shorter version of the movement, which was recorded for electroplating purposes, to enable the moulding of a ten-inch resin record playable on a gramophone.
Unlike the original Nikisch recording where the entire symphony was recorded in one day, the re-enactment took two days to record two of its movements. The first day was almost entirely taken up with establishing the orchestra’s positions, adjusting the recording equipment and making sound tests to establish the optimal recording set-up. However, long periods of experimentation during recording is entirely commensurate with the studio practices of the acoustic era where recording tests were being performed on a daily basis (Batten, 1956, p 34). The second day was devoted to recording each movement on two wax discs.
© Aleks Kolkowski
RCM Chamber Orchestra with Robin O’Neill
Component DOI: http://dx.doi.org/10.15180/150302/003