by Charles P. Repka
Since the introduction of stereophonic recording, there have been three basic philosophies with respect to microphone placement. These are:
1) multiple mikes,
2) spaced omni mikes (using two or three mikes), and
3) coincident stereo mikes.
After the initial experimental period in stereo recording had passed, the coincident stereo mike was just about abandoned in the U.S. with most American recording engineers choosing some sort of multi-mike technique. In Europe, coincident microphones became the technique of choice for two major reasons:
The conviction that it produced a more realistic stereo effect and, more importantly, that it produced a compatible signal easily converted to mono.
In recent years, however, there has been some degree of role reversal with a renewed interest in coincident mikes in the U.S. and the acceptance of the multi-mike techniques in Europe.
(Look at the session photos found on the jackets of some of the more recent Philips and DGG recordings, and you will see a veritable forest of micro phones.)
As a recording engineer, I've had to use all three techniques and sometimes i combination of them. For a professional engineer, each technique has its uses in the recording studio.
However, for the amateur recordist, and in particular with recording a live event, coincident mikes have several major advantages.
How It Works
The basic coincident technique was developed in the 1930s (along with the first stereo recordings) by an English engineer Alan Blumlein. Blumlein used two figure-eight pattern ribbon mikes mounted so that their pattern lobes were at right angles (90°) to each other, as shown in Fig. 1. The stereo effect is produced, primarily, by the difference in amplitude generated in the two mikes by the sound source. A sound on the right generates a larger signal in mike B than in mike A. A sound directly in front produces an equal signal in both mikes and a sound on the left produces a larger signal in mike A than in mike B. The same process takes place with spaced omni mikes, but because of the spacing, there is also a time delay between the two signals. While this time delay causes no real problem when recorded on a tape machine, it does generate an "out-of-phase" signal (usually at low frequencies, but dependent upon the space between the mikes) that results in a large vertical modulation component on a stereo disc which is difficult to cut, plate, and play back. It can also produce a loss in gain and unpleasant sound if the two channels are combined into a single mono signal. (This is of great concern to European broad casters. Although most broadcasts are made in stereo, over 85 percent of the receivers are still mono.) Since the coincident mike has both its transducers mounted on the same vertical axis (see Fig. 2), the arrival time is identical in both channels, thus reducing this problem to a large degree.
Fig. 1--The basic "Blumlein" pair of mikes.
At this point we should stop and consider that studies of the hearing process have shown the ear to use intensity, phase, and time delay to produce the spatial image. The fact that a coincident mike produces a realistic image means that a) the brain is very clever in constructing a stereo "illusion" from a minimum of information, b) we do not fully understand how the coincident mike really works, c) we do not fully understand how the hearing process works, or d) all of the above. I, personally, vote for d. (Editor's note: Me too)
Other Coincident Techniques
While Blumlein was limited to using figure-eight pattern microphones, modern coincident mikes can offer a choice of multiple patterns allowing the engineer to use the cardioid or hyper-cardioid patterns (see Fig. 3).
These patterns work as well as the figure-eight pattern microphones in producing a stereo image, but pick up less of the ambient hall sound.
These other patterns were used in the development of two variations on the basic coincident technique. These variations are the M-S technique and the O.R.T.F. technique. The M-S (mid-side) technique uses a forward-facing cardioid mike and a side-facing figure-eight mike, as shown in Fig. 4. The mike outputs form a signal matrix with the cardioid mike supplying the sum L+R signal and the figure-eight mike supplying the difference or L-R signal.
These two signals are then added and subtracted electrically (either before or after recording) to produce pure Left and Right signals. This sum and difference technique can be used to pro duce signals that are electrically equivalent to the various X-Y patterns normally used in coincident recording (see Fig. 7).
The advantage of this method is that the acoustic characteristics of the recordings can be controlled electrically without having to physically move the mike. The amount of separation can be varied from pure mono to pure stereo, or somewhere in between. It also allows the amount of hall sound to be increased or decreased without changing the amount of presence (distance from the performers).
As originally developed (also by Blumlein), the M-S matrix decoder was a pair of transformers with split secondaries connected as shown in Fig. 5.
But for this circuit to work properly, the transformers have to be of high quality and closely matched with respect to frequency and phase response, something that is very hard (and expensive) to do in a transformer.
These days, however, it is much easier to do the necessary additions and subtractions with electronics. A simple circuit has been devised by my friend Reg Williamson in England using two op-amps to replace the transformers.
The basic schematic for this op-amp is shown in Fig. 6 with the construction details along with other circuit applications to be found in issue No. 1, 1977 of The Audio Amateur. (Back is sues are available from The Audio Amateur, P.O. Box 176, Peterborough, NH 03458.)
The O.R.T.F. method uses two cardioid mikes facing outward with an angle of 110° between them (see Fig. 8). This method is said to approximate the effect obtained with dummy-head (Kunstkopf) recording and also produces excellent results when played back through loudspeakers. Because of the spacing between the transducers, the O.R.T.F. method does not have the time-coherence properties of pure Blumlein. The initials O.R.T.F. stand for Office de Radiodiffusion Television Francais, the French government radio network that developed this technique.
Fig. 2--Construction of two typical stereo mikes.
Probably the strongest virtue of the coincident mike technique is its simplicity under actual working conditions. One simply places the mike in a central location that gives a good balance between the musicians and the acoustics of the hall. It is this simplicity that makes coincident mikes a favorite of broadcast engineers recording (or transmitting) live symphonic concerts (I should say of European and especially BBC broadcast engineers, as we Americans have not yet seen the light).
The trick is in finding the right spot, but this is always a problem no matter which technique is used. The stereo mike is somewhat easier to locate than spaced omnis (I always worry about the spacing; are the mikes too far apart or too close together?), and it is infinitely easier to use than the dozen or more mikes used in a typical multi-
mike recording. Locating and balancing the levels from a large number of mikes takes a great deal of time and skill, and don't forget the logistics problem of setting up all those stands and mike cables. Of course, many of the balance problems can be delayed by using a multi-track tape machine (don't worry, we can fix it in the mix!), but that is not an option available to most amateur recordists.
Fig. 3a--X-Y crossed cardioid pair of mikes.
Fig. 3b--X-Y crossed hyper-cardioid pair of mikes.
Fig. 4--The M-S (mid-side) mike pick up pattern.
Placement of a coincident mike is not all that difficult if one keeps in mind that musicians have usually solved their own balance problems, providing that they have had at least one rehearsal at the recording location. Thus, when recording a large orchestra, a good place to start is with the mike above and slightly behind the conductor's head. With a string quartet or small chamber orchestra, choose a point equidistant from all players. With a solo instrument, slightly above and in front of the player is a good bet, but keyboard instruments, such as piano or harpsichord, can cause problems because of reflections off the lid. The amount of hall ambience can be controlled by the pattern shape on multi-pattern mikes as well as the distance from the musicians, with figure-eight mikes allowing the most ambience and cardioid mikes the least. I've given no exact placement distances because so much depends on the kind of sound you and the musicians want to hear, the acoustics of the hall, the exact instrumentation of the orchestra, and the pattern used on the mikes. What works in one hall with one set of musicians will not work in another hall (or even the same hall) with another group of musicians.
You have to learn to use your ears and make adjustments for each situation.
Often, in a live concert situation, the locations where you are permitted to place a mike are limited to one or two places that produce minimum interference (visually) with the audience (or cameras in a televised event) but do not necessarily have the best sound. A recording session situation gives you more chances to experiment, but even then there is no guarantee for success if the musicians don't understand what you are trying to do. John Woram (well-known audio consultant and recording engineer) related to me how he was once asked to make a special demonstration recording using a single stereo mike. The small group of professional musicians hired for the session were shocked to discover that they would have to control their own balances. If they wanted to hear more piano, the piano player would have to play louder; for less bass, the bass player would either have to play softer or move further away from the mike.
They had never been recorded this way before and would not believe that the recording could be made with only one mike. After all, records are made by having a mike for every player and letting the engineer figure out the balances, aren't they?
Fig. 5--The Blumlein sum and difference matrix technique using transformers.
Coincident mikes have one major weakness in that if the hall sounds bad or the musicians' sense of balance is poor, there is not too much the engineer can do about it. You simply can not fix it in the mix! Good coincident mikes tend to be very truthful in what they hear. They cannot make a bad hall with mediocre musicians sound like anything but a bad hall with mediocre musicians. But, at the same time, it is this sense of reality that makes the coincident technique so at tractive to many engineers.
Fig. 6--Circuit for Williamson's M-S matrix decoder.
Quadraphonic Coincident Mikes
There is still enough interest in quadraphonic recording for two manufacturers (AKG and Neuman) to develop their quadraphonic coincident mikes (four separate cardioid capsules mounted at 90° to each other), an English manufacturer (CALREC) to develop a tetrahedral "Sound Field" micro-phone, and for CBS to modify the QM-69 mike into the so-called Ghent microphone that produces a SQ-encoded signal. But it is not really necessary to use any of these mikes to record in quadraphonic because a true "Blumlein" figure-eight pair of mikes will produce an encoded quadraphonic signal if used in the proper manner.
A sound source located at the front or rear of the mike pair (0° and 180°) produce an electrically in-phase signal and sound sources located at the sides (90° and 270°) produce signals that are out of phase. Careful positioning of musicians around the mike and play back via the Hafler system can have spectacular results. Playback with other matrix systems will have a good quadraphonic effect, but will not recreate the true location of the players. A Blumlein recording is a worthwhile experiment for anyone interested in quadraphonics.
Coincident Microphone Discography
The manner in which an engineer uses his microphones has been com pared to that of an artist using various pigments and brush strokes in creating a painting. And just as a good artist chooses his paints and technique according to available light and subject matter, a good recording engineer selects his mikes and technique ac cording to the acoustics and type of music being recorded. However, the renewed interest in coincident mike technique has created a cult who have declared it to be "the one true way of recording" and have produced endless reams of faulty mathematics as proof.
This is the equivalent to saying that there is only one way to paint a picture. This level of silliness has reached such a level in the U.S. that a major producer of direct-disc recordings will no longer discuss publicly the details of his latest symphonic recordings, declaring that the information is a "trade secret," while, at the same time, letting it be known "unofficially" that the recordings were made using a single stereo mike.
To help the Gentle Reader arrive at some opinion on this matter, I have compiled a list of recordings which demonstrate the capabilities of the three major mike techniques. While not being, by any means, a definitive list, it will allow the reader to become aware that sonically excellent recordings can be made irrespective of the mike technique employed.
Coincident Microphone Recordings
Angel SFO-36034, Rimsky-Korsakov:
Scheherazade, Chicago Symphony, Ozawa (cond.).
Angel SFO-36035, Bartok; Concerto for Orchestra, Chicago Symphony, Ozawa (cond.).
Angel SFO-36039, Stravinsky: Petrouchka Suite & Firebird Suite, Chicago Symphony, Giulini (cond.).
Angel SFO-36040, Brahms: Symphony #4, Chicago Symphony, Giulini (cond.).
Angel S-36047, Mahler: Symphony #1, Chicago Symphony, Giulini (cond.).
Angel S-36048, Beethoven: Symphony#7, Chicago Symphony, Giulini (cond.).
Angel S-36043, Dvorák: Symphony #8, Cleveland Orchestra, Szell (cond.).
Angel SFO-36032, Brahms: Double Concerto, Rostropovich, Oistrakh, Cleveland Orchestra, Szell (cond.).
Angel S-36093, Classical Virtuoso Guitar, Angel Romero.
Angel S-36096, Bach: Italian Concerto, French Overture, Igor Kipnis, harpsichord.
All of the above recordings are the work of one recording engineer, Car son Taylor. During his long career at Angel, Carson insisted on using only one or two coincident mikes on his classical recordings. This list represents those recordings that Carson feels are some of his best works.
Sheffield Lab-7, Wagner: Orchestral Excerpts, Los Angeles Philharmonic, Leinsdorf (cond.).
Sheffield Lab-8, Prokofiev: Romeo & Juliet Excerpts, Los Angeles Philharmonic, Leinsdorf (cond.).
Sheffield Lab-3, Harry James: King James Version.
Sheffield Lab-6, Harry James: Comin' from a Good Place.
These Sheffield records are, of course, direct-to-disc recordings and have all the benefits attributed to that recording method, i.e. wide dynamic range, silent surfaces, and, in general, superior sound. The Los Angeles Philharmonic recordings are "rumored" to have been made with a single AKG C 24 stereo mike. I find the acoustics a bit too dry, but as a result the recordings are a model of clarity and definition. I also find the Prokofiev to have a much warmer sound than the Wagner, even though they were made in the same hall a day apart. I don't know if this is due to a change in the mike position or pattern or is just due to a difference in the scoring.
The Harry James records are de scribed in the liner notes as being made with a single C-24 mike (al though one or two additional mikes can be seen in the session photos), and it does not sound at all like the typical direct-to-disc big band recording made in a dead studio. These re cords realistically recreate the sound of Harry and the band in a room of just the right size.
Spaced Omni Mike Recordings
Nonesuch H-71269, Music of Edgard Varése, Contemporary Chamber Ensemble.
Nonesuch H-71291, Percussion Music, N.J. Percussion Ensemble.
Nonesuch H-71284, Gershwin: Piano Music. Bolcom (pf).
Nonesuch H-71304, After the Ball, Morris & Bolcom.
Vox QSVBX-5133, Ravel: Complete Orchestral Works, Minnesota Symphony Orchestra.
Vox QCE-31098, Prokofiev: Alexander Nevsky, St. Louis Symphony & Chorus.
All of these recordings were produced and engineered by Marc Aubort and Joanna Nickrenz. Marc believes in using a basic omni pair of mikes along with an occasional spot mike to solve certain balance problems. This list contains only a few of the many re cordings they have done over the years for Vox, Nonesuch, Vanguard, DGG, and others.
Mercury SRI-75003, Saint Saéns: Sym phony #3, Detroit Symphony, Paray (cond.).
Mercury SRI-75028, British Band Classics, Eastman Wind Ensemble, Fennel (cond.).
Mercury SRI-75030, Prokofiev: Love for Three Oranges, Scythian Suite, Lon don Symphony Orchestra, Dorati (cond.).
Mercury SRI-75057, Hindemith, Schoenburg, Stravinsky, Eastman Wind Ensemble, Fennel (cond.).
Mercury SRI-75058, Stravinsky: Firebird, London Symphony, Dorati (cond.).
All of the above are Philips Golden Import versions of the classic Mercury Living Presence recordings in the early 1960s. All these recordings were made by an engineering team led by Bob Fine who used only three spaced omni mikes. Even though some of the recordings are nearly 20 years old, they are still equal to anything being made today.
Multi-Mike Recordings
Columbia M-33523, Ravel: Daphnis et Chloe, New York Philharmonic, Boulez (cond.).
Columbia M2-33303, Schoenberg: Gurre-Lieder, B.B.C. Symphony & Chorus, Boulez (cond.).
DGG 2565 022/4, Schutz: Psalms of David, Regensburger Doomspatzen.
DGG 2530 619, Saint Satins: Symphony #3, Chicago Symphony, Barenboim (cond.).
DGG 2711 012, Bach: St. Matthew Passion, Berlin Philharmonic, Vienna Singverein, Von Karajan (cond.).
Both Columbia recordings used in excess of 25 mikes. The Ravel piece was recorded on eight tracks with Bud Graham as engineer, and the Schoenberg was recorded on 16 tracks with Bob Auger as engineer. The Saint Saens recording is a bit unusual in that the orchestral part was recorded in Chicago (on eight tracks, I believe) and then sent to Europe where the organ was overdubbed in a cathedral in France. The Schütz recording uses a combination of stereo mikes and spot mikes; and the Bach recording, as well as any piece of music that calls for three choruses, two orchestras, and six soloists, simply cannot be recorded using two mikes.
Coincident Microphone Directory
AKG Model C-422 FET, variable pat tern (X-Y or M-S), $1600.00.
AKG Model C-34 FET, variable pat tern (X-Y or M-S), $1100.00.
AKG Model C-33 FET, X-Y cardioid pattern only, $650.00.
AKG Model C-424 FET, quadraphonic coincident mike with four cardioid patterns, $1700.00.
AKG Model C-24, variable pattern (X-Y or M-S), discontinued.
The C-24 is considered by many to be the classic stereo mike. Unfortunately the last 12 units of this model came off the production line in June, 1978, and were immediately shipped to waiting customers. AKG has re placed the C-24 with the C-422 FET which uses the same diaphragm capsules, but has a modern FET preamplifier circuitry instead of tubes. The C-33 and C-34 are based upon the popular 451 microphone series.
The Bang & Olufsen BM-5 has been discontinued for over five years now, but it can still be found, occasionally, in the "used equipment" lists. This is a true "Blumlein" stereo mike consisting of two figure-eight ribbon mikes in a single housing. An extremely mellow sounding mike, the output is very low (-80 dB) requiring the use of a low noise, high gain preamp. The original selling price was $100.00, a real bar gain.
Carson Taylor's favorite mikes are the Neuman Model SM-69 FET variable pattern (X-Y or M-S), $2229.00 with the 48V phantom power supply, $166.00 if required, and the Model QM-69 quadraphonic coincident mike with four cardioid patterns, including power supply, $2303.00.
Schoeps mikes are not as well known in this country as AKG or Neuman, but they are well known in Europe for their high quality. They are the only manufacturer of O.R.T.F. stereo mike in a single housing. The O.R.T.F. method usually requires the use of two separate cardioid mikes properly oriented on a stereo bar.
The Schoeps Model CMTS-301 has three patterns (X-Y or M-S), $1585.00 and $216.00 for the power supply. The CMTS-501 also has three patterns (X-Y or M-S), and costs $1535, plus $216.00 for the power supply. The MSTC-44K is an O.R.T.F. stereo mike at $950.00, and $192.00 for the power supply. The Model MSTC-34U is an O.R.T.F. stereo mike at $1000.00 and $216.00 for the power supply.
The Studer SKS-501 mike has three patterns (X-Y or M-S) and sells for $980.00, plus $215.00 for the power supply and is manufactured for Studer by Schoeps.
The Sony ECM-99 has two cardioid pattern capsules and sells for $50.00, while the Sony Model ECM-990 with two cardioid-pattern capsules is $100.00. Both of these are electret mikes with outward-facing cardioid capsules, but spaced much closer together than the O.R.T.F. type. Al though these mikes are not as quiet or have as smooth a response as the other mikes on this list, their low price more than compensates for these shortcomings. I've used both mikes on many remote recordings with excellent results.
Although I've only listed stereo microphones, the coincident techniques I've discussed can be achieved using separate mikes of the appropriate pattern mounted so that their capsules are on the same vertical axis. Several companies make stereo mounting bars that can be used for X-Y and the O.R.T.F. methods, and there are also special adapters available for the M-S technique. The only restrictions on the mikes selected is that for best results they should both be made by the same manufacturer and be of the same transducer type, i.e. both be dynamic, condenser, or ribbon types. The same signal fed to both a condenser mike and a ribbon mike will result in signals 90° apart in phase. If these two mikes were used in a M-S technique, the results might be a little strange.
It is now possible to produce very high quality, low cost electret con denser mikes as the recent Equipment Profile by Jon Sank on the Nakamichi CM-700 mike system (Audio, September, 1978) has shown. All that remains to be done is for some manufacturer to produce a truly low cost, high quality coincident mike.
Fig. 7--The equivalent X-Y and M-S patterns.
Fig. 8-- The O.R.T.F. stereo mike setup.
above: Neuman SM-69 B&O BM-5 AKG C-422
(Source: Audio magazine, Sept. 1978)
Also see: The Compleat Microphone Evaluation--An Update (Sept. 1978)
Nakamichi Model CM-700 Electret Condenser Microphone Systems (Sept. 1978)
= = = =