Topic: Matching Mics

[ol' dirty taper]

I have three, same model, mics with these different frequency read outs printed out from the factory. Based on the reports, I see two I would pair (1 & 3), but then looking at the frequency chart, the peaks are wildly different on the high end.
Reading a bit I saw you can record the same source, invert phase/mono then playback to see how well a pair they truely are, but can you pair based on these frequency charts?

Mic 1:
Sensitivity (1 kHz) : -30.5 dB re 1V/Pa
Equivalent to : 30.0 mVPa

Mic 2:
Sensitivity (1 kHz) : -32.0 dB re 1V/Pa
Equivalent to : 25.0 mV/Pa

Mic 3:
Sensitivity (1 kHz) : -30.4 dB re 1V/Pa
Equivalent to : 30.1 mV/Pa

[Gutbucket]

There are a number of different parameters that can be matched.  The primary ones used to match microphones are:
Sensitivity
Frequency response
Phase response (we'll ignore this one)

Sensitivity difference is easiest to correct for. Just apply a bit more or less gain to one to match the other. Checking and correcting the stereo balance of the resulting recording should be part of your processing chain anyway. However having to correct for extreme differences in sensitivity might lead to noise or overload problems.

Frequency differences are more challenging to correct.  You can EQ by ear so one side sounds like the other, or do something more involved like place both mics immediately adjacent to each other and record noise so as to to derive two response curves, then use that to correct one to more closely match the response of the other.  Some EQs incorporate a curve matching function that can be used to do this far more precisely than by ear.

Because of all that, I'd typically choose the microphones that sounded closest in terms of frequency response.  Then adjust recording gain to get similar recording levels in use, and tweak stereo balance further if needed in post.

If you go the curve matching EQ route, that is able to modify both frequency response and overall level (which corresponds to sensitivity) at the same time. You'd need to save the correction curve and mark the mics so that you always apply the correction to the correct microphone. You'd then always apply the resulting correction curve to that microphone channel as an initial step of your post processing.

You could even take that a step further if you wanted to correct or modify the overall response of both microphones used together as a matched pair.  In that case you would first manipulate the response curve of the microphone you are matching the other one to, then match the second one to that, and apply the resulting EQ corrections to both channels.  But tread carefully there because any change you make to both mics becomes "baked in" and its no longer just about matching one to the other, but about changing the overall character of the mics.

[ol' dirty taper]

Gutbucket,

Thanks you for the thorough response. Mics 2 & 3 have the closest frequency response, very similar, just a 2 dB difference. After reading your reply a couple times I have settled on keeping those as a pair and adjusting the gain as you suggested. Thanks again.