by Robert Auld
Because we can
The November 1992 issue of Audio magazine featured an interview with RCA producer Jack Pfeiffer. For a man renowned for his charm and his ability to work with temperamental artists, he had some surprisingly pungent things to say, especially about recording techniques. On the subject of multi-miking, for instance:
"...the fewer microphones you have, the more likely you are to get a really first class recording. Microphones are stupid. They pick up everything that comes their way. So the more mikes you have, the more phase differences you get, plus you pick up all the reflections from the acoustical environment. It all adds up to a mess."
And his description of how the use of more and more microphones crept up on producers and engineers:
"...then we began to think that sometimes the center of the orchestra, which was behind the soloist, sounded a little subdued-- that it wasn't being picked up properly. So we thought, let's put a couple of mikes up for the woodwinds, just to have a little more control. And then, well, maybe we don't hear the percussion quite enough. Eventually it just got out of hand."(1)
This is a good example of what I call the "because we can" syndrome. We figure out how to do something. We become intrigued by the possibility of taking the idea even further. We do so and become caught up in doing whatever-it-is. At some point the process becomes more important than the reasons for doing it. The original impulse was to solve a problem, but now we continue because we can, and because it seems easier that way. I think this is how we ended up recording big bands in multi-track studios, as if they were rock bands.
In audio recording, the trend in the 1960's was towards increasing complexity. First it was large mixing consoles, made possible by the introduction of transistorized circuitry. Then came the multi- track revolution, with 8, 12, 16, and eventually (in the 1970's) 24 tracks available on a single tape machine. The possibilities opened up by these new devices caused profound changes in how music was recorded in the studio, including:
1. Many more microphones were used in order to close-mike all of the instruments in an ensemble. The purpose was to establish enough separation between instrumental tracks so that electronic editing of individual tracks would be possible.
2. Often, close mikes alone would not establish the desired separation. Therefore, acoustical panels, isolation booths and general deadening of studio acoustics were also used.
3. The measures described above made it difficult or impossible for the musicians to hear each other properly. Therefore, headphone cue systems were used, with each musician on a set of headphones.
4. As the above measures eliminated any sense of room ambience in the recorded sound, artificial reverberation was used as a substitute.
All of this separated the musicians from the sense of giving a performance when making a recording. Correction of mistakes could be left to later when only the musicians involved need re-record their parts. Ensemble balances and "room acoustics" were all determined later, during mix-down. The musicians need not even record as an ensemble; different parts could be over-dubbed at different times.
By the 1970's, multi-track recording methods had become the norm for most kinds of music recording. For rock and pop musicians this was not a problem; the new studio techniques had been devised to suit their needs. But for the jazz big band, the new techniques posed many problems. These could be grouped into two areas: problems of performance, and problems of sound.
The performance problems are obvious enough: an ensemble of musicians used to hearing each other acoustically must now hear each other over headphones, with the balance they hear determined not just by them, but by the engineer doing the headphone mix. And for some instruments (such as the brass), the experience of hearing one-self over headphones while playing can be so different from the acoustical experience that it can affect one's ability to play normally, and so require a difficult period of adjustment.
The sound problems derive from the nature of the ensemble: the big band contains several groups of instruments that are balanced in a particular way. A good analogy is the vocal choir. When recording a choir of, say, 40 singers, it is not considered good engineering practice to separately mike each singer. The sound of the choir will not be properly reproduced by 40 separate close mikes; the blend and balance of the ensemble will be compromised. This is why, in recording a vocal choir, we use a few microphones placed some distance from the ensemble, so as to pick up the blend and balance that it already has.
In its own way, the big band resembles the vocal ensemble: we have a section of trumpets, a section of trombones, a section of saxophones, all supported by a rhythm section. The more astute arrangers and composers have always written so as to take advantage of the way these sections balance and blend, both internally and with each other. Older recording methods were designed with that in mind.
Multi-track techniques, however, are designed to provide separation between instruments for reasons of production convenience. When this objective is achieved, the natural balance and blend of the ensemble is stripped away. It is then the task of the engineer to restore those qualities by electronic means during mix-down. It is here that we face an engineering dilemma: the tools that are most commonly available are not adequate to the task. If it were just a matter of setting balances correctly, or of some re-equalizing, or of adding a little reverberation, perhaps an exceptional engineer could produce completely satisfactory results. But the sound of the ensemble consists of all of these things and more in a complex interaction. With current tools and methods the best engineer in the world can only provide an approximation of what we would like to hear.
There are often acoustical and ambient sound problems in live performance venues (especially in night clubs), but the immediacy and spontaneity of live performance cannot be obtained elsewhere. Multi-track techniques have been applied to live recordings of rock concerts with great success; the additional control over the sound that the method affords can compensate for a multitude of sins. However, using multi-track techniques to record the jazz big band in live performance introduces problems beyond those I outlined for studio recordings.
The main problem is the large number of microphones in use, and their tendency to interfere with each other. In the studio, careful placement of the musicians, acoustical treatment of the space and other measures can be used to control problems of sound leakage between mikes. But in live performance those options are often not available. If the space is acoustically small (as on a night club bandstand, for example) the recording can become a study in the effects of phase interference between multiple microphones. Recording all the mikes to separate tracks is no help, as the problem is acoustical. Indeed, recording to separate tracks may conceal the extent of the problem until mix-down, when it is too late to do much about it.
I do not consider the multi-track method to be the most suitable method for recording the jazz big band. The practical advantages do not compensate for the loss of sound quality. The vocal choir analogy points to a more suitable method: using a small number of microphones placed to pick up the natural balance and blend of the ensemble. Of course, this is not exactly an original idea. The first modern production team to try it out was led by Doug Sax and Lincoln Mayorga of Sheffield Lab in their 1976 recordings of Harry James and his Orchestra.
Sax and Mayorga had been known for the direct-to-disk recordings they made in the early 1970's. Those records, however, were made with conventional studio multi-mike techniques (very expertly applied, to be sure) with the elimination of the tape recorder as the chief novelty. With "The King James Version", the first of their recordings of Harry James, they broke with conventional studio setups completely.
The band was recorded in a medium sized chapel with moderately live acoustics. The band setup was approximately the same as for live performance. A single stereo microphone, centered on the horn sections, provided the main pickup. Two spot mikes were used: one on the acoustic bass, the other on the piano. Harry James and the other horn soloists simply stood at an appropriate distance and played at the stereo mike. The mix of, essentially, four microphones (counting the stereo mike as two) was then sent directly to the disk-cutting lathes (and to analog 2-track tape as a backup) without further modification. (2)
As heard on the 1996 CD reissue (Harry James and his Big Band, "The King James Version", Sheffield Lab 10068-2-F) this recording has a kind of clarity and palpable sense of reality that is all too rare in commercial big band recordings. And it is one of a very few recordings to accurately reproduce the wide dynamic range of this formidable ensemble.
If the secret to making a great sounding record was to keep the number of microphones to less than the number of fingers on one hand, everybody would be doing it. But recording with simple methods has other requirements, including:
1. The room has to be suitable to the music. The main pickup mikes will be hearing a good deal of the room. If it is too small, or too reflective in the immediate performance area, or has other obvious defects, a change of venue may be the only sure remedy.
2. The musicians have to be good. A band that plays with poor balance, no sense of dynamics, and is out of tune to boot, will sound every bit as bad as it is.
3. The engineer has to be good. He or she has to be expert at selecting and placing microphones for best advantage in various rooms. A clear concept of how a good big band should sound on record is a prerequisite.
Beyond these basics, there are methods for dealing with the variables presented by real-world recording sessions. The main microphone pickup can be configured a few different ways--coincident stereo with x-y cardioid patterns, mid-side stereo (a figure-8 mike for the sides, a cardioid for the middle), or a spaced array of two or three omnidirectional mikes. If the room has particularly spacious acoustics, additional mikes can be placed out in the room to pick up this ambience. Added instrumental sections (french horns, mellophoniums, added percussion) might benefit from section spot mikes. And the use of spot mikes on horn soloists is appropriate, as this practice is normal in most live performances.
Yet another way of keeping it simple is what I call "conservative multi-miking". As developed by the engineers at Capitol Studios in the late 1950's, this technique splits the reeds and brass to left and right, with the rhythm section in the middle. Single microphones are used on the saxes and on the brass (which are seated in two rows, trumpets on a riser behind the trombones). These two mikes are panned left and right to define the ends of the stereo stage. The rhythm instruments do leak into the two main microphones to some degree, but their placement and presence in the stereo image is determined by spot mikes, one for each rhythm instrument. Horn soloists and/or vocalists may also get their own mike. An advantage of this technique is that it preserves the internal balances and blend of the horn sections, while allowing well defined rhythm section sound.
Whatever the additions and variations of technique, the goal is the same: to produce a convincing illusion of the ensemble sound as it might be heard in a good live performance. In this respect, recording the big band is conceptually similar to recording the symphony orchestra. We should note that classical recording engineers, after passing through a phase during the 1960's and 1970's of using 20 or 30 microphones to record an orchestra, have since returned to using simpler techniques that better withstand the clarity of digital recording.
There are situations where trying to apply a simple main pickup will be an exercise in futility. One is when the room has inappropriate acoustics. If the recording must be made in such a room, switching to a conservative multi-mike setup (like that I described above) might help, as the ratio of direct to reflected sound will be higher. The loss of ambience, if a problem, can be made up with careful use of artificial reverb.
Then there are situations where amplified sound being produced in the venue is a problem. In particular, on-stage monitor speakers can bleed into any microphones that are not close to the intended sound source. This problem can be corrected by careful attention to the placement and operating levels of the monitor speakers, but there are situations where this will not be possible. (Example: the band is accompanying a famous vocalist, whose demands for monitor volume can be met only at the expense of the band's sound.) Then, close mikes on all the instruments may be the only workable solution.
I am not opposed to multi-track techniques when the alternatives are all worse. With care, a good recording can be made that way. Here are a few suggestions for avoiding the worst problems:
1. Do not over-mike. For example: many rock engineers will habitually use 8 or 9 microphones on a drum set. For the purposes of rock, where the drums are front and center in the mix, this is fine. But the jazz big band drummer has a supportive role in the ensemble. The exaggerated imaging made possible by the rock pickup works against that role. We need to pick up the sound of the set approximately as the drummer hears it. This can be done with 3 or 4 mikes (say, stereo overheads and a spot mike on the kick drum).
2. Be wary of interference effects among all those horn mikes. The saxophone section is a particular trouble spot. Because the saxes are not as loud as the brass, each sax will likely get an individual mike. The placement and polar response of those mikes are important. In this particular case, mikes with bi-directional (figure-eight) patterns may be the best solution; the narrow acceptance angle and pronounced 90 degree rejection can be used to advantage.
3. As a matter of course, you should monitor the session in both stereo and mono, so as to catch any microphone interference-related problems. If certain mikes seem especially problematic when included in the mix, record them to separate tracks. This will, at least, preserve the option of not using them in the final mix.
4. If the placement of the instruments in the stereo mix corresponds to how they were placed on the stage, many imaging and phasing problems can be reduced, or even eliminated. Disregarding this rule when mixing live concert material almost guarantees that such problems will occur.
5. If you are doing the final mix-down, try and do it in a facility where you are intimately familiar with the sound of the room and the monitors. The mixing flexibility that multi-track methods permit make it easy to make bad mixing decisions. If you are being fooled by unfamiliar monitors, a bad mix is likely.
By "recent" I guess I mean, of the digital era. This one qualifies, having been recorded with the Soundstream system. I have not heard the CD issue (which may be available at Amazon.com as of this writing) but as the quality of the LP is exceptionally good, I would not expect drastic differences.
This recording is another important example of the use of a simple main pickup. It was made in a medium-sized theatre with a nice, resonant acoustic. The band was set up on the stage as if for a concert: three rows of horns, rhythm section to the left. Jack Renner placed a spaced array of 3 omni-directional mikes directly in front of and about 7 feet above the saxes as his main pickup. A pair of mikes were placed out in the hall to pick up additional ambience. Spot mikes were used on the piano, the bass, and on the various horn soloists. All of this was mixed live to a 2-track digital recorder.
In comparing this recording to "The King James Version", the main differences are in how the two bands balance themselves, and in the use of spot mikes for horn solos. A lot of the difference in balances has to do with the arranging; Harry James used Ernie Wilkins and others who wrote in the acoustic style developed before electronic amplification became common in live performance. Thad Jones (who wrote all the charts for "Naturally") came out of that tradition (he wrote many arrangements for Count Basie in the 1950's), but adopted a more modern style when writing for the Thad Jones/Mel Lewis Orchestra. That style tends to rely on the sound system to balance soloists against the band, which allows backgrounds to solos to be louder. It is also a denser style--there is often more going on at once--and one that encourages powerful brass playing. We certainly get that from the Mel Lewis Orchestra, and the recording reproduces it all to great effect.
I have a minor nit to pick: Mel Lewis' pianist, Jim McNeely, is an exceptional soloist and is featured on three of the six tracks on the record. The rather plain mono pickup of his piano does not really do him justice. I would prefer to mike the piano in stereo and spread it a bit in the stereo image. Since the piano would be amplified in live performance, I don't think this would violate the spirit of this recording, and the tonal quality would likely be better. However, this is only a detail in a generally superior recording.
This is a live concert recording, at the Montreux Festival in Switzerland, mixed live to digital 2-track (the Sony system, probably the PCM 1610). It is a very interesting study in the problems of a multi-mike pickup in a live performance venue.
Dave Richards evidently did not have a full sound check before the concert began, so we get to hear him building his mix during the opening number, "Wind Machine". He starts with just the rhythm section mikes open, as Basie does one of his classic piano-and-rhythm introductions. Then, as the other sections of the band start to join in, the mikes on those sections are brought up, section by section. This allows us to judge the amount of leakage between the various section pickups; it is significant. By the end of the second number, "Blues for Stephanie", things have settled down enough that the remaining changes in the mix are largely confined to fine tuning.
Throughout the record, the balances between the horn sections are a problem. The saxes, especially, tend to loose clarity and at times all the horns tend to mush together in the mix. Integrating the close-up stereo drum mix into the general sound also proves difficult. There are constant little adjustments to the mix throughout as Richards wrestles with these problems. As good an engineer as he is, he can only succeed to a limited extent, given the effect of so many mikes interacting in a small space.
Fortunately, the performance captured here transcends any problems with the sound. If you want an example of a great big band in full cry, inspired by an appreciative audience, this record is hard to beat.
Diane Schuur and The Count Basie Orchestra
(CD, GRP records GRD- 9550, recorded 1987)
This should have been a great sounding record. Diane Schuur is a powerful singer and the Basie band is still a first-rate ensemble, even without Basie. The recording, performed in front of an audience, was made in a studio (A&M studios, at the time one of Hollywood's biggest) and the engineer for the live pickup was Alan Sides, who is highly respected in the business. And, co-producer Jeffrey Weber was an assistant on some of Sheffield Lab's early direct-to-disk recordings. So, why don't I like it?
The chief reason is the balances, which are very strange and have nothing to do with how the Basie band sounds live. For example, in much of track 7, "You Can Have It", the loudest instruments in the mix are the bass (really in-your-face) and...the drummer's high-hat cymbal! The rest of the band is wailing away (this is definitely not a ballad) somewhere off in the background.
The session photos on the album show the band set up in normal concert fashion: three rows of horns with the rhythm section to the left. But what we hear on the recording is a huge drum set spread from left to right (with that high-hat poking out of the center), the piano about in the middle, the bass out in front and much too loud, and the saxes and brass both to the left and right, set well back. The brasses hardly ever sound like trumpet or trombone sections--individual horns tend to stick out and recede--and they may as well be in the next room for all the presence they have. The overall effect makes it sound as if Diane Schuur is being backed by an energetic combo of a half dozen players.
While it would be tempting to blame it all on the mix-down engineer (Don Murray of Capitol studios), he had to answer to two producers, two executive producers and a star singer. I suspect all those cooks fixed it in the mix, really fixed it. In a perverse way, this record is a good argument for live to 2-track recording.
This is another big-band-with-vocalist affair, recorded to digital multi-track in Toronto. The performers are certainly fine; Mel Torme is in good form, McConnell is an expert arranger and The Boss Brass are a collection of the best studio musicians in Toronto. The production team is also confidence-inspiring: producer Carl Jefferson has made dozens of albums for Concord Jazz notable for their fine sound. All the ingredients were in place for a spectacular sounding record. But the result does not sound great; it is merely adequate. What happened?
This record is a perfect example of a good engineer and producer struggling with the limitations of multi-track methods. The worst problems are avoided; there are no strange balances, ugly tonal qualities, or other such distractions. There is always reasonable clarity, and nothing is overdone. Yet, the mix is oddly two-dimensional with little sense of depth. If you were in the room with this 22-piece big band the impact of the sound would just blow you away, but the recording is tame, to the point that one becomes aware of such details as the artificial coloration of the reverb.
Comparing this recording with "Sinatra's Swingin' Session"(3) is an eye-opening experience. The older recording features a wide stereo spread that sounds wonderful on loudspeakers, and the clear divisions between instrumental sections, the way little details of the arrangements are highlighted but never distract--well, this is how a recording of this kind should sound. Only a slight edginess in the treble and a bit of tape hiss give away the recording's age.
"Mel Torme/Rob McConnell and The Boss Brass" was recorded with, I think, about 20-odd microphones. "Sinatra's Swingin' Session" uses about half that number and sounds much better, even though the number of musicians in the ensemble is somewhat more.
It is clear that we have to stop using so many microphones to record big bands; that we need to pay more attention to what the music and the ensemble really sound like, and use methods that will reproduce that sound. The alternative--just continuing to use multi-track methods because "that's how it's done"--will not do justice to the music, or to our ability to record it.
1. The quotes are from "The Audio Interview: Jack Pfeiffer",
by Susan Elliot, Audio magazine, November, 1992.
2. My thanks to Doug Sax for responding to my questions
about these sessions.
3. "Sinatra's Swingin' Session" has been released
on CD, Capitol CDP 7 46573 2. The recording method used for this album
was described in Part 1 of this paper, under the heading "The engineering
style of the three-track era: some examples". I also described the main characteristics of that studio method as "conservative multi-miking" in The requirements of simplicity section above.
copyright 2012 by Robert Auld