Part I appeared in the February issue, pp. 16-18.
The "art" of sound recording consists of selecting the proper microphones for a given situation and placing them in the most advantageous position. We will look at three basic techniques--coincident, near coincident and spaced omnidirectional--and then discuss which might be more beneficial given the specifics of organ layout and room acoustics.
Coincident Microphone Placement
We've probably all been to a concert where a professional recording engineer has set up one very large and impressive microphone on an equally large and impressive stand with which to make a stereo recording. Within that large microphone were actually two directional microphones which the engineer, with an amazing amount of flexibility, can select, position and modify by remote control. Coincident means "to occupy the same area in space," and that's what a stereo microphone has: two mono mikes occupying the same space within the microphone housing. One of the characteristics of all coincident techniques is that the sound arrives at the left and right microphones completely "in phase."28
Figure 10 shows how you can position two cardioid (unidirectional) microphones in a coincident position. The strength of this technique is that it gives a fairly realistic stereo image when played back through speakers (i.e., the first violins seem to be on the left, and the double basses seem to be on the right). The weakness is that the stereo image seems to lack a "sense of space."29 Since cardioid microphones are directional, they accept sound from the source in front of them and reject sound, such as reverberation, coming from the room behind the microphones. This may be a plus in an extremely reverberant room.
Professionals may also choose to use two "figure of eight" directional microphones30 set in an "X" pattern at 90° to one another, each of which picks up not only sound from in front but some from behind as well. This coincident technique, invented by British scientist Alan Blumlein in the 1930s, can give very natural sound in some circumstances.
Another coincident technique favored by some professionals is the "M-S" system31, which requires a special processing network to resolve the recorded sound into left and right stereo signals. An advantage here is that it gives the mixing engineer greater control of the stereo image from the mixing desk than is available with any other technique.32
Near-Coincident Techniques
In a successful attempt to improve the stereo illusion, sound engineers began to separate the coincident microphones ever so slightly so the sound arrives at the microphones just slightly out of phase, thus contributing additional information which enhances the stereo image.33
We'll discuss two similar setups, the ORTF system from the French National Broadcasting Organization and the NOS system from Dutch Broadcasting. Both of these use cardioid (unidirectional) microphones. The ORTF system splays the microphones out at a 110° angle and separates the recording capsules by 17 cm (63/4"), whereas the NOS has the mikes at a 90° angle with a 30 cm (113/4") separation.34 These near-coincident techniques are superior to two strictly coincident cardioid microphones. Professional audio stores sell inexpensive adjustable rigs to hold two cardioid microphones on one mike stand in a near-coincident configuration similar to NOS (see Fig. 11). A near-coincident variation of the Blumlein technique places two figure of eight mikes at 90° to each other in an "X" configuration, but separated by about 7".
Spaced Omnidirectional Mikes
In many of the coincident or near-coincident configurations we just discussed, you are recording primarily the sound of the organ alone. With a spaced pair of omnidirectional microphones, however, you are recording not only the direct sound from the source, but also the room's response to the sound--reverberation--which is a big plus in organ recording. Under the best circumstances, the sound of spaced omnis can be very open and sensual indeed.35
How far apart should the microphones be spaced? The minimum is about 4'--that is, 2' on each side of the centerline drawn between the sound source and the microphones. Omni mikes are typically spaced 1/3 of the way in from the edges of the sound source. For example, if the organ is 18' wide the microphones could be placed 6' apart--3' on either side of the centerline (see Fig. 12).
If the sound source is very wide, however, two omnidirectional microphones may be spread so far apart that an aural "hole in the middle" becomes apparent. This is alleviated by placing a third omnidirectional microphone directly in the center, and then with a mixer adding just a bit of its sound to the left and right channels.26 If the volume of the center mike isn't kept quite soft compared to the left and right mikes, however, you will kill the stereo effect. A variation of this "center channel" technique provides a third mike to accent a soloist.
Spaced Pair of PZMs
Spaced PZM microphones behave very much like a spaced pair of omnidirectional mikes. The bass response of PZM mikes is enhanced when they are resting on a surface at least 4x4'--thus the floor is an excellent place for them. However, you don't want to bury them in the shadow of a pew or other obstructions, as this will modify their hemispherical pickup pattern. The author's favorite PZM setup uses two 4x4' pieces of masonite37 which are stored at the church and placed on top of the pews as needed. For flattest frequency response, place the PZM 1/3 of the way off center--8" off center on a 4x4' panel38 (see Fig. 13). For personal analysis recordings, you may be able to position the mikes on the console (see Fig. 14).
Which Is Better?
There is a spirited debate in the audio world between the proponents of coincident or near-coincident techniques versus the advocates of a spaced pair of omnidirectional mikes. The coincident techniques--which ensure that the left and right channels are in phase--used to solve problems that no longer exist today: the difficulties of cutting the master from which LP recordings (remember LPs?) were stamped, the difficulties of phono cartridges (remember them?) tracking low frequency sounds on LPs, and the problem of sound cancellation on mono radio stations (a rare breed) as out-of-phase stereo signals were summed to mono.
Further, as Edward Tatnall Canby observed in Audio, the bureaucracy at National Public Radio mandates coincident recording techniques (especially M-S) and gives them a hard sell in spite of the fact that many listeners find something important missing in the resulting recordings.39 Agreeing with Mr. Canby, Christopher Czeh, Technical Director of WNYC Public Radio in New York wrote:
The phase differences between spaced omnidirectional microphones help the listener in mentally recreating the spatial acoustics of the original performance. I have used spaced omnis for classical recordings for six years and have obtained excellent results. The major reason I prefer spaced omnis over coincident mikes is that they sound better in most circumstances.40
David Wilson of Wilson Audiophile Recordings, agrees and notes the crucial difference between the ears and microphones:
Microphones "hear" very differently than ears do. The microphone is very literal in what it picks up. There is no neurological ear-brain system that compensates for ambiance and perspective. For most recording, I prefer omnidirectional microphones because they are more natural sounding. That is, they more naturally integrate the sound of the instrument with room acoustics, and that's very important with pipe organs. In almost every organ recording I've made, however, I've experimented with a coincident pair of directional microphones, almost out of a sense of technical duty. After listening to the test results, I've almost always gone back to a spaced pair of omnis.
Frederick Hohman of Pro Organo has a different view:
My personal preference is for good directional microphones--not omnidirectional. A pair of these can be placed in any conventional pattern and configuration one desires. A single stereo mike could be the easiest way to do a quick setup, since this eliminates the factor of microphone spacing.
Jack Renner of Telarc, who has recorded Michael Murray in many diverse situations, looks at the broad picture:
The thing about coincident or near-coincident microphone techniques such as the ORTF configuration with directional mikes, or the crossed figure of eights, or the M-S systems, is that while they may not be everyone's cup of tea in terms of finished sound--I myself like the sound of a pair of spaced omnis--the coincident techniques will give you a perfectly acceptable recording and are a very safe way to approach a recording of anything.
How Far from the Organ?
How far the microphone(s) are placed from the sound-producing elements of the organ is one of the critical decisions in any recording setup, and it won't be the same for all circumstances. If an organist is making personal "analysis" recordings, a relatively close microphone position will give increased clarity, especially in a reverberant room. If the goal of the recording is to show the organ/room combination to its best advantage, a more distant position will increase the proportion of room (reflected) sound. Pipedreams' Michael Barone, who has probably listened to more organ recordings than anyone and who has made quite a few organ recordings as well, has some definite opinions:
A lot of people think that to get a sense of space they need to record from the back of the hall, and so many organ recordings are made miserable by this "gray tunnel" effect. But you don't want to put the microphones within two or three feet of the chamber either. You want to place the microphones where there is an obvious focus of the sound, but where the sound has begun to excite the room and participate in the acoustics of the space.
John Eargle of Delos agrees that most people tend to place the microphones too far from the organ, and describes how he decides where to place the microphones:
First I walk around the room while listening to the instrument. The best place for the mikes is within a zone where the direct sound of the organ and the reverberant sound coalesce. What you have at this magic point is a very natural blend of room sound, plus good articulation from the instrument.
David Wilson is a firm believer in recording some "tests" to determine the best place for the microphones:
Generally I will start testing with a very close placement, say perhaps 10 or 12', which is closer than I believe is ideal. We will record 30 seconds or so of music and move the microphones back--generally I move them back in 3' increments--and record another test. We repeat this procedure five times. I also vary the height, starting with a height which is less than ideal--I believe 8' or so is the minimum satisfactory height--and go up from there to perhaps 20' or higher. I also vary the spacing between microphones. I start with the microphones closer together than I think they should be, say 4', and separate them further. By listening to the playback of these tests, we discover the best distance from the organ, height and between-mike spacing.
Jack Renner also stresses listening:
In placing the microphones, a lot of it is experience and a lot is listening. I have the organist play with various combinations of stops and I walk around the room listening until I find a place that sounds focused and blended--a place where all the registers seem to come together and where the bass pipes especially sound good and solid. You will find a point where there is good balance between the direct sound from the pipes themselves and the reverberant sound of the room, where you have a pleasing mix and where you don't hear various voices "popping" in and out, which is one of the biggest pitfalls in organ recording.
Aesthetics and Mike Distance
Crucial factors in deciding how far the microphones should be placed from the organ may well be the type of organ, the type of room and the type of music to be recorded. You might expect one type of presence, articulation, clarity and room sound for an all-Bach program on a tracker organ in a moderate-sized church, and have completely different expectations for a program of Romantic music on a large Romantic organ in a reverberant cathedral. Personally, I think a good number of recordings of the latter type have been ruined because the engineer was striving for too much clarity. These misguided attempts often have harsh, close-up organ tone and inadequate reverberation from mike positions that were too close. In this context it is very educational to listen to the same organ played by various artists and recorded by different engineers.41 Despite what the "experts" say, only you can decide if you like cathedral music to wash over you in a sea of reverberation.42
More than two Mikes?
When the sound source is very wide, for example a symphony orchestra or an organ that is quite spread out from left to right, you may have to spread a pair of omnis so far apart that you begin to lose sound from the middle--giving rise to the expression "the hole in the middle." Some recording engineers solve this problem by placing a third omni mike directly on the center axis of the sound source and mixing it on site into the left and right channels at a much softer level. This is Telarc's standard three-mike setup for symphony orchestras, although for concertos they will use additional mikes if necessary to highlight the soloist. Telarc's standard organ setup is two spaced omnis. However, they used a three-mike setup to record the wide organ at Methuen Music Hall, with the mikes about 35-40' back from the organ. When John Eargle recorded Robert Noehren on the large Rieger which sits front and center in the chancel of the Pacific Union College Church in Angwin, California:
We used three spaced omni mikes, 15-18' from the organ case. This case, like most trackers, is fairly shallow--eight feet deep at most. If there is a magic zone for mike placement that seems to work with this type of instrument, it is in the 17-20' range.
Other recording engineers, David Wilson included, do not use this technique because they feel that mixing a centrally-positioned monophonic mike into the left and right channels dilutes the stereo effect.
Recording the Reverberation
In order to capture the way an organ really sounds in a room, it is sometimes necessary to add additional microphones to record the reverberation. Few American churches have an excess of reverberation, but many have more than would be captured by the setups we have described thus far--two or three microphones placed relatively close to the organ. So a pair of microphones at some distance from the organ, with a small amount of the output of the left "reverb" mike mixed into the left channel and vice versa, does the trick. One might think that a single mike placed at a distance with the output shared between the channels--a variation on the "hole in the middle" technique--would suffice, but this is not usually done:
Reverberation from a single [distant] source divided between the left and right channels is unsatisfactory because the resulting sound, which, to give a natural effect, should be distributed across the space between the two loudspeakers, appears in this case to emanate from a single point.43
When John Eargle recorded Robert Noehren playing the organ he had built in 1967 for The First Unitarian Church in San Francisco:
I wanted to accurately portray the physical layout of the organ--which is arranged left to right in the rear gallery--so the primary mikes were a pair of directional cardioids splayed in a near-coincident configuration. The room is not reverberant, but there is enough room sound to give a nice glow and enhance the music. So we used an additional coincident pair of directional mikes, aimed more or less at the side walls, to capture this glow.
When Michael Barone recorded the Fisk organ at House of Hope in St. Paul, Minnesota, he encountered a similar situation:
The organ, which has a Rückpositiv, is located in the rear gallery. It generates a lot of bass energy, but that is not apparent in all areas of the room and generally not along the center aisle as the bass energy tends to hug the walls. So we placed a single stereo mike in the center aisle on a stand tall enough to get it well above the Rückpositiv. We also placed a pair of omni mikes a little further back from the organ closer to the side aisles, and then mixed the four inputs together until it sounded good--it's a little like cooking!
John Eargle describes his technique recording the large encased Rosales tracker organ at Trinity Episcopal Cathedral in Portland, Oregon:
The organ is located at the back of a rather deep chancel. Two omnidirectional microphones were used for direct pickup of the instrument in the chancel area, while a coincident pair of directional mikes was placed out in the church for reverberant pickup.
Improving the Room
There are basically two things you can physically do to the room before recording: decrease the noise and increase the reverberation. Potential noise sources that you may be able to do something about include: ventilation and heating systems, buzzing fluorescent lights, open doors or windows, etc. You may have to work around other noise sources like vehicular and air traffic, school children, and even expansion sounds from the roof as the sun heats it up mid-morning and it cools down in the evening.
It will increase the reverberation in an empty church significantly if the pew cushions can be removed. This is John Eargle's standard practice and he gets a lot of benefit for a reasonable effort. If the church is large and storage of the cushions is a problem, try stacking the cushions from two pews on top of the third, etc., etc. This will expose two-thirds of the hard pew surfaces (see Fig. 15). Or if the church has theater-type chairs with plush cushions, flip all the bottoms upright to minimize the absorptive surfaces.
Some Typical Solutions
The following are some microphone selection and placement solutions for various types of rooms:
Excessive reverberation--Use a pair of cardioid (unidirectional) microphones in a near-coincident configuration such as ORTF or NOS.
Minimal or average reverberation in a large room--Start with a pair of spaced omnis or PZM mikes and then, if you have mixer capabilities,44 try an additional coincident pair of directional microphones further back in the room mixed very subtly into the main pair (left into left and right into right).
Very wide sound source--Use a pair of spaced omnis or PZMs 1/3 in from the edges of the sound source. If necessary, a third omni in the center can be very subtly mixed in if there is an audible "hole in the middle." Alternatively, experiment with a splayed pair of directional cardioid mikes in the ORTF or NOS configuration.
Divided organ on the left and right sides of chancel or gallery--Try a pair of spaced omnis or PZMs. At Grace Cathedral in San Francisco, David Wilson recorded the huge Aeolian-Skinner which is divided in left and right chambers in the chancel plus a Bombarde Division at the rear center of the chancel. He used just two omnis spaced 8' apart, on stands about 20' high placed in the nave about 15' from the organ.
Rear gallery placement or organ high in the chancel--Unless the rear gallery is very deep (potentially allowing microphone placement within the gallery), you will need stands that allow you to get the microphones well up in the air.
Gallery placement with a Rückpositiv--The mike stands must enable placing the mikes well above the Rückpositiv if the correct balance between divisions is to be recorded (review the section on mike stand safety).
Organ is in a chamber on one side of a large chancel--The "standard" placement of a pair of spaced omnis on either side of the center aisle or a pair of coincident mikes in the center aisle pointed toward the rear of the chancel will pick up too much sound in one channel and not enough in the other. If the chancel is big enough, you might try a pair of spaced omnis within the chancel, each of which is the same distance from the organ.45 Alternatively, you might try a pair of cardioid directional mikes in the ORTF or NOS configuration within the chancel placed opposite the organ chamber and pointing at it. A third possibility is a pair of PZM mikes taped to the chancel wall opposite the organ. With these solutions, the reverberation component will likely be nil, calling for reverberation mikes further back in the nave.
Organ is in a chamber on one side of a small chancel--If the chancel is not that large, try to adapt either of the above alternatives through placement within the nave. For example, if the pipes are on the left of the chancel, place a near coincident pair of cardioids on the right side of the nave pointing towards the organ. Or if using a spaced pair of omnis, keep the left and right microphones approximately equidistant from the pipes. Always avoid placing an omni mike too close to a wall to prevent hard reflections.
Modifying Registrations
If the purpose of the recording is to hear the effect of a piece you're learning or to document a recital performance, then the registrations are chosen for the live performance and the recording is secondary. But if the primary purpose is to create a recording which shows the music, artist, organ and room off to best advantage, the question of modifying registrations to serve that end is legitimate. English recording engineer Michael Smythe offers this advice:
One must keep a keen ear open for stops that do not record well. What may sound fine in the church may come through the loudspeaker as an opaque noise, for example, the booming sound which 16' pipes quite often produce on certain notes. Therefore the organist has to rethink his registration for recording, which may be totally different from a recital. Sometimes one can do nothing about it, however, there being no suitable alternate stops.46
The late Michael Nemo of Towerhill, who made numerous recordings of John Rose on the huge Austin at St. Joseph Cathedral in Hartford, concurred:
From a technical point of view, there are some problem stops. For example, 32' flues like a Bourdon or Open Wood can be quite pleasing in person. As most stereo systems won't reproduce anything at all from the bottom range of a 32' stop, however, it doesn't mean much on a recording. And by virtue of strong, low-frequency fundamental, these stops often create enormous standing wave problems in the room. No two 32' stops are alike in the way they record, however--some can be quite delicious and others only cause problems.
Excessive Dynamic Range
In addition to eliminating problem stops, there is the question of the dynamic range of large, Romantic organs. Consider Dupré's Cortège et Litanie, which begins very quietly on a solitary Choir Dulciana (sans pedal) and ends fff with a page of crashing chords over an octave pedal point. While this enormous dynamic range can sound glorious in person, if the recording level is set as it should be for the fff climax, the pp sections on tape will recede into inaudibility. If you turn up the playback volume so you can actually hear some detail in the pp sections--which you certainly can in a live performance--when the piece gets to the ff and fff sections you will be blasted into the next county unless you turn the volume back down again.
In the analog days when recording was done on magnetic tape, you would have a good bit of tape hiss competing with the Dulciana and thus there was motivation to avoid excessively soft sounds. But now that professional recording is done on hiss-free DAT,47 many engineers--reveling in the huge dynamic range of DAT recordings released on CD--are creating recordings of large, Romantic organs that virtually force listeners to keep their fingers on the volume control, especially when using headphones.
There are two ways around this. One is for the organist to compress the dynamic range of the organ by, in the Cortège et Litanie, for example, leaving the sub and super couplers off48 for the climax and substituting the Geigen Diapason for the Dulciana at the beginning--at that volume level the Geigen will sound like a Dulciana and the climax will be good and loud nonetheless. Another option is for a recording engineer who reads music and can follow the score to increase the volume level of the very quiet parts at the mastering stage.49 The final recording should not simply enshrine the technical capabilities of the DAT/CD medium but should be a reasonable facsimile of the way the performer's artistry actually sounds in the room.
Conclusion
Making recordings can be a useful tool for self study, a means of communicating with potential employers and professional competitions, a satisfying hobby, a part-time career, or the means to artistic fulfillment. We have endeavored to explain the bare minimum required for an understanding of the process. We have given some "quick and easy" prescriptions for personal recording. And finally, we have explored professional recording techniques used by some of the top pros in the field, whom we sincerely thank for their time and generosity.