Microphone frequency response and proximity effects

I start a new thread on microphones to avoid messing existing threads...
I had written:

I was about to post the same question on a microphone to record organs. However, my attention was drawn towards (a pair of) Shure SM81. These cardioids are advertised as having flat response down to 20 Hz, good for 32' notes.
Has anyone experience with them?

pwhodges replied:

Directional microphones are subject to the proximity effect, which raises the bass response at close distances. An unqualified frequency response will typically be measured at 1m or less; and a mic quoted as flat to 20Hz will usually be starting to roll off at 150Hz at a distance (this is true of pro mics too; it is a matter of physics). This is why many people prefer to use omnis for recording organs in particular. I use calibrated equalisation to correct the response of directional mics at a distance, but getting good enough info can be hard.

Paul, I have checked the Shure website, and they say that proximity effects arise when a mic is placed one foot from the sound source, or closer. If a mic response is defined at 1-meter from the source (as is said in the SM81 case), shouldn't it be free from such effects? If so, this problem is therefore relevant only for "dry" sampling at close distances, not for ambient sampling. Am I missing something?

Also browsing through the Shure site, I noticed the KSM109, whose response is not as flat as the SM81 down to 20 Hz, but is not dramatically dropping either. How can one obtain the "true" response for a mic, to apply "calibrated equalisation" post-facto as you say? Is it factory-tested and accompanies each microphone (as e.g. in research-grade optical filters)? Has someone more specific suggestions for mid-price mics suitable for organ sampling?

Thanks!

Sorry for the delay - I had a concert this weekend.

The theory of directional microphones gets more difficult the more you look at it, and this forum is not the place to go into the physics anyway. It helps a little to remember that a cardioid response is a combination of an omni and a figure-of-eight. The basic theory relates to these patterns separately. A curiosity, though, is that one effect of the different requency response characteristic of the figure-of-eight part of the response and the omni is that the directional diagram of a cardioid changes with frequency.

A figure-of-eight ribbon microphone has a response which is basically flat within its operating range; but we needn't consider this further, as that's not what you have, nor are they often used as cardioids. However, the unmanipulated response of a figure-of-eight condensor microphone falls towards the bass at 6dB/octave, and it is necessary to modify the design to keep the response level in the main frequency range, usually down to about 150Hz-200Hz for plane waves (i.e. at a long distance).

Proximity effect causes a 6dB/octave rise below a certain frequency - this frequency varies with distance. At 1m it is about 150Hz, and so compensates for the fall-off usually designed into the microphones (giving nice flat graphs, too!). For use with instruments or a voice, this is a typical operating distance, and all is fine. When used closer, as with a singer holding it, there is a substantial shelf rise in the mid bass range, which is often thought desirable, but at larger distances there is a shelf down to the lower frequency at which proximity effect comes in.

Another reason to avoid holding the response level to too low frequencies is that there is often substantial energy at very low frequencies which is emphasied by proximity effect even at considerable distances - draughts and wind noise! This is why directional microphones are more sensitive to these, and yet another reason to use omnis in a space like a church which can have substantial air movements.

Because a cardioid is the combination of omni and figure-of-eight responses, the slopes will be less than the 6dB/octave I quoted above. By how much depends on the tightness or otherwise of the cardioid.

As for getting the "real" response of a mic; if you can trust the published curves to nmean something (and they are actually quite idealised in most cases), the crucial thing to know is the distance of measurement, because everything else can, with experience, be deduced from this, as the explanation above is meant to have shown. However, bear in mind that it is really hard to measure microphone responses accurately at LF in any case.

Paul