----- Build your last loudspeakers yourself -----



What's new



Stereo Recording and Rendering

Audio production



Room acoustics

Issues in 
speaker design












dipole speaker

    Dipole models

    Active filters



Three-Box active
system (1978)



& Room












Sound recordings


Other designs

My current setup

About me

Site map









Users Group



Digital Photo



What's new



Recording & Rendering

--- Recording & Rendering 101 --- Acoustics vs. Hearing --- Subjective evaluation --- 
--- Room optimized stereo --- Sound reproduction --- Recording what we hear ---
--- Experimental results --- Theory --- SRA --- Sound field control --- 



Experimental results

Don Barringer reports:

Initial experiments with the previously described four microphone array have resulted in a change to the front pair, with the main objective being a more synergetic or complementary mix with the rear microphones. 

The Mk 41 supercardioid capsules have been replaced with Mk 4 cardioids, and the separation between them reduced so that the pair is now in ORTF configuration (an included angle a of 110 degrees with 17 cm separation D1). The ORTF technique has been proven over decades and under a wide variety of conditions. Schoeps participated in its initial realization and manufactures the ORTF stereo microphone, as well as an ORTF stereo mounting bar allowing flexible use of the components, which I use here. All four microphone amplifiers are the CMC 6 "linear" type.

The rear omnis, LA and RA are mounted vertically, pointing to the ceiling, due to the increasingly directional high frequency polar response of the MK 2S capsules. The roughest - and smoothest - response is thereby relegated to the vertical axis; the horizontal axis, whatever its imperfections, remains uniform through 360 degrees. In a perfectly diffuse field, this orientation would not matter, but in locations where both direct and diffuse properties are expected, the horizontal axis was thought to be the more consequential. In a search for possible improvement, an experiment was conducted replacing the Schoeps with Earthworks TC20 omnis which have a better polar response due to their small 7mm diameter design. That theoretical advantage proved difficult to discern, probably owing to this particular application, although they remain a viable alternative. The rear microphones were placed a distance D2 behind the cardioids and separated by D3, where D2 = 9 m and D3 = 12 m. These distances are under investigation and may be reduced in the future.


The mixing sequence

An integral part of this recording technique is to record to 4 channels so that combining the microphone pairs can be done post session under optimum monitoring conditions (in this case the Orion++ in my living room). The absolutely critical nature of this requirement is addressed in detail below. I use the Edirol R-4 (battery powered) for recording, and Adobe Audition 1.5 software for processing the recording.

During mixdown, the low frequency roll-off of the cardioid mics is equalized flat to around 20 Hz with a shelving low frequency equalizer. I found that using the MK 4ís published frequency response curve to visually derive the inverse curve was preferable to setting it solely by ear. The results now sound correct, but further experience would be welcome here. The software provides an ample number of turnover frequencies with levels available in about 1 dB increments.

Experience has shown that it is also necessary to have high frequency shelving equalizers available for the rear omnis. The Schoeps MK 2S mics have a high frequency rise of 4dB on axis, intended for use in somewhat diffuse fields. Following their use in a recent session at Washington National Cathedral, I discovered that an additional boost of 2dB at 7 kHz made a significant overall improvement. The cathedralís lengthy and memorable reverberation was helpful in identifying this effect. As a result of this experience, I remixed a previous session at a small church that used TC20 omnis (flat in the direct field), and arrived at different settings: from +1dB to +4dB at 8 kHz, and from -5dB to -3dB for the level relative to the front mics, confirming the influence of this detail on the overall scheme. 

Whatever theories may be at work here, I think the predominant one arises from the very nature of this array. Recall that the pickup from the rear omnis is intentionally unrelated to the front pair, yet importantly related to them when combined. For optimum integration of these two very different sonic scenes, it appears that the rear scene needs to spectrally blend with the front one at the upper frequencies. Experience to date suggests that each recording session has to be treated as potentially unique. If adjustment is required, as it probably will be, shelving hf equalizers appear effective at shaping a compatible response. There is sufficient flexibility in the software for this, and 1 dB level increments are available. 

The most critical processing stage of course is combining these two stereo pairs. The results are a novel aural experience and a lengthy period of learning and acclimation is necessary to become familiar with this uniquely interactive behavior. My initial impression was that there is no precise "correct" level for the rear mics but instead a permissible zone of several dB. After taking into account the influence of the omnisí high frequency response, I learned that the zone for "correct" level of contribution from the omnis can narrow to less than a dB. (Available level increments are several tenths of a dB.) As a guide, their level will be at least several dB below that of the main pair. Also, this level adjustment should not be viewed as a forward/backward control. The entire array must be moved to achieve that effect.

The mixing sequence is:

  1. Apply low frequency eq to the front pair,
  2. Correct for any left to right imbalance within each pair,
  3. Add the omnis to the front pair, and begin the learning process,
  4. Choose a preliminary level setting, and then experiment with the hf eq of the omnis,
  5. Alternately recheck the level and hf eq until satisfied with the result.

The only reason to listen to the rear omnis separately, is to check channel imbalance. I have not found it a useful condition at all for adjusting the hf eq.


Microphone placement

Placing the microphones is a relatively uncritical procedure, and starts with locating the main pair, which sets the audio stage. The ears serve as an excellent guide in this process. Walk backwards from the musicians until a balanced and reasonable representation of the entire ensemble is heard. Put the stand there. In my experiments, I have not in retrospect desired a closer placement, but could on occasion hear the possibilities of a more distant one. This placement is not precise; it is a matter of meters. Now, moving backwards 9 m, the rear omnis were placed 6 meters to each side of this point. I did not find that height is critical. The first experiments were done with the mics at ear height (standing), and subsequent higher positions did not affect fundamental impressions. If the mics are raised Ė particularly for large ensembles Ė 3 or 4 meters, then perhaps some additional clarity will be gained that obviously canít be heard from the floor. Experience will have to accumulate to anticipate the quality and any importance of this effect.



An audiophile vocabulary has been accumulating for many decades, and it mainly owes its existence to electrical, mechanical, and acoustical artifacts found in the reproduction of music. So what use is it when these artifacts are not present, such as in live music? The vocabulary becomes irrelevant when the ears and brain, no longer vexed by unnatural phenomena, are able to resume a more familiar function. That is where we are in discussing this mic technique. The acoustic event heard in the vicinity of the main mics at the session is what is heard at home. The reality and the illusion appear too much alike to notice dissimilarities. The vocabulary used to describe live music applies, in this case, to its reproduction as well.

Though related to the above, I noted as a separate impression the absence of distractions. I was habitually listening for imperfections, it seems, but none stood out, leaving only the performance. This suggests that whatever imperfections exist, they are perhaps inaudible or at least easily ignored under the circumstances.

There is a strong you-are-there impression and the importance of the stereo seat has declined. It is no longer an exclusively rewarding location. I find that I am no more disturbed by moving off axis at home than I am moving to an equivalent position at the recording session. Here again, the illusion during reproduction matches the reality.



Accurate monitor loudspeakers are indispensable. Without them, how is a recording engineer to know what is possible, what the true effect of his every decision is, what accuracy sounds like in stereo reproduction? Without accurate monitors, the approach described here would not have occurred.




What you hear is not the air pressure variation in itself 
but what has drawn your attention
in the two streams of superimposed air pressure variations at your eardrums

An acoustic event has dimensions of Time, Tone, Loudness and Space
Have they been recorded and rendered sensibly?

Last revised: 06/28/2014   -  © 1999-2014 LINKWITZ LAB, All Rights Reserved