Introduction to Computer Music: Volume One

3. Microphones | page 5

Stereo and multiple microphone techniques

A real challenge to a composer, who may not have the same skills and training as an audio engineer, is to use more than one microphone to create a coherent sonic image. Pitfalls abound — an image with a "hole in the middle," weird frequency cancellations from time/phase issues, too much or too little room ambience. I list some tried-and-true stereo techniques below in the hopes that they will help the composer to decide what technique best fits his/her need.

The first documented stereo mic'ing is attributed to Clement Ader at the Paris Electrical Exhibition in 1881, using the fledgling telephone equipment of the time. More seriously researched efforts using spaced microphones are attributed to Harvey Fletcher of Bell Labs and Alan Blumlein working across the pond at EMI in the UK. (In fact, Blumlein's contribution to stereo were so impressive, it is a wonder we don't just call it Blumlein — but we at least have a mic pattern called a Blumlein pair.) Both worked on their mic'ing schemes in the 1930's, and both schemes had their pluses and minuses.

Fletcher's method used spaced omnis, where the two omnis are placed several feet apart in front of a sound source, such as an orchestra. Because the omnis will pick up both the source and any reflected ambience, the closeness to the source determines how much of the room environment is picked up. Differences in the time and phase a source sound strikes each microphone gives it its stereo directional image. This method is also called Time Difference Stereo or A-B Stereo. The use of omnis rather than directional mics such as cardioids allows the pair to have less loss of low frequencies at greater distances. Often you will see a pair of spaced omnis hung further back in a hall. A 3:1 rule — suggesting that for every unit of distance from the source to the mic, the mics should be spaced 3 units apart — helps to avoid some phasing problems.

A completely different approach was taken by Blumlein and then followed by a huge number of variations. Rather than using the time difference between mics to produce a stereo image, Blumlein researched a technique in which microphones were placed very near, or in some cases, directly on top of each other. This is called coincident mic'ing. If you remember the psychoacoustic mechanism for directional cues discussed in Chapter One, both time and relative amplitude give us cues. Blumlein discovered that by using only amplitude differences between speakers, he could create the illusion of timing differences and thus a robust directional image. By using directional patterns, such as cardioid mics, the off-axis attenuation provided the amplitude differences, as illustrated on the Flash example of page 3.

Some of the following is repeated from the information on page 3 so that readers may print it out. (Flash examples normally don't print.)

The most common coincident pairs are:

X-Y coincident pair: two cardioid mics aimed across each other at an angle between 90 and 135 degrees and less than 12 inches apart to recreate accurately the way a listener hears with directional cues. (A distance of more than 12 inches apart creates phase cancellation problems.) This pattern is very useful for many situations, but it may not provide as wide a stereo image as some other techniques. The image, however, is extremely mono-compatible, and that is why it was very popular in the radio/television broadcast world.

Near coincident pair: similar to X-Y coincident, except the microphones face outward at 90-135 degrees. One of the most well-used near-coincident patterns is the ORTF (Office de Radio Television Francaise), where two cardioid microphones are spaced 17 cm (6.7") apart and are angled outwards 55° from each other (110° total). Again, this may not provide as wide a stereo image as spaced omnis, but it is excellent for smaller environments and is also mono-compatible. A variation of this is NOS (Netherlandshe Omroep Stichting) in which the mics are angled at 90 degrees outward but at a distance of about 12" (30 cm), which the Netherlands Broadcasting System believes captures more ambience than ORTF, but still has fewer phase problems than a widely spaced array. Other variations (usually small distance and/or angle differences) you may come across are RAI (21 cm, 100 degrees), DIN (20 cm, 90 degrees), Olson (20 cm, 135 degrees).

Both coincident methods capture less of the room ambience than spaced omnis, but often this is a good thing. Stereo bars are available for ease of set-up in spacing and angling the mics.

A Blumlein pair uses two coincident figure-8 microphones at a 90-degree angle. The microphones are sensitive to the room ambience as well as the signal, so differences in distance from the source are critical. As shown on page 3, sounds from the rear image on the opposite side.

Decca tree: More complex, multi-microphone methods for optimum recordings in concert halls exist, such as a Decca Tree, developed by the recording company of that name for orchestral recordings. A Decca Tree bar has three omnis. The outer two omnis (could be cardioids as well) are positioned about 2 meters apart. The middle microphone may be closer to the ensemble and may have difference directional characteristics. By controlling the relative level of the side vs. middle mic(s), more or less ambience and directionality can be achieved during mixdown. Additional variations (shown to me by Konrad Strauss, the I.U. Recording Technology director) involve cutting down the omni field of the outer mics with acoustic baffles. Having had my orchestral music recorded with a wide variety of techniques, my own personal choice for orchestral recording would be a Decca Tree.

Another recent, popular technique, particularly with the ease enabled by digital consoles and digital audio workstations (DAWs) is M-S stereo or Mid-Side stereo. A cardioid mic is positioned on-axis (directly facing) the sound source and a figure-8 mic is positioned sideways to the source. The figure-8 is brought in to two channels of the board, one of which has its polarity reversed. Each of these channels, when mixed with one of the figure-8 channels produces either a left or right image. Again, when mixing down, the degree of separation can be controlled by varying the amount of signal from either mic. Outboard M-S decoder boxes or M-S plug-ins for DAWs are available as well.

Spot mics are used in multitrack recordings to allow engineers to feature a particular instrument or instrument group. Several of the problems that crop up with spot mics are time/phase differences from the house mics (this can be easily corrected with digital editing), differences in ambience (artificial reverb can help) and bleed from other sources that you didn't want to highlight (acoustic baffles can help). There is also the problem of mechanical transmission of sound from foot tapping or stage vibration, so shock mounts are helpful. Even correcting for these difficulties, this author has experienced 'harshness' and other disagreeable issues when using spot mics for all but limited use.

Outriggers (hung off to the sides parallel to the center array) are sometimes used in multitrack recording to pick up signals from a wide ensemble such as an orchestra or choir when using a narrow center array such as a Decca Tree.

Newer configurations for recording 5.1 surround images are evolving all the time, such as the Fukada Tree and OCT patterns, with 5 and 4 cardioid microphones, respectively, and additional imaging accomplished at the mixing console. See Mark Ballora: Essentials of Music Technology p. 192-197 for further details.

1 | 2 | 3 | 4 | 5 | 6

| Jacobs School of Music | Center for Electronic and Computer Music | Contact Us | ©2013 Prof. Jeffrey Hass