by Rainer Maillard © 2023 

DG's first recording with the Boston Symphony Orchestra in January 1970 marked the entry into the new recording process of quadraphony and thus into analog multitrack technology. In the decade before the introduction of digital technology, DG produced in this way many legendary recordings. It is worth taking a look at the technology used, the working methods and the people involved.

Björn Blüthgen, Paul Meister and the development team prepare the 18-channel tube console for the first Quadro recording at Boston Symphony Hall, Hanover-Langenhagen January 1970. ©  EBS

Studio versus Live

Unlike today, all of these quadraphonic recordings in the 1970s were, without exception, studio productions - recorded without an audience, in empty concert halls. Today, mostly live orchestra recordings are released. These live recordings, however, go through a sophisticated post-production process to meet all the required sonic and musical demands. Today, the term “live recording” does not refer to an un-edited version of a single performance, but to a product that is a compilation of usually several concerts, rehearsals and even a follow-up “patching” session.

Live recordings place higher demands on recording technology. The difference in acoustics between a full hall vs. an empty one must be compensated for, and audience noises need to be removed. In the 1970s, analog editing technology with scissors/razor blades and glue posed so many challenges that detailed editing of live recordings was simply unthinkable. With today's digital recording technology, it is entirely possible to overcome these challenges. Musicians now have the option of recording "live" as opposed to "in the studio".

With studio recordings there are many possible solutions to balance issues not available to someone recording a live concert. The orchestra setup can be altered to be ideal for the microphones rather than for the audience. For example, the harp can be placed in front of the first violins, directly by the conductor, so that the main microphone captures its sound clearly. This reduces the need for additional spot microphones; consequently, the miking can be more purist.

When it comes to recording, less can be more. For example, the acoustics of an empty hall are usually quite satisfactory for an appropriate sound image, without an audience that reduces the reverberation time of the concert hall. Should the acoustics still turn out to be too dry, rows of chairs can be removed to get more reverberation. Today, this is simply done through digital sound processing. The lack of audience noise, in turn, allows the audio engineer to set up fewer microphones at a greater distance from the instrument groups; in any case, the mixing consoles in the 70s had fewer channels.

Another very crucial aspect is that the sound image created during the recording was final. Adjusting the balance between individual instruments after the session was impossible.

Any sonic comparison between historical and modern recordings is therefore multi-layered. Both the techniques and the equipment used are different depending on whether the recording is generated from “live” concert or studio sessions.

Recording setup for Hector Berlioz' "La Damnation de Faust" with Seiji Ozawa and the Boston Symphony Orchestra. Boston Symphony Hall, 1973. © DG

Microphones

It is interesting to take a look at the microphones used in this period. Since the 1970s, the technical characteristics of microphones in general have evolved. Yet, the types used at that time already had a very high tonal quality. Tube condenser microphones from Neumann (M 49, U67, SM2, KM 53, KM 54, KM 56) and Schoeps (M221B) were used for the recordings. The switch to phantom power fed condenser microphones (Neumann U87, KM 84, SM69) and the Schoeps Colette series (MK2, MK5) occurred around 1974. The sonic differences between these and  modern microphones is smaller than the time span of 50 years would suggest. It is therefore not surprising that even tube microphones, such as the Neumann M49 and U67 types, are still in demand and being manufactured today.

Like many things in recording studio technology, it is not the quality that has changed, but rather the operational reliability, convenience, size, price and quantity. From a measurement standpoint, today's microphones might be more linear in frequency response and lower in noise, but ultimately it is the sonic impression of the finished recording by which any final judgment must be made. The amount of noise perceived by listeners usually stems from completely different factors: air conditioning, outside noise, heating and, in the case of live recordings, the audience. These sources of noise cannot be read off any data sheet. If they are kept to a minimum, as the sound engineers of the time did with great care, the recordings sound extraordinarily clean.

Günter Hermanns ^[1] in the control room of Symphony Hall Boston, February 1971. © EBS

Mixing consoles: tube versus transistor

The mixing consoles used at that time had a smaller number of microphone inputs. To this day, apart from the general switch to digital technology, the number of channels has increased rather than the quality of the individual components. If fewer microphones are used, they usually have to be placed at a somewhat greater distance from the instruments. Oftentimes, this is actually appealing in terms of sound. However, it takes away the possibility to manipulate each instrument or voice individually via the mixing console.

The change from tube to transistor technology happened even earlier in mixing console construction than it did with other equipment. The key factor here was not so much the sound quality either, but rather the requirement of 4-track mixing consoles for the newly introduced quadraphonic technology. The first quadraphonic recordings were still produced with the 18-channel stereo tube mixer developed by Deutsche Grammophon in 1960. It was replaced in 1971 by a 23-channel transistor mixing console, which had been specially designed for quadrophonic recording. They called it the "Boston console" because it was first used in Symphony Hall there. In 1974, the PolyGram Module Mixer followed. This joint development of DG and Philips at PolyGram was built in Baarn. It was also designed for quadraphony, but the number of channels could be expanded due to its modular construction. This console was in use until the introduction of digital mixing consoles in 1992. The mixing consoles had similar functions to modern analog mixing consoles and contributed significantly to the outstanding sound quality of the recordings made at that time.

Mobile versus studio
Recordings of classical orchestral works require concert halls with suitable acoustics. Since digital reverberation systems – like the ones that are being used today - were not available at that time, the original acoustics of the concert halls played a greater role than they do today. Therefore it's no wonder that most recordings of the time were also made in halls with good acoustics. Many DG  recordings were made in three legendary halls: Munich's Herkulessaal, Boston's Symphony Hall, and Berlin's Jesus Christ Church. DG set up permanent control rooms in these venues. For recordings in other halls, mobile equipment was used, which in principle was not much different from the permanent installations. The choice of the Jesus Christ Church is particularly interesting, since it was chosen for recordings with the Berlin Philharmonic Orchestra, despite the high logistical effort involved in recording there, and even though the Philharmonic Hall (the Berlin Philharmonie) was also a very good option. The acoustics of the church have often been praised. The often-mentioned explanation that the relatively short reverberation time in the low frequencies was a deciding factor in producing good sound, is only half of the story, because this "deficiency" in the low frequencies was often compensated for by the audio engineers with the help of an equalizer. The fact that the Church acoustics allowed for a very transparent sound image and thus the possibility to record large orchestras with very few microphones seems more relevant.

Mixing
Quadraphony required doubling the mixer outputs, but today's practice of recording each microphone on a separate track was not possible at the time. Mixing was done directly in 4-track technology. This had a major impact on the production process. Since the recorded format was a finished quadraphonic mix, it was impossible to make any changes afterwards. Accordingly, the conductor and orchestra musicians listened back to the recording takes during production, the balance was double-checked, changes were discussed, then the music was rehearsed again and re-recorded if necessary. If, for example, one wanted to make the secondary part of the second oboe more prominent, this had to happen "in front of the microphone". Today's attitude of "we'll fix it in the mix" did not exist. However, whether you raise the level of the second oboe in post-production, or whether the oboist plays louder during the actual recording, the end result will sound different. It may turn out that it is not the oboe that should play louder, but rather that the other musicians should play softer. Relying less on post-production to adjust balances etc. might actually be why recordings from that time have such a big advantage. 

Tonmeister Hans-Peter Schweigmann ^[4], recording producer Werner Mayer ^[5] and Edith Mathis at the new PolyGram module mixer during a recording session in Munich 1974. ©EBS

Recording
Quadraphony required tape machines with four separate audio tracks. DG used Telefunken M15 and Studer A-80, as well as Ampex machines for their Boston recordings. However, the different types of magnetic tape had a greater influence on the sound. The dynamic range of audio tapes is smaller than that of today's digital recorders. Tape hiss can be irritating at low program levels, while saturation effects and distortion can occur at higher levels. The art of the audio engineers was to make the technical limitations as inaudible as possible through exact calibration for each type of tape, and by finding ideal level control. The characteristics of the human ear, which is not a linear system in terms of frequency response and sensitivity, were also taken into account. From a measurement perspective, magnetic tape performs worse than today's digital recorders, but considering what our ears actually perceive, the difference is smaller than the measured values would suggest.

Noise reduction
With the decision to produce quadraphonic masters, the question of how to continue serving the stereo format arose. It would have been possible to record a stereo mix next to the 4-track tape. However, both the 4-track and the 2-track tape needed to be edited later, which would have resulted in a significant amount of extra work. Creating a stereo downmix from an edited 4-track master seemed to be more practical. The downside however, is that an additional tape generation is created, which would not be necessary with a “direct to 2-track tape”. However, each tape generation means additional noise that had to be kept as low as possible. That is why all 4-track tape machines were outfitted with the Dolby noise reduction system. This noise reduction system makes clever use of the human ear's perception of sound. At moderate and high levels, the music itself covers any tape noise, but not during quiet passages. The Dolby system thereby keeps the program level as high as possible above the noise floor, depending on the frequency as well. Accordingly, the Dolby process amplifies the recording level during quiet passages, but not during loud ones. In this way, the dynamics of the program on tape are reduced significantly. When playing back the tape, quiet passages that were boosted during recording are lowered to regain the original dynamics. In this process, the noise floor is lowered as well during quiet passages and thus falls below the perception threshold. Each technique has its own advantages and disadvantages, and so does the Dolby system. One possible disadvantage is that errors in dynamics and transient response can occur if the calibration is not exact. For this reason, the Dolby system was partially abandoned later, when even lower noise tapes became available.

Recording engineer Jürgen Bulgrin, tonmeister Klaus Scheibe, Eugen Jochum and recording producer Günther Breest ^[8].  © EBS

Editing
At the time of the Quadrophonic recordings, editing the tape using blades and glue was understandably limited in its possibilities. The technique involved splicing the tape in the appropriate places after selecting the best recording takes and re-assembling them. However, the likelihood for an analog tape edit to be totally "inaudible" was not always high. An edit might not be successful at any point in the music. The levels of both takes had to match perfectly, otherwise sudden changes in loudness would occur and the edit would become audible. Recording producers and audio engineers at that time developed a good intuition during the recording sessions to determine where edits might work well.

Today, one often hears conductors and musicians say that they prefer live recordings to studio recordings. To be honest, however, one should not forget about the fact that conductors and musicians today know exactly that the possibilities of editing in the digital domain are almost unlimited. Nowadays, the term "live recording" usually means that these releases are a collage of several concerts, rehearsals and retake sessions. Thanks to modern digital technology, edits can now be made at almost any point, even if the level, tempo, pitch, phrasing, and acoustics between takes are not homogenous. There is no chance of doing this with analog tape. With the "old" studio recordings, the conductors and musicians knew that edits might not be possible at every point. Accordingly, the tension was greater during studio recordings as well.

Karajan versus Kleiber
Ever since the introduction of magnetic tape, editing could be used as a tool to “design” the perfect (or near perfect) recording. Unlike in film, where the picture cut always remains visible to the viewer, the edits in audio must be inaudible and appear as if the artists are playing in one take. Any audible editing would irritate the listener and be perceived as a mistake.

On the one hand, audio edits are intended to satisfy the demand for creating a perfect performance. Unlike a concert, a recording can be reproduced over many decades. No one wants their legacy to contain small mistakes that might hardly be noticeable in a concert. On the other hand, a studio production offers the possibility to interrupt the process at any time, check the quality of what has already been recorded, and eventually to continue "fine-tuning" the performance. There are certainly differences between performers’ work methods. This is reflected, for example, in the amount of splices visible on the original tapes. Herbert von Karajan's legendary recording of Death and Transfiguration and Four Last Songs by Richard Strauss contains a total of "only" 13 edits. However, in the case of the symphonic poem, a full four, non-contiguous recording days are noted in the recording protocol. The type and manner of the recording tapes’ batch numbers and splicing material indicate a longer production period. Karajan probably re-recorded entire passages and had them inserted into the existing master. Karajan could afford the luxury of repeating the recording of a work until the result was satisfactory. The same goes for Kleiber's recording of Beethoven's 7th Symphony. Recording takes from an additional recording session were inserted after the editing had already been completed. However, this is yet another completely different way of working. In this case, the edits show how Kleiber must have struggled to find the "correct" solution in some places. More than 120 splices are visible throughout the tape. Interestingly, these are concentrated in certain specific passages. Many of these splices merely represent unsuccessful attempts to make an edit. A tape, once cut, remains separated. If you want to undo the edit because the result is not yet perfect, you also have to glue the tape together again and the splice will remain visible. The fact that many splices were rehearsed until they hit the perfect spot shows Kleiber's uncompromising approach to his recordings.

The people behind the scenes
As interesting as the consideration of all the technical devices might be, the manner of how they’re being used within the recording chain is even more crucial for quality. This is where the human factor of shaping the sound comes into the equation. Just as the sound between one orchestra and another might vary, recordings made by different recording producers or engineers do sound different if you start listening closely. Even recordings from different labels - independent of the people involved - show differences. Therefore, it is absolutely worth looking at how recording teams were organized at Deutsche Grammophon Gesellschaft in the 1970s.

As a result of global political and company-specific decisions by the three leading business personalities Joseph Berliner, Bruno Borchardt and Ernst von Siemens, the members of the recording teams found themselves based in two locations and in different organizational units around 1970. DG employed the musical recording producers under its then president Kurt Kinkele. The Hamburg-based operation dealt with all artistic matters concerning conductors, orchestras and soloists, plus the advertising and distribution of the recordings. The tonmeisters and audio engineers, on the other hand, operated out of Hanover and reported to Peter Burkowitz, the technical executive director of the group that formed PolyGram. In Hanover, the recording equipment was put together and shipped around the world, the original tapes were edited and mixed after the recording sessions, and the records were pressed.

This company-specific split had its consequences. The artistic recording producer and the tonmeister were able to work together as equals and as a team, since neither one was subordinate to the other in the company hierarchy. At the same time, DG’s Hamburg branch concentrated on forming long-term artistic relationships, whereas the Hanover branch focused more on technical developments. This is the only way to explain why a project like “Quadrophony” could be pursued for such a long time, without rendering concrete products for the consumer market. For many years, Peter Burkowitz -  being a recording engineer himself, as well as an inventor and designer of mixing consoles - was able to convince his superiors of the importance of investing time, money and talent in the recording medium’s future.  Kinkele, on the other hand, was able to nurture long-term relationships and collaborations with the greatest performers, and did not have to deal with technical issues.

In the older DG tradition the recording producers, such as Hans Weber and Rainer Brock, had been trained as conductors, and the audio engineers, such as Günter Hermanns and Hans-Peter Schweigmann, had been trained in electrical engineering.  In the 1970s this was now subject to change. Younger tonmeisters, trained at the Detmold Academy of Music, were recruited. The world’s first tonmeister course had been established in Detmold after World War II by Erich Thienhaus, a recording producer and engineer at Deutsche Grammophon.  Generations of tonmeisters were trained there, such as Heinz Wildhagen, Klaus Scheibe, Günther Breest, Klaus Hiemann, Karl-August Naegler and Wolfgang Mitlehner, who later worked both in Hamburg as recording producers and in Hanover as tonmeisters.

Following a suggestion by Peter Burkowitz, Kurt Kinkele signed an exclusive recording contract with the Boston Symphony Orchestra in 1969. New York based producer Thomas Mowrey was hired to support these productions. Mowrey had already come into contact with quadraphony during his own training. Therefore it is not surprising that the first quadophonic recording by DG took place in Boston.

Gernot Westhäuser, Joachim Niss and recording producer Thomas Mowrey ^[13] in Boston. © DG

Quadraphony then and now
DG's decision to record in the quadraphonic format from 1970 onwards was in every respect an investment in the future, because the question of how the consumer would be able to hear the product at home was initially completely unanswered. Different quadrophonic systems were competing with each other, each with its own amplifiers and decoders for home entertainment, each with its own advantages and disadvantages. For reasons of sound quality, Deutsche Grammophon relied on the CD4 format developed by JVC until 1974. It offered - as a major advantage over other systems - four discrete channels. Nevertheless, the challenges were immense. In order to be able to cut the signals for four independent channels in a single groove of the record and later be able to reproduce them, the rear channels needed to be transferred into the inaudible range above 20,000 Hz by means of frequency modulation - similar to FM radio. This range, however, can only be used to a limited degree with records and leads to a reduction in quality for the rear channels. LPs produced with CD4 are not stereo-compatible either. If you play them on a regular record player without a CD4 amplifier, only the front channels will be tracked, which makes the sound much too dry overall. In the end, the technical difficulties were too great and too many, and DG refrained from a market launch.

This did not mean the end of 4-track technology in recording - quite the opposite. On the one hand, engineers were convinced of the sonic advantages, hoped for future solutions for the consumer, and continued to create quadraphonic masters. On the other hand, multitrack technology offered the advantage of being able to change the balance later. This aspect actually led to an expansion to eight tracks as early as 1972. Here, too - in accordance with the DG philosophy - the 8-track tape was spliced to create an 8-track master (“Original”). Both 4-track quadrophonic masters and stereo versions were then created from these originals/masters.

With the advent of the CD, the question of which source to use arose. In some cases, the 2-track downmixes were used, but it quickly became clear that re-mastering the 4-, 8- and 16-track originals led to much better results. Remixes on CD, for example, were made for the Original Series of CD reissues, but only in 16bit and 44.1 kHz (at that time).

Around the year 2000, the SACD offered DG the chance to release the quadraphonic masters in the 5.1 surround format. SACDs are still quite popular, particularly in Japan. Emil Berliner Studios has been producing high-resolution remasterings in 192 kHz/24 bit by explicitly using the originals as a source. Today, the immersive format Dolby Atmos represents yet another possibility to showcase these recordings. Emil Berliner Studios also creates high-resolution stereo versions for all Dolby Atmos projects that are based on analog sources.

DG Managing Director Hans-Werner Steinhausen ^[14],  PPI Managing Director Coen Solleveld and DG President Kurt Kinkele in conversation. ©private

The Original Source Series

50 years of dealing with the analog multitrack archives led to this very special vinyl project, in which the stereo LPs are cut directly (with no intermediate step, neither analogue nor digital) from the multitrack originals for the first time in history.