The following is a draft of an article I published a while
back in The Sensible Sound (issue 92). I also covered the
topic with less depth in my book, The Home Theater
Companion, and much of what follows comes from my contacts
with quite a large number of recording engineers. I also
posted this article in response to a comment in another
thread.

The article draft:

Two-channel purist microphone techniques, that is, those
that minimize the amount of electrical, multi-microphone,
and mixer-related diddling required to make two-channel
recordings, have traditionally come in two different
flavors.

Some adept classical-music engineers prefer the simple,
two-microphone, spaced-array technique often used by
individuals like Marc Aubort in his work for Delos and
Centaur. Others, like Peter McGrath in his work for Audiofon
and Harmonia Mundi USA, also prefer this technique, but with
the option of employing additional support microphones. A
technique like this seems logically sensible, since with
stereo playback the speakers are themselves in a spaced
configuration.

Decades ago, Mercury elaborated on this technique and filled
in the space between the two main microphones with a third -
allowing for a more stable, mixdown-controllable image in
the center. The technique resulted in their vaunted Living
Presence recordings, still admired today for their
soundstaging qualities and sense of stage depth.

A center fill has been used to good effect for some time in
an adapted form for a large number of Decca/-London
record-ings. The highly regarded "Decca Tree," pioneered by
Roy Wallace, perfected by Kenneth Wilkinson, and used by
such Decca engineering notables as John Dunkerley, Stanley
Goodall, James Lock, and Philip Siney, consists of three
omni-directional microphones. Each is attached to the end of
a T-shaped assembly that is suspended above the conductor's
podium. The left/right capsules are about 2 meters apart and
the center unit is about 1.5 meters in front of the axis
between them. The resulting geometry imparts a quite warm
and spacious sound to the recording.

With still wider spacing, the spaced, three-microphone array
has in the past been used by engineer Jack Renner of Telarc.
Many of the classic Telarc releases were products of this
arrangement. When recording large ensembles, Renner often
employs omni-directional microphones, with the left and
right modules up to nine feet on either side of a centrally
posi-tioned one. Keeping the microphones at a
wider-than-usual distance apart minimizes the impact of
comb-filtered interference effects that some individuals
feel color the sound of spaced-array record-ings. Telarc's
most basic recording technique has long been very popular
with a number of knowledgeable listeners and record-review
critics and the company's material has become a benchmark
for many serious music lovers.

Not every spaced-array advocate likes the wide-track
approach, however. Harry Munz, who has recorded some notable
material for Gothic records (many of which have been given
rave reviews by yours truly), often employed a pair of very
high quality omni's spaced only a few feet apart. Actually,
the spacing required for best results will often be
determined by the size of the ensemble to be recorded and
the desired sound-stage spread and focus.

The spaced-microphone arrangement (sometimes supported by
ambience-augmenting microphones located out in the audience
area), whether utilizing a moderate distance between units
or a very wide one, imparts a spacious, open, and sometimes
pleasantly diffuse sound to the music when it is played back
in typical home listening environments. Many
classical-recording enthusiasts in America seem to like
this, particularly if they own speaker systems that present
a very focussed soundstage image. The technique compliments
what truly exemplary soundstaging/imaging speaker systems
such as the Dunlavy Cantatas (reviewed by me in issue 87),
Waveform MC/MC.1 sub/sat systems (reviewed by in issue 84),
and Triad InRoom Silver sub/sat systems (reviewed by me in
issue 93) can deliver.

Not everyone is enthralled by this technique, however.
Dissenters, such as researcher and mathematician Dr. Stanley
Lip****z, have in the past indicated their strong preference
for the coincident or near-coincident,
directional-microphone technique. This is sometimes used by
engineers working for Chesky, Opus 3, Nimbus, Hyperion,
Teldec, Reference Recordings, and on occasion, Delos and
Harmonia Mundi USA, as well as by a number of others.

One configuration, the XY, double-figure-8 approach
(sometimes called the Blumlein array), requires the use of
directional (front/back sensitive, side insensitive)
microphones located extremely close together and aimed at
each flank of the orchestra. By having the direct signals
respond only to level differences, as opposed to the
spaced-array's response mainly to timing differences, this
two-capsule, dipolar-microphone technique insures that the
recorded sound is kept comfortably phase coherent.

Apostles of this and several other "intensity stereo"
techniques believe that any spaced-array microphone
arrangement, be it utilizing two (or three, or more)
microphones, will add annoying, phase-dominated,
comb-filtering effects to the sound. They note that while
the reflected, ambient "hall" sound on any recording should
have a somewhat diffuse and phase-dominated quality, the
"direct" sound that is coming from the ensemble itself
should be coherent and lacking in the time-of-arrival,
phase-difference anomalies that a spaced-microphone
configuration will by its very nature produce.

They point out that piano recordings are particularly prone
to bloating and indistinct focus when recorded with a spaced
technique, as evidenced by numerous examples that give the
impression that the piano is twenty feet long or that the
keyboard is spread out between the speakers. They also claim
that coincident-source recordings have the ability to
project a valid front-to-back depth with any type of
ensemble or solo instrument that spaced-microphone
recordings can imperfectly fake - but cannot duplicate.
Finally, with small-ensemble performances, they claim that
left-to-right sound-stage imaging is going to be more
accurate with coincident techniques, particularly when
listening from the audiophile-preferred, central "sweet
spot."

Phase problems and the nature of a proper sound-stage
presentation are only part of the ongoing debate about
microphones and their placement for two-channel recordings.
Supporters of the coincident technique note that central-ly
located images (usually involving soloists) are more sharply
focussed. Spaced-array enthusiasts will counter that
although the central and near-central images produced by
their technique may be more diffuse and less stable when
listening from the sweet spot than what is possible with
coincident-pickup techniques, the negative effects can be
minimized by using the previously noted blended center
microphone. Accent microphones on individual instruments can
possibly do an even better job of leveling the playing
field.

What's more, they point out that the result-ing lack of
sweet-spot listening tightness with central, half-left, and
half-right images that can show up with spaced-microphone
techniques is subjectively no worse that what is often
encountered in a typical, live classical perfor-mances -
even ones happening in excellent concert halls. In those
live-music situations pinpoint imaging is impossible to
experience for anyone but the person conducting the
orchestra.

Spaced-microphone advocates will point out that their
favored technique ensures that under home-listening
conditions, sound-stage images will be less likely to shift
radically toward the nearer speaker if the listener moves
away from the central sweet spot, making recordings made
that way more suitable for social listening. As they see it,
coincident-source microphone recordings require that
"serious" listening be done from a small, central area if
full advantage is to be taken of what that technique offers.
Only one person gets to experience the recording at its
best, with everybody else in the room getting a substandard
sound stage.

Most spaced-array advocates feel that even though the direct
signals produced by the technique are less phase coherent
than those which result from coincident-source practices,
the net result is more subjectively realistic when
experienced in a typical home-listening environment. Spaced
speakers and spaced microphones complement each other. In
addition, a few enthusiasts claim that spaced-array
techniques, although ideally not in the same purest-oriented
class as coincident-microphone practices, allow the
recording engineer to do a more customized job of dealing
with concert-hall size, shape, and reflectivity
deficiencies, particularly with large-ensemble recordings.

Perhaps some kind of compromise is in order, and to this
effect a number of other recording engineers and astute
listeners swear by variations on the two basic techniques.
The technical people at Nimbus records, for instance, made a
point of combining the coincident technique with Ambisonic
surround process-ing. Although not very well known in the
USA, with proper decoding, this British system supposedly
can simulate a reasonably accurate concert-hall environment.
The Soundfield, four-capsule microphone used with this kind
of recording process was originally designed with Ambisonic
recording in mind.

Some experts try to compromise between the spaced- and
coincident-array techniques by using the French-pioneered
"ORTF," or the Dutch "NOS," near-coincident systems. The
former places outward-angled, directional microphones only a
few inches apart, while the NOS technique uses somewhat
wider spacing and a slightly wider pickup angle between the
microphone capsules. (The individual microphones are
cardioid models that have progressively weaker sensitivity
as the recording angle widens, until there is a near null
directly behind the capsule.) The result is quasi-coincident
behavior at lower frequencies, due to the long wavelengths
involved, with shorter-wavelength frequencies from the
midrange on up being given enough of an inter-channel time
delay to impart a degree of spaced-array openness to the
sound.

With large-ensemble recordings, even engineers dedicated to
spaced-array and/or coincident-source techniques usually end
up augmenting their main-pickup configura-tion with
addi-tional micro-phones out in the audience area to pick up
ambiance. Others may flank the main array with widely spaced
microphones, in order to highlight the left and right sides
of the orchestra or add "bloom" to the sound. The degree of
enhancement can be adjusted later, during the editing
sessions.

In addition, many engineers, even those with the purist of
motivations, use additional pickups near particularly
hard-to-record instruments or vocalists to make them more
audible in the final mix. Ironically, this is one way to
reduce the dynamic range of a recording, because instruments
that are difficult to hear with fully purist techniques -
and difficult to hear during live performances, too, for
that matter - are increased in level to make them more
audible during home-playback situations.

Practices of this kind are almost mandatory if the hall in
which the recording is being made has acoustic deficiencies.
Recording engineer John Eargle, who has made many superior
recordings for both Delos and ProArte (and who is a
specialist in the ORTF technique), is particu-larly well
known for his ability to modify orthodox procedures to
compensate for hall/orchestra anomalies. He frequently
employs level-adjustable, spaced-omni microphones well off
to the sides of his main ORTF array to add a controllable
amount of spaciousness to his record-ings.

Craig Dory, of Dorian, is also well known for his innovative
microphone techniques, as are Keith Johnson of Reference
Recordings, Ralph Couzens of Chandos, Bob Katz of Chesky,
Paul Goodman when he was working for RCA, and John McClure
when he was working for Columbia. Notable freelance
classical engineers such as Tony Faulkner, Ben Connellan,
Mike Clements, Mike Hatch, Don Hartridge, Christopher
Greenleaf, and Jonathan Stokes are also masters of purist
innovation, as were renowned Decca engineers John Dunkerley,
John Pellow, and Simon Eadon.

Needless to say, not every recording engineer is enamored of
simple, or semi-simple, microphone techniques, be they
spaced-array or coincident. Many recent classical releases
by Decca/-London have been masterpieces of advanced
multi-microphone tech-niques, and Jac Holzman, founder of
Elektra and Nonesuch, pioneered the intelligent use of
multi-microphone techniques for classical material in the
1950s and 1960s.

Deutsche Grammophon has used as many as 32 channels to
record a symphonic work, mixing their inputs down to 2
channels by deftly blending the total. Columbia and Sony
have also made a large number of multi-microphone classical
recordings. While the results of heavy-handed
multi-microphone use by these and a few other companies have
sometimes been disappointing, a fair number of such
rigorously controlled record-ings, even classical ones,
sound quite good. And of course, some pop transcriptions
have sounded downright spectacular.

Indeed, in the pop-music realm, multiple microphones and
elaborate mixing techniques are the norm. Recording
engineers such as Ed Cherney, Frank Filipetti, Chip Davis,
George Massenburg, Roger Nichols, Hugh Padgham, Bill Schnee,
Alan Sides, Bruce Swedien, and Al Schmitt are past masters
of multi-microphone usage. And of course, during the 1950s,
60s, and 70s, Bill Putnam and his Universal Recording
Company set the standard for refined pop-recording
techniques for a generation. Ironically, the use of multiple
microphones and complex mixers has allowed some pop-music
engineers to move ahead of their classical counterparts in
some respects, because those technologies lend themselves so
well to the creation of discrete-channel surround-sound
recordings.

With classical music, steely, harsh-sounding violins are the
most conspicuous indication that the multiple-microphone
technique has been executed incorrectly, because the
procedure nearly always puts the directional microphones
close to the direct field of the violins. When this is done,
the mellow, mostly off-axis blend of the sound that reaches
an audience during a live concert may not be properly
reproduced. Correctly locating a micro-phone for the best
blend can be a time-consuming job, because the unit must be
placed in that part of the direct field that most closely
simulates the reverberant sound encountered at a live
concert.

Many multi-capsule classical record-ings also come across as
one-dimensional sounding because proper timing clues are not
reproduced. Sounds picked up by microphones placed very
close to instru-ments at the rear of the orchestra are added
to the mix simulta-neously with close-miked sounds picked up
from instruments at the front. However, at a live concert
the rear-most sounds arrive later than those from up front,
which is one of several ways that we sense the depth of the
orchestra. In addition, at a live concert the sounds coming
from the rear of an orchestra tend to generate a more
reverberant soundfield and blend than what is produced up
front. This also aids the listener in sensing depth. It is
difficult to simulate these effects with multi-miking,
although individual-track, digital-delay systems can help.

Nevertheless, given that time-delay phase artifacts and
levels can be precisely controlled during the mixing
process, when placed under skillful control,
multi-microphone techniques can do a remarkable job of
simulating an "intensity-stereo" recording. Indeed, the
technique can greatly reduce many of the phase problems
disliked by coinci-dent-source advo-cates - while still
allowing the kind of final control that recording engi-neers
and produc-ers (and sometimes even conductors) can feel
comfortable with.

Whether the use of scads of microphones and extremely
complex mixing consoles will be the norm as the world of
hi-fi sound reproduction further advances into the realm of
surround sound is a question that nobody can answer right
now. It is likely that at the very least a center-channel
microphone, in combination with hall-ambiance microphones
will become mandatory, as will careful post-production
mixing. For better or worse, it appears certain that the era
of absolutely pure two-channel recording techniques is all
but over.

Howard Ferstler