Topic: microphone matching

[SonicSound]

I recently purchased a new "matched pair" of microphones and wanted to confirm if they are really matched.  According to the response curve / general data sheet, the mic's were measured as follows:

Mic "A": 21.6mV/Pa = 0dB
Mic "B": 21.3mV/Pa = 0dB

I am assuming this is the microphone sensitivity but what are realistic/acceptable delta values to consider microphones "matched"?

May the schooling begin…

[Church-Audio]

I recently purchased a new "matched pair" of microphones and wanted to confirm if they are really matched.  According to the response curve / general data sheet, the mic's were measured as follows:

Mic "A": 21.6mV/Pa = 0dB
Mic "B": 21.3mV/Pa = 0dB

I am assuming this is the microphone sensitivity but what are realistic/acceptable delta values to consider microphones "matched"?

May the schooling begin…

I would say that .3 mV difference between two mics to be very VERY close. I think you also have to consider your signal chain and know that there is always going to be differences between left and right no matter how well your capsules are matched particularly in mic bodies and preamps. Even in preamps with detented gain knobs unless there is a way to do a fine adjustment Ch 1 and Ch 2 Could be way off.. So in the end response curves are important and so is relative sensitivity. You can go nuts trying to perfectly match a pair of mics.

[DSatz]

I agree with the other replies you've received so far. The difference in sensitivity is <1/8 dB, which is a closer match than you'd probably get if you bought two calibrated measurement microphones from the best manufacturer in the world.

Enjoy your luck while it lasts--these things drift over time.

The one caution I'd raise is that ~21 mV/Pa, while certainly not out of range, is about twice the sensitivity of many other studio condenser microphones. People around here--including me, even though I'm a classical guy--like to record things that GET LOUD sometimes, but many preamps and recorders (especially small, portable ones) are still designed for consumer-grade dynamic or electret microphones. Professional trappings such as balanced inputs with XLR connectors don't always mean that what's inside can handle professional microphone signal levels without distorting. Just be sure that any recorder/preamp you buy for use with these microphones can handle input signals of several hundred millivolts without clipping.

--best regards

[SonicSound]

Thanks all around!

[illconditioned]

I agree with the other replies you've received so far. The difference in sensitivity is <1/8 dB, which is a closer match than you'd probably get if you bought two calibrated measurement microphones from the best manufacturer in the world.

Enjoy your luck while it lasts--these things drift over time.

The one caution I'd raise is that ~21 mV/Pa, while certainly not out of range, is about twice the sensitivity of many other studio condenser microphones. People around here--including me, even though I'm a classical guy--like to record things that GET LOUD sometimes, but many preamps and recorders (especially small, portable ones) are still designed for consumer-grade dynamic or electret microphones. Professional trappings such as balanced inputs with XLR connectors don't always mean that what's inside can handle professional microphone signal levels without distorting. Just be sure that any recorder/preamp you buy for use with these microphones can handle input signals of several hundred millivolts without clipping.

--best regards

In many cases these "hot" microphones can be useful for recording directly into mini "flash" devices.  For example, I've had great luck running Beyerdynamic MC930 directly into mic in on the R09 (original model).  I power the mics with a (self-made) 18v phantom supply (18v through a 2k2 resistor).  This is a perfect match for the somewhat noisy front end of the Edirol.

  Richard

[illconditioned]

I agree with the other replies you've received so far. The difference in sensitivity is <1/8 dB, which is a closer match than you'd probably get if you bought two calibrated measurement microphones from the best manufacturer in the world.

Enjoy your luck while it lasts--these things drift over time.

The one caution I'd raise is that ~21 mV/Pa, while certainly not out of range, is about twice the sensitivity of many other studio condenser microphones. People around here--including me, even though I'm a classical guy--like to record things that GET LOUD sometimes, but many preamps and recorders (especially small, portable ones) are still designed for consumer-grade dynamic or electret microphones. Professional trappings such as balanced inputs with XLR connectors don't always mean that what's inside can handle professional microphone signal levels without distorting. Just be sure that any recorder/preamp you buy for use with these microphones can handle input signals of several hundred millivolts without clipping.

--best regards

Hey David, while I've got you "on the thread" may I ask how mics are mathced?

You have to match over a range of frequencies, right?

My "cheap and dirty" technique, for lav mics at least, is to put both mics side by side and record pink noise, approx 12" in front of a speaker.  Then I play back and look at the spectrum (the Waves plugin in Wavelab for example).

I've also heard of putting the capsules "head to head", with a bit of space in between so they phase cancel.  I think someone said Countryman demo their mics this way.

Anyway, what I've noticed is some mics are dead on (eg., Sanken COS11, DPA406x, Countryman B6).  Some others (eg, Countryman B3, Nevaton MCE400) sadly, vary quite widely.  They typically "cross over" somewhere in the middle, but typically one falls off faster at low or high frequencies.  I suspect this is dependent on manufacturing technique.  These are all mass produced devices in the lav market, right?

  Richard

[DSatz]

Richard, for one thing "matching" really means selection--no manufacturer (that I know of) trims or tweaks anything so that two microphones will come out better matched. Rather, the regular quality control procedure includes a precise measurement of sensitivity at 1 kHz and a frequency response curve across the audio spectrum, and capsules are then selected for the best (or a good enough) match from among the current production batch.

There's no agreed-upon or standardized definition for the maximum allowable deviation; that is up to each manufacturer. Some manufacturers' "selected pairs" may well have more difference between them than the random output of a more careful manufacturer, for all we may know.

One thing that's for sure is that the ads occasionally seen on eBay for "matched pairs" of vintage microphones are a total lie. For one thing the tolerance limits of even the best manufacturers were wider in the old days; for another, they didn't used to sell matched pairs of microphones (it's a relatively recent phenomenon) and finally, as I indicated earlier, microphone response drifts somewhat with age and use, so even a perfectly matched pair today may no longer qualify in a few years--let alone a generation or two later.

The midrange frequency response shouldn't be an issue; it'll normally be almost identical in all capsules of a given type. But the frequency extremes, and the high frequencies in particular, may need a closer look. The two flattest capsules may not also be the two closest-matching capsules; that's a matter of probability and statistics. And since at any given level of manufacturing quality the capsules generally have wider deviations than the amplifiers, as a rule it's the capsules that are selected.

--best regards

[ghellquist]

DSatz pretty much has said it. I will only add some small collected tidbits.

First of all, as already said, what the manufacturer does when "matching" mics is not defined anywhere.

I would guess that in some factories the junior janitor simply takes any two mics with consecutive serial numbers and calls them matched. He might of course check that they are the same model but you cannot really be sure about that. In this case matching does not really guarantee anything at all. But as you might expect some of these manufactures to buy the lowest price of components in batches, matching will probably improve the chance that the two mics at least use components from the same manufacturer which probably is a good thing.

In other instances the manufacturer claims that the tolerances are so small that any two mics coming out of the factory are matched. All others are discarded early on and never leaves the factory. Neumann claimed this for the small condensors for a long time, KM184 or KM140 beeing examples. Market pressure forced the US importer to put the mics together two and two and call them matched. Many years later, even the conservative Germans started putting them in boxes calling them matched. I am not at all sure that any real "matchning" is done for these mics.

In other instances there are stories of manufacturers that take lots of mics produced and carefully evaluates which two that not only measures closest but also sounds closest, not necessarily the same thing.

Others simply measure the sensitivity of the mics at one frequency and from the ones they have around select the two closest.

Bottom line is that you need to check what the manufacturer really means by "matching" mics. It most probably is a good thing, but how good is difficult to know.

My suggestion for checking your mics matching is to put them very close to each other on a stand in the middle of a large room. Now walk around the mics while recording in steroe. Talk all around the mic, whisper, jangle a set of keys (making high frequency sounds). If the stereo image wawers a lot your mics are not matched, if the sound stays square in the middle the mics are very well matched.

In my experience the best place to check if mics are matched is from the "back side", not from the front. This is where, to my experience at least, the difference is largest and easiest to hear.

// Gunnar

[DSatz]

Gunnar, the term "matched pair" is certainly open to abuse, and it actually is abused sometimes, particularly by sellers of used equipment. But for a serious manufacturer such as Neumann it isn't a casual matter or just a slogan to be thrown around without any meaning. I think that we benefit greatly from the attitude which certain manufacturers take toward their work, and I like to see that appreciated. Companies such as Neumann are still operating in a professional manner even though many of their customers nowadays aren't professionals. And when they say "matched," they actually mean something by it.

As far as I can tell, pair matching wasn't an explicit concern before the late 1950s when stereo recording became common. Before then the issue was the more general one of consistency in production, i.e. if you've used one U 47, do you then know what you can expect from another one? As Neumann began to produce stereo microphones starting with the SM 2 in 1957, they began to select the capsules for those microphones, since the entire basis of coincident stereo recording is a very close match between the microphone channels. This practice continued with their later stereo and quadrophonic microphones.

With the introduction of the K 67 capsule in 1960, Neumann's capsules themselves were now being built from halves which required selection and matching so that the synthesized omnidirectional and figure-8 patterns would come out correctly. (That had always been a problem in the U 47 and particularly the U 48.) And the multi-pattern KM 66, and later the KM 76 and 86, were made with a pair of cardioid capsule heads enclosed within a larger, screened enclosure, and those two internal capsule heads were specially selected as matched pairs. If you sent Neumann one of those microphones for repair and one of the capsules was defective, they would replace them both, and they would charge extra for their selection as a matched pair.

Similarly, by this time Neumann would provide selected, matched pairs of any of their microphones for an additional fee (or sometimes, I suspect, simply as a favor to an important client). They didn't publicize this, however; they have long been one of the manufacturers that claim their microphones to be so consistent that specially matched pairs aren't required. But very small shifts in the balance of a stereo recording can have a significant influence on how the sound is perceived, and I think that if they could economically produce their microphones within even narrower tolerance limits, they would be glad to do so.

Anyway in the late 1990s Neumann began to introduce factory-matched "stereo sets" of some models; if a match could involve consecutively-numbered microphones, that was done. Then the U.S. branch began to sell their own "stereo sets" of additional models in which the microphones weren't selected as matched pairs (something which only the factory could do) but which had consecutive serial numbers. The success of that program persuaded the company as a whole to start selling "stereo sets" of further models--and with the factory now involved, it became possible to provide both matched capsules and (quite often) consecutive serial numbers. So that is how things stand today.

--best regards

[illconditioned]

Thanks for that info, DSatz.

I think some (most?) companies (DPA and Beyerdynamic, at least) also provide frequency response charts.  These are measured only at zero degrees incidence (ie., front of the microphone), but at least you can see visually how close the microphones are.  I also saw (on a DPA chart posted here) a difference plot for the mics.

It seems reasonable to get a matched pair, at least from the major manufacturers.

  Richard

[stevetoney]

While I understand and respect all the input, I'd really like to hear what people have to say about why they feel that matching of capules is really worth my concern, or whether it's kind of a red herring subject.   

I've given this a bit of thought over the past couple years.  I guess what I've wondered is that, if preamp bodies of a mic pair aren't similarly matched, and I've never specifically heard that preamp channels are always precisely matched, not to mention the other devices in the chain such as the ADC, then why should we be concerned if a capsule pair is or isn't matched? 

Not only this, but as Chris Church pointed out earlier in this thread, we twist/turn/tweak away on the knobs of each channel separately anyway, in effect un-matching the channel response to a specific sound source, so what difference does a couple of db make if the capsules aren't matched precisely? 

Finally, with the great dynamic range offered by enhancements in recording technology (24bit word length), it seems to me that perhaps there's no real need to have to squeeze every single db out of a microphone/preamp/ADC signal chain in order to get the hottest/best sound possible, like some might have wanted to in the days of 16bit recording.

I'm not suggesting that capsule matching isn't a worthwhile service being offered by manufacturers (especially when clients are asking for and willing to pay a premium for matching), but I've had the above thoughts for awhile and I'd really like to hear input from people about why they feel that capsule matching is worth our concern, in light of the above factors.

[stevetoney]

I don't think it has anything at all to do with use or abuse of dynamic range, it's about integrity of a stereo image.  It's true that variations in preamps can cause an issue (slanting of the stereo image to one side or the other), but preamps aren't subject to the same sort of variation in frequency response that transducers are.  Thus, even if you cannot calibrate your preamps, if you have a closely matched pair of microphones you can use the microphones to calibrate the recorded stereo channels in post-production (it helps to generate a dead center test tone at the start of the session, but it can be done through analysis as well).

I am guilty as any at the preamp problem, given that I produce a preamp with independent channel pots.  Even so, visual matching of the gain setting results in a 0.3dB tolerance, not too bad I don't think.

Yeah, I understand about the dynamic range portion of your response.  When crafting my question, I was going through a mental checklist of all the reasons, perhaps, of why channel matching might have validity.  So, let's scratch preservation of dynamic range from the discussion.

OK, your response makes sense and I also see how matching the frequency responses to get two capsules with graphs that parallel each other will help to keep the stereo imaging consistent for all frequencies.  (In my earlier response, I was focused on the mic sensitivity aspect of matching and not the frequency response aspect.)

Question:  Can the stereo imaging change alot from DFC calibration between unmatched vs. matched capsules such that there would be very obvious changes in imaging? 

I ask this with consideration of the idea that I'm not sure that I would care that much if my audio image put, say, a kick drum in the absolute center of my soundfield versus ever-so-slightly shaded to one or the other side in the image.

Having said this and somewhat answering my own question, I suppose the issue in matching is getting accurate reproduction of the stereo image, so whether the stereo image shifts a little one way or the other in the soundfield due to un-matching, the point is that matching would tend to eliminate that from happening. 

[ghellquist]

Same sounding mics in the two stereo channels is indeed one of the aspects that help to create a good and stable stereo image. It makes much more sense when you record acoustic instruments -- recording sound from a PA is already artificial.

What you can experience when recording acoustic instruments with mics that differ too much is that the position of the instrument may wander in the stereo position depending on what tones it plays. A low tone on the cello might come from center while higher tones (in frequency that is ) might come more and more from the right. So in my mind, if you plan on hiring the London Symphony orchestra for a session then you should use well matched mics, if nothing else because the extra cost incurred is tiny compared to the session cost.

You can actually try this yourself very easily by using two different mics in a stereo recording. I once had to resolve to using a cardioid on one channel and an omni on the other, and the result did come out useable.

As has been said ( as always DSatz has a lot of background info ) , high end mic manufacturers do make mics that are generally good enough without any extra selection steps such as matching. When you ask them to do it they will do a respectable job of matching your pair.

So the interesting effect is that when you stay with these top end manufacturers there is not really any need for a normal user to order matched sets, but they actually do a good job when selecting the mics for you.

On the low end of things, you can never be sure. Have you ever really understood what the low-price manufacturers does when they sell you a "matched set" of mics. Select consecutive numbers or anything more? And yet, these mics are where things really differ so here is where matching really would make a difference. I know that in older days, as an example, the Octave mics from Russia often needed to be matched, and importers in many countries did this as a service. Quality has gone up in Russia so the selection is less needed now.

The large difference is bound to be in capsules. Making capsules is still a bit of "art" as they are mechanical things that all differ. The electronics is in that respect easier - use low tolarance high quality components and they all pretty much comes out the same. The problem though is that low priced manufactures has to use low-quality, high tolerance components so even there they might differ a lot.

Bottom line:
 - buying high quality mics : they are good enough, no matching needed but the factory does a high quality job in selecting pairs. My suggestion is that if you plan to buy the likes of Schoeps, Neumann, Sennheiser, DPA, Microtech Gefell ( non-exhaustive list ... ) then pay the small premium for a matched set.
 - buying low price mics : they can differ a lot but what the factory does to match is a mystery so if they are matched in sound is anybodies guess. So my suggestion is not to pay anything extra for matching.

// Gunnar

[SparkE!]

I've got to agree with Gunnar's "bottom line" conclusions on this, but I'd like to add a bit of my own opinion as to why those are the correct conclusions.  Unfortunately, we tend to find explanations for our observations that may fit those observations, but are nonetheless based on junk science.  We humans like to be able to explain everything we encounter, even if we don't fully understand what's going on.  That's not to say that our badly founded explanations are not worthwhile.  Usually there is a strong correlation between our explanations and our observations.

That being said, I agree with DSatz that the concept of matching mics became important when stereo recording became more common.  The easiest thing to match is sensitivity, and that is usually done by measurements at the single frequency of 1 kHz. That is usually a reasonable surrogate for matching the amplitude response at all frequencies. What's ironic is that amplitude plays a relatively minor role in how we hear stereo images.  It is actually differences in the time of arrival of audio signals between the two stereo channels that most strongly affects the stereo imaging.  So insisting that the phase response is consistent from mic to mic is actually more important than matching the amplitude response.  Fortunately, the phase response of mics generally doesn't change significantly from one mic to another mic of the same design.  So in a weird way, it actually is amplitude response that is important.  Phase response kind of takes care of itself without matching.

Thank god for simple results!

[stevetoney]

^ Not only that, but most people on this list are hobbyists recording in settings where very few of the variables leading to the end result can be controlled.  In those settings, I can't see where a matched capsule pair would ever need to be a requirement.  Granted, all things being equal, it would be nice to have a matched pair though.