Brief Discussion of Stereo Microphone Arrays for Ambient Field Recording
Like most people in the pro audio industry, I've learned over the years how to work with a handful of common techniques for creating stereo audio: AB, XY, M-S, ORTF/NOS, Blumlein, Jeklin, binaural, panning of multiple mono tracks. Each has certain strengths and weaknesses; each has its place in the grand scheme of things.
During the 1990s I experimented casually from time to time with ambient field recording using some of these traditional techniques. In 2004, after a long hiatus, I took up the challenge much more seriously, beginning with something that had worked fairly well for me in those early days: a pair of Shure Beta58 microphones that I owned, situated in an ORTF array. I also built a couple non-traditional experimental mic rigs that year. They were colossal failures. I learned a lot from them, but the memories are too embarrassing to repeat here.
At that time I also joined the Nature Recordists e-mail group, where a lot of discussion soon developed about Shure WL183 lavalier microphones. These small omnidirectional electret condenser mics seemed to be well regarded by several group participants, so I ordered some and began tinkering with literally dozens of stereo mounting configurations.
Little by little, through seemingly endless sequences of trial and error and hundreds of hours of careful listening, I began to understand some of the subtleties I was hearing from these and other stereo arrays and how they translate to both headphones and loudspeakers. The more I understood, the more I wished I could capture greater depth and spaciousness in my recordings.
In March 2006 I posed a question about this to the Nature Recordists group. Rob Danielson, a leading participant, responded with a key insight about human auditory perception that triggered an important "ah-hah" moment for me -- a critical "what if" hunch -- that soon led to a crude configuration similar to the one pictured below, using forward-facing Shure WL183s attached to small parallel wooden barriers, with the mic capsules approximately 5.5 inches apart:
...
Sample audio clips:
Sandhill Cranes :: Listen (1:00) >>>
Sandlot Volleyball :: Listen (1:28) >>>
Spaciousness and depth were so dramatically improved that I immediately abandoned all previous experiments and adopted this as my new default approach.
Unwilling to put out thousands of dollars for some of the most highly esteemed microphones on the planet, but still wanting to move toward higher quality and lower self-noise, I paid close attention when members of the Nature Recordists group began discussing AudioTechnica's modestly priced AT3032 mics. These small-diameter omnidirectional studio condenser microphones have extended frequency response characteristics, high SPL handling capabilities and very low self-noise (tested by group participant Eric Benjamin to be considerably lower than the manufacturer's published specs -- lower, in fact, than some of the most highly esteemed and high-priced microphones on the planet).
I picked up a pair, mounted them in a head-spaced parallel barrier array, and was very pleased with the results. Below are three variations on the theme:
...
...
...
Sample audio clips:
Frog Chorus :: Listen (1:22) >>>
Spring Thunder w/ Siren :: Listen (1:31) >>>
Evening Twilight Over Beaver Pond :: Listen (2:53) >>>
Freight Train and Evening Thunder in Hinckley, MN :: Listen (4:31) >>>
With arrays of this type, stereo localization is primarily a function of the setback distance from mic capsules to the leading edge of the barriers. Overall barrier size is also important -- too small and the desired barrier effect is diminished, too large and unpleasant mid-frequency "bumps" begin to appear. Distance between the microphones will primarily affect the perceived width of the soundstage; there's room for tinkering here, but roughly "head-spaced" makes undeniable good sense and delivers the best results. Any angle between the barriers, inward or outward, will also have an effect. Each of these variables overlap with and affect all the others, so season to taste... wisely.
I currently prefer this approach over coincident stereo arrays because it captures important time-arrival differences. I also prefer it over simple spaced omnis or any ORTF/NOS-type arrays because these small-ish barriers seem to dramatically stabilize the stereo image and nearly eliminate unpleasant "wavy" microphone interactions that usually accompany spaced pairs. (This improvement in clarity and transparency is an amazing side benefit that I did not anticipate!) Being highly phase coherent, these rigs also provide excellent mono compatibility. Most importantly, as I indicated above, this approach seems to capture considerably greater depth and spaciousness than any other stereo technique I've worked with to date.
Curt Olson
Minneapolis, Minnesota, USA
August 2007 (updated: April 2008)
Notes:
1. The photos above are intended merely to illustrate the basic concept and to show some construction possibilities. I offer no plans or specifications, as each particular combination of microphones used and construction methods pursued will require its own adustments. I've enjoyed hearing from a number of people who have come up with some very clever and successful adaptations of their own.
2. It turns out that a solid-body mount, such as the short piece of 4x4-inch cedar fence post shown above, eliminates certain rare and extremely subtle resonances that have been known to occur between two spaced barriers. The trade-off, of course, is added weight.
3. Counterintuitive as it may seem, cardioid microphones can also be used successfully in these rigs. Cardioids will, by nature of their design, yield a "flatter," more "compressed," less spacious image than omnis, but clarity and transparency across the soundstage are greatly improved compared to more traditional M-S and ORTF/NOS-type arrays.
4. Rob Danielson has tested some of the variables associated with head-spaced boundary and other stereo arrays, resulting in sound samples you can evaluate for yourself in this blog. In particular, there are sound samples comparing different omni mic capsule setback distances in this post.