story of a long, careful record production
Whenever we make a music recording, we deal in virtual reality. Sometimes we can imitate a real acoustic space, such as a concert hall. But, whatever we do, listeners will relate it to their own physical surroundings (if we do it well). Any music recording is an evocation of something which doesn’t even have to exist in the physical world. A live recording of a group or an orchestra still has to exist in its own right and on its own terms in a different environment. It’s never the sound that you heard in the performance. It evokes it.
Once we establish that a recording is a different thing, there are alternative ways of assembling the sounds of many people and instruments. All fall logistically between two extremes. In approaching a project which is to end up as a grand-sounding CD, you will always choose a production method at some point on the sliding scale between them.
One end of the scale is the simultaneous recording, such as of a symphony orchestra or a string quartet. Many jazz recordings are done complete in one take, without adding anything later, because of the need for interactive spontaneity and the difficulty for musicians in synchronizing with a previous performance where visual clues and spontaneous collective musical decisions are absent. Obviously, a live recording of a band falls into this category, although some changes are often made later, typically substituting a more considered studio vocal performance for a live lead track.
At the other extreme is the ‘layered’ production typical of electronica but which is also used widely in contemporary commercial pop music production, from hip hop to American Idol manufacturing. An eventual record might combine electronic drums, bass, guitars, synthesizer, vocals and other any other instrument, but each line is generally created and recorded at a different time from the others, and often modified as subsequent layers throw the music into a new perspective. This approach has only been possible since the early seventies, when multitrack tape machines with 16 parallel recording tracks were introduced.
When making a recording to project an artist in the strongest possible way, a producer will choose how much ensemble and how much layer might give the best, most convincing musical results. What matters, ultimately, is the sound from the speakers, which must project an illusion of simultaneous performance. That is how the listener hears it. Production of a rock band would lie in about the middle: the rhythm track is typically laid at once, to capture that hard-to-define ‘lock’ that musicians achieve when they start to breathe as one. On this foundation, vocals and other embellishments will be overdubbed, with few practical limits.
Like any large-scale work, the Universe Symphony presents considerable musical-technical difficulties. Those inherent in a large-scale Mahler symphony are surmountable within the concert hall, which is where the composer designed it to be realized. The implications of Ives’ radical piece, and the further inferences by Johnny Reinhard from the composer’s original notes, raise the stakes considerably. On several occasions, he maintains, Ives desired practical realizations which could not be undertaken in his time (even speculating about performing piano works on an instrument in non-equal tuning (whole-number ratio), impossible on the standard keyboard). The première of his new realization at Lincoln Center in 1996 was a notable achievement, but to raise the technical performance standard to be acceptable in repeated playback from a CD is next to impossible in live, simultaneous performance.
Johnny is a conductor, composer and bassoon virtuoso with a strong Brooklyn accent. The Stereo Society’s second release had been Raven, an exotic collection of his compositions performed by various unusual ensembles which mostly included him. These ranged from Atlantis (for five conch shells, tuba and chimes) through Raven (after Poe, for speaker, bassoon, clarinet/tarigato, trombone, saxophone, contrabass and gong) to Dune (after Frank Herbert, for bassoon solo).
The last piece on Raven is Circle. Its basis is a live performance by a trio of bassoon, didjeridu and multiple percussion. After the initial recording, two more performances were superimposed. This was Johnny’s introduction to hands-on layered production. Shortly after the Universe Symphony premiere in 1996, he phoned. No-one would commit to recording this huge new performing version of the work, which was understandable in the darkening economic gloom of the classical music world. Would it be possible to record using layers, for 70 minutes and 71 parts (eventually to be played by 18 different musicians). I took a deep breath and said most definitely maybe. We should talk. After all, the twentieth century still had a little while to run.
It is not practicable to record a typical symphony by layering, adding one or a few instruments at a time. Crucially, the music would probably by unable to achieve a coherent ensemble or lock, even if a conductor was synchronized on video, since too much of the music depends on instruments playing in unison and the slightest timing deviation will be obvious and detrimental. The first violins follow the leader, who follows the conductor. Visual cues are important, in several directions at once. There is nothing to be gained by diverting from the orchestral recording technique that has given us much beautiful musical sound. However, Ives didn’t score his piece in this traditional ensemble way. In his radical music lay a radical recording solution.
Although he had not always done so in previous orchestral works, in the Universe Symphony Ives was able to allocate just one instrument to each musical line. Also, much of the writing features complex rhythmic relationships which are held together only by the fixed points of an underpinning 16-second cycle, the Basic Unit (subdivided into pulses of two seconds). Such music can successfully be executed in either simultaneous or layered ways, since players follow their own different rhythmic paths between the two fixed points in time often with reference to a relatively small number of other players. The conductor acts as a traffic cop, indicating the two-second pulses, while cueing the instrumental entries (still a very significant task) as well as fulfilling the myriad other simultaneous conducting duties such as controlling balance and dynamics, and ensuring that intonation is correct.
There were still issues to be resolved on a medium time scale. Ives scores the percussion in recurring cycles whose instrumental exponents coincide every 16 seconds but who subdivide the period into segments of between 1/2 and 1/43, usually the reciprocal of a prime number. This generates the extraordinary effect of different pulses on their allocated instruments sounding as if coming in and out of focus as they converge on/diverge from the inevitable and dramatic point of coincidence every 16 seconds.
It’s very difficult to play 43 equal-length notes and to arrive at a coincidence with other players 16 seconds later, especially when your colleagues are dividing their time very differently. An earlier version of the piece had provided click tracks for guidance, at the appropriate temporal subdivision. Unfortunately, these can bleed from the performers’ headphones into the microphones, and it’s certainly not much fun to play to them. Further, this performance seemed to us rather stiff. Better we believed, to allow the musicians to ‘breathe’ in more intuitively human fashion (in more traditional music, even triplets, three beats in the time of two, tend not to be played as equal thirds of a beat). It helped considerably that we were working with progressive New York musicians, all of exceptional virtuosity, who had participated variously in the American Festival of Microtonal Music concerts over many years.
Now: how to start? The world has seen enough earnest layered productions starting with a kick drum at the bottom end (great sound!), adding the bass (great sound!), then a few other instruments (great sound!). Then: what happened to the kick drum? It’s hard to anticipate the appropriate sound in a familiar, tested context like basic pop recording, let alone with a piece as radical and different in sonic concept as the Universe Symphony. This would clearly have a sound difficult to anticipate. What was needed was a musical core, a collection of instruments whose first contribution to the eventual huge recording would have, and therefore define, a musical feel in its own right. We needed a soul.
We settled on an instrumental quartet: low bell (Skip La Plante), trombone (Julie Josephson), cello (David Eggar) and viola (Anastasia Solberg). Johnny stood before them, conducting the 16-second Basic Units in straight 4/4 time, with the video camera recording him simultaneously. We recorded this foundation sequentially through the symphony, with Johnny conducting and also cueing the invisible players who would perform in the future.
When the music was being recorded the only arbiter of tempo was Johnny and his own musical sensibility. Eventually, the music would phrase quite naturally, and this ebb and flow would influence all the performers to come. The result of these three or so initial sessions was a piece of spare musical texture which, even in its primitive form, we found ourselves listening to with real pleasure.
Issues of perspective and ambience in orchestral recording have never been less than controversial, so the sound and space of the individual instrumental performances were a major concern. Despite all sounds being generated in a room roughly 8’x20’, we would eventually have to mix them all together to give a satisfactory acceptance by a listener of the virtual image of 74 instrumentalists playing together in a shared concert hall space.
In a concert hall, where instruments have the space to sound good, in the place in which they were designed to function, it’s possible to put up just two microphones, typically together and angled at 90 degrees (a ‘coincident pair’) above and a little back from the conductor, and capture a reasonable recording of a performance. However, you are at the mercy of the room’s acoustics, which are often not ideal. Concert hall sound has to ensure that the audience can hear the players and the players can hear themselves, which is achieved through a little compromise all round. That’s if the architect can successfully predict the eventual acoustics in the plans, not a common achievement. (The best classical recordings are often made through use of some unforeseen environments, such as English town halls.)
Extra ‘spot’ microphones are used in a typical classical recording to emphasize sections that might sound under-represented in the main microphone pair. Starting in the 1960s, with the availability of large, high-quality sound-mixing desks, the coincident-plus-spot technique evolved (for some producers) into an array of mics spread around the orchestra, each set designed to capture a particular section. This gave the producer better control over the overall balance, but was a method demanding good ears and a firm grasp of the room acoustics. Arguably, the conductor should be controlling the balance in real time, so the control achieved using multiple mics may be redundant. Incompetence has produced some truly dreadful recordings, even of artists of the highest level and with a big budget, where the instruments sound thin as well as right in front of your nose. Without great care and skill, depth of sound and ambience could disappear entirely.
In multitrack studio recording, a mic is often as close as six inches to the instrument, and with a little care you can capture a surrealistically heightened sonic color. This always needs additional artificial reverberation to place it in context in the final two-channel stereo mix. However, this technique always places the perceived sound up front, and even with the most careful artificial reverberation this can sound uncomfortably artificial and congested. Our 74 players would not line up at the front of the stage without bumping elbows and getting bad tempered.
As can be seen in the session photos, the solution was to pull the mics back from the performer, capturing something of the room’s ambience (a helpful, short reverberation) each time an instrumental line was recorded. This contributed substantially to the openness of the finished recording. Artificial reverb was still used, but you have the aural sense that the instruments are arranged in a physical space with real depth. When defining the sonic perspective for the first few players, it was difficult to anticipate the spread when all sounds were in. However, applying experience, care and with some good luck, we hoped eventually to hear our orchestra of Lincoln Center.
One of the observations that has stayed with me from my studio apprenticeship in the mists of antiquity, 1970/1, is that ‘you can tell a good engineer by the way he uses bad separation.’ If one instrument's sound is bleeding into a mic dedicated to another, some control over balance is lost. With a room full, you might face disaster. But a good engineer can handle this matrix of interacting influences and often make a better recording than if each sound were conveniently quarantined, thanks to the added spatial openness resulting from the multiple manifestations of each instrument.
As the instruments built up over the months, we always tried to record in groups of at least two. This could be frustrating for the performers, since one person’s mistake in this demanding music would nullify another’s flawless performance, but the depth of sound which resulted from cumulative leakage contributed significantly to the heightened sense of perspective in the finished recording. (Our small recording area had started as an office but had morphed into a studio many years before with the encouragement of the demanding vocalist Carmel who had so appreciated the acoustic response and recorded sound of her voice.)
This office had not originally been designed for critical recording with wide dynamics, and for the Universe sessions it had been necessary to lower the background noise (remember, this is New York) with lavish applications of glue, foam and particle board. We persisted because something in these unprepossessing surroundings sounds nice and stimulates performers, a happy chance discovery. In this tiny space, we had accommodated seven musicians for Johnny’s Raven session, even delivering a gothic spoken vocal disturbingly close to your face. With care, and with all cramped in the same acoustic space, that particular recording sounds exceptional and spacious and is boosted by that overall, indefinable blend of bleed.
The first Universe recordings were into the disk recorder module of the Synclavier, a venerable and great-sounding (if operationally clunky by contemporary standards) classic machine. Its capacity is just 16 parallel tracks of up to three hours each. We would eventually sprawl over 98 audio tracks, which would eventually be gathered together and mixed down to just the stereo two. As a way station, we might have gathered them together, in sub-mixes, but in our uncharted territory it seemed impossible to achieve away from the context of the larger music. Which we couldn’t yet hear. We were heading into the unknown and would have to keep our options open. All performance recording strands had to remain distinct and sonically malleable.
Having filled up the Synclavier, the next recordings were spilled into our early version of what has become the industry standard, Digidesign’s ProTools system. While sonically slightly compromised, we could adjust on playback to be close to our ideal (for technical reasons, audio quality on individual tracks in a multitrack environment is far less crucial than on stereo program material). That gave us 24 more tracks. We now had 40, way to go. Fortuitously, we would be empowered thanks to a 21st century breakthrough which hadn’t existed when we started the project.
So much contemporary music recording computers is confused by the bells and whistles of a bloated computer application, the BYC (‘because you can’) syndrome. Much fancy stuff in the menu bar doesn’t matter in everyday practical recording. It was the audio hardware manufacturer Mackie’s (among others’) master stroke to manufacture a big electronic blob smaller than a microwave oven that is closer in spirit to an old 24-track tape machine. Our first purchase, at $5000, offered comprehensive audio editing. Our second, at $2000, had the same recording capacity and quality without the option of manipulation. In the olden days, a tape recorder of similar quality would cost some $50 000, but here was a machine that we might buy for the duration of the project and then sell by mail order. Even three years before our eventual gathering of all these instrumental tracks, the cost and logistics of maintaining this musical flexibility would have been prohibitive. We got lucky.
The musicians’ performances were recorded over a full year. Beside the obvious logistic difficulties of scheduling highly-rated professional musicians, who were joining the project mostly for the love of it, we had to move unprecedentedly huge amounts of digital audio data around the studio. Ives could only be recorded a few times a week, but the studio continued busy with other projects which, since they depended on days of block time, had necessarily to take priority. Since the capacity of computer hard drives even three years ago was far less than now, we routinely had to clean out and reload the music we were working on, a tedious chore. Eventually, we would have around 100 streams of audio lasting 64 minutes, or about 4.4 track.days, equivalent in music content to around 40 full stereo compact disks.
One of the most difficult tasks was editing these many sections into one continuous performance. Remember, at the start of the project, to make the backbone of four instruments, we began at the beginning and worked through the piece chronologically. In that way, we could feel closer to its development and ultimate progressive unfolding. Unfortunately, there were sections where none of the first four players made a sound. Rather than have Johnny conduct silence, and look even more surreal than he did when cueing non-existent players, we decided to build them later on a different foundation. Eventually, this left us with an enormous patchwork quilt, which required days of careful final assembly to transform into a continuous whole. It was easy to make a mistake, and a timing error of a twentieth of a second in the sonic layout would be destructive. Finally, we were able to play all 98 tracks from beginning to end. Now we could reduce this monster down to size: just two stereo tracks.
Mixing any large multitrack production follows a clear procedure. For the Universe, plugging up the sound sources alone, routing them to the desired channel on the mixing desk through whatever processors we thought necessary, took a couple of days of planning and setup. The main mixer has 80 automated channels, where balance and cuts (mutes) can be fine-tuned in non-real time. We used the ancillary non-automated 40-channel mixer to submix groups of instruments, such as the nine flutes (which for logistical reasons had taken ten tracks), before feeding a smaller number of channels at the main mixer.
Once all were gathered, we then had to mute any channel when it was not playing, since the noise build-up could be substantial and unacceptable in the demanding sonic world of classical recording. At the symphony’s height, all instruments would be playing. At the quiet extreme would be just one solo low bell, naked apart from 99 other flowing audio streams of hiss, distant passing trucks and people coughing or rustling paper. This preparation, for the 64 minutes of all tracks, took some three days. The textures are so exposed most of the time that the slightest oversight could compromise the cleanliness of the background, or leave in a distracting accidental noise. It’s a tribute to engineer James Rosenthal’s dogged persistence that the final result is cleaner than most conventional orchestral recordings.
The last stage, the sound mixing itself, took almost a month. With such complex textures, getting all the sounds to blend, lie in the right perspective and relate meaningfully to each other was uncharted territory. If you are mixing a rock+roll track, there are many norms to relate to, at every stage. Even if you choose to follow an unconventional path, there is still a clear reference point for you: the convention that you choose not to adopt. In mixing the Universe, the only references we had were previous examples of persuasive, large-scale orchestral recording. Our ears only: there was no procedural precedent.
The characteristics of each instrument were found highly (and surprisingly) interdependent, another potentially disruptive feature of our constantly-moving target. While I was hovering anxiously in the background, Johnny and James initially balanced the instruments in a style very characteristic of working musicians who are used to being in the middle of the action. Such people necessarily have an acoustic perspective very different from that of the audience. Consequently, this can make them quite removed from the virtual reality space that is evoked by a good stereo recording.
Johnny’s initial preference for the instrumental perspective was close and dry, and while it had clarity it remained unconvincing. Reverberation and instrumental perspective were needed. That was my starting point, to glue it all together by building on the enormous amount of work that had already gone into the mixing preparation.
Natural reverberation gives us clues about size and position. Artificial reverberation does the same. But an extraordinary, and very helpful result of adding reverberation to a studio recording is greater clarity and distinction between instruments, perhaps because it responds differently to each class of sound. Adding different flavors (varying in richness and decay time) to these 98 distinct tracks opened up everything, and made the soundscape far more exciting and clear. Unfortunately, by the same token it changed the overall balance dramatically, losing some instrumental distinction, much to Johnny’s alarm. So we had carefully to work through the piece once more from beginning to end: another several days’ work.
Finally, the Universe was finished. We rested over many seventh days. At home, we were all overwhelmed by the sound: its color and clarity in the service of the larger musical entity. Whether percussion-only, the pure ear candy, or stacked instrumental music, all was in the place of our choosing, with no compromise. I had experienced the relief and satisfaction at the conclusion of many ambitious album projects, but absolutely nothing approaching this scale.
Many times over the last five years, we all asked ourselves whether we would do it all over again. It’s hard to imagine a project on a similar scale. In the early stages, part of the terror for me was the unknown, possibly unsatisfactory entity that we would only hear clearly at the end, and my only comfort could be from the demonstrated consummate skill of both Charles Ives and Johnny Reinhard: four generations apart, these two. What kept me going was trusting in that combined skill and insight. Yet it was still a shock when I finally heard real music emerging from the control room speakers, truly the new planet condensing out of the primeval dust cloud.
Certainly, we would do it again. Not in the same way, of course. Even as we worked, available technology was improving in very helpful ways, some of which sped our plow over rocks that would have broken it in earlier studio configurations. The scope of the computer disk-based recording with which we started was transcended by that available at the finish, and had that been on hand at the start we could have moved far more quickly. But this is the story behind a radical endeavor, where we were in constant danger of being over-ambitious when confronting the scale of the music assembly. Perhaps it’s always like that. But we couldn't tackle the piece any other way.
If you got it right before you started, it wouldn’t be fresh. There’s a parallel with making the music itself, on which many great exponents in many styles have remarked. If you knew where you were going at the start, you would be on a path already traveled. You wouldn’t be doing anything new. This recording effort boils down simply to sound out of two loudspeakers, which must convey the musical vision of the composer. As for its effectiveness, that's just for you to judge with your ears, carefully forgetting everything I've written here. Listening is all, and no amount of words will change that.
- Mike Thorne May 2005
Charles Ives at the Stereo Society:
To Charles Ives' Stereo Society home page
Tradition And The Universe Symphony, by Johnny Reinhard
Previews and Reviews:
Thorne production commentaries
Home Albums Artists Contact Downloads Help Links New Shopping Words
We encourage shopping:
site maintenance by empathydesign, York, England