I get the impression that manufacturing tolerances for mid-range and cheaper microphones are not always as tight as one would desire, something that’s reflected in the fact that such microphones need to be individually tested and matched if a ‘matched pair’ if desired. This made me wonder whether, if I were to try out a mic made by, say, Rode, MXL or SE Electronics and found that I liked it, and were then to go out and buy that model of mic, might the one I tested and the one I bought sound significantly different?
SOS Forum post
SOS Technical Editor Hugh Robjohns replies: The three companies you cite here are far from being in the lowest echelons of manufacturing tolerances these days, and my limited experience of them is that they are all generally quite good as far as tolerances go, if not quite in the same league as the likes of Neumann, Schoeps or Sennheiser. Different mics of the same model often do sound slightly different when compared in critical A/B listening tests, although the variations tend to be far less with the higher-end manufacturers for the reasons you’ve mentioned. Nevertheless, there are always tolerance limits in any manufacturing process, and two different mics of the same model could easily be at opposite extremes of those tolerances. Whether such inevitable variations (however large or small) are significant for any specific purpose depends on the application’s requirements and user expectations.
In my experience, close matching is only really critically important when using the mics for coincident (or near-coincident) stereo pairs and, to a slightly lesser extent, in spaced stereo pairs. Fortunately, there’s a very simple way to check for the accuracy of matching between any two mics:
1. Mount the two microphones directly above one another in a studio or large-ish, well-damped room, with their capsules as close together as possible and facing forwards. You’ll need an assistant to talk and walk around the mics, so set the mics up so that they are level with the assistant’s mouth. The assistant should remain at least two feet away from the capsules — ideally more if the acoustics allow — to ensure that both mics receive the same voice level at the same time.
2. Connect the mics to a mixer or other recording system, and pan one mic hard left and the other hard right. With someone talking directly in front of the mics (at a sensible distance to ensure both receive the same level), adjust the preamp gains to give precisely the same output from both mics. My referred way to do this is to set the preamp gains roughly for a sensible output level, then switch the monitoring to mono, flip the polarity of one channel, and fine-tune the gain of one channel only for the deepest cancellation null.
3. Reset the monitoring to stereo and remove the polarity reversal.
4. With your assistant still talking, their voice should appear to come from a narrow point in the centre of the stereo image. If different frequencies (such as sibilants) appear to flick the image towards one side or the other, the mics’ on-axis frequency responses aren’t well matched. For example, an out-of-tolerance peak in the response of one mic will pull the image towards the side to which it is panned, with the size of the peak determining how much the image shifts. A difference of 4dB is enough to move the image about a third of the way towards the edge.
5. If you have an acceptably stable centre image when talking directly on axis to the mics, ask the assistant to walk and talk slowly in a circle around the mics (facing them at all times). The voice should still remain stable in the centre of the stereo image throughout (because the two mics are facing in the same direction), although the level will fall towards the rear if you’re testing directional mics due to the reduced rearward sensitivity of the microphone polar pattern.
6. Any substantial image shifts towards one side or the other indicate unmatched polar-pattern sensitivity and/or different off-axis frequency responses. Imaging shifts will tend to get worse towards the rear, but such errors are not particularly relevant once you get beyond 90 degrees off-axis (which is the maximum limit for stereo acceptance angles in coincident pairs). However, any movement within the front ±60 degrees region indicates differences that will significantly compromise the stereo imaging if the mics are used in a coincident pair.