Topic: Schoeps matched pair of caps vs non?

[Cobiwan]

Thanks Scooter! I hadn't even considered the coolness factor when asking about sonic performance/difference between the two. I already own matched MK 4s, so I guess that answers my question for when I buy another set of caps.

[DSatz]

To take this out of the theoretical and into the real world: Take three minutes from any good stereo recording that you're familiar with, bring it up in an audio editor and save out those three minutes. Now make a copy of that file, and then in the copy, change one channel but not the other--say, reduce its overall level by 1 dB, reduce the bass by maybe a dB, and raise the treble by a dB. Save that file, then compare the two recordings. If they sound the same to you, you don't care about pair matching.

(If you use microphones from manufacturers that aren't among the very top handful, you might want to double or triple those differences and try the experiment again. A "matched" pair from one of the big Shanghai factories might well have greater differences between capsules than the worst combination possible for an entire production year from Schoeps or Neumann.)

Note: Don't expect to hear the direct result of the changes you made. Just listen, then switch back to the unaltered recording and listen again--follow your subjective impressions. Our ears and brains respond to differences between channels as small as ~1/4 dB in a stereo recording, not by registering them consciously, but by perceiving different subjective qualities in the overall recording. And 1/4 dB is more stringent than the pair matching criteria of any microphone manufacturer (or loudspeaker or headphone manufacturer, for that matter). I therefore believe that any improvement in microphone consistency is to be welcomed for two-mike stereo recording methods--provided that you like the microphones in the first place! (Having two identical, shitty-sounding microphones isn't a whole lot better than having two different, shitty-sounding microphones.)

We're only a tiny slice of the overall microphone market, though--and that's why pair selection/matching exists as an extra-cost option (in the case of Schoeps and, if I recall correctly, Microtech Gefell) or as part of certain sets (in the case of Neumann). It's not worth the extra cost for a studio that's buying a dozen or so microphones to use in random combinations, where each mike's signal will be going through a fader and EQ on a mixing console anyway.

But it is quite possibly worth it for people who have only two microphones of any one kind, and who use them as coincident or closely-spaced pairs for stereo recording.

--Something to think about, for the philosophically minded: The "vintage" microphone crowd takes a contrary viewpoint. One of the reasons they revere the old, manual production methods (in addition to romanticizing what really came from a dire lack of funds, especially in the ten years or so following World War II) is that those people almost never use microphones in pairs, and manual production inevitably leads to greater variation--which then allows those people to each pick their favorite from a wider array of possibilities than would be produced today. And no two aficionados of "vintage" microphones would pick the same favorite U 47, C 12, M 49 or Ela M 251 E from any batch of twenty.

[stevetoney]

To take this out of the theoretical and into the real world: Take three minutes from any good stereo recording that you're familiar with, bring it up in an audio editor and save out those three minutes. Now make a copy of that file, and then in the copy, change one channel but not the other--say, reduce its overall level by 1 dB, reduce the bass by maybe a dB, and raise the treble by a dB. Save that file, then compare the two recordings. If they sound the same to you, you don't care about pair matching.

Hmmm, some additional thoughts and questions. 

I get your point about A sounding different from B, but I don't get why it should matter (at least most of the time). 

Again, my thinking is that unless you've got control of the source, which you may have in a studio environment, you may not (or should I say probably won't) have balance anyway at the point of capture.  For example, what difference would it make if I have well matched capsule pairs when I'm recording imperfect sources?  Like for example when you've got a sound guy in the back of the room tweeking knobs changing the EQ or the pan by at least that same 1 db in your example above every few minutes throughout a show?  Or say you're recording an orchestra acoustically where there's an extra clarinet on the right, so maybe the clarinets sound 1db louder on the right than those on the left?  Similar thing, isn't it?  Either way, if this is a concern to me, seems like I'm doing the work in post to EQ one of the channels regardless of how good my capsules match up.

My point here is that I understand that you'll hear a difference between the before and after when you moved your EQ 1 db in one direction.  But that's just because you've altered your reference point so that A sounds different than B.  Different in sound doesn't necessarily equate to worse in sound, it just means it's different...just like it would sound different after the sound guy tweeks the knobs during a song. 

I guess I just don't understand why the recording made by a matched pair would be considered better than a what you get with an unmatched pair, just because it may sound a little bit different, unless you're really concerned about accurate imaging.

[DSatz]

tonedeaf, yes, the stereo imaging is one place where the effects of a greater mismatch are likely to be especially noticeable. But it's not the only one, since even with only one channel boosted only 1 dB, your average levels will have increased enough that some listeners would typically report hearing more detail in the recording ("I heard things in the music that I never noticed before" or "It was just a shade clearer" or "I could follow the individual voices and instruments better"). That's the old hifi salesman's trick; it doesn't require an increase to both channels at the same time. The playback that is just slightly louder will sound better and not louder.

If you're willing to accept more variability than the minimum necessary, knowing the effect it can have on the impressions made by your recordings, there's no argument that anyone can possibly make against that. Also, if you have the opportunity to set up your mikes during a rehearsal or warmup, and test your setup with the same live music that you're going to record for real later on, you can even tweak some of the variability away--especially where differences in gain and sensitivity are concerned.

For that matter, in any given pair of microphones one capsule is (at least marginally) more sensitive than the other while one amplifier has (at least marginally) less gain than the other. So if you pair the right capsule with the right amplifier, you'll come closer to having identical sensitivity between the two microphones than you would get by combining the capsules and amplifiers the other way around. If you can find out or figure out which is which, that will help even (or maybe especially) if you don't have matched capsules.

--best regards

Edited later to add: What I should have mentioned up front is that in a number of respects, our brains are more sensitive to the difference between what our ears are hearing than they are to what both ears are hearing in common. For example, directional microphones often have rather twisted phase response, but that generally passes unnoticed because both microphones in a stereo pair have phase response that's twisted very similarly to one another.

[stevetoney]

Thank you.  This is an interesting subject.

[Cobiwan]

Dsatz,  I was curious how you can check response without sending to the factory for a checkup. What equipment/tools are needed? I won't pretend I know much about polar patterns but was interested in learning more about "matching" responses.
Thanks for all the great input in this thread; it is truly enlightening.

[Gutbucket]

How about doing a null-test and looking at a frequency plot of the differential?  Probably would need to run a base-line test to check subsequent tests against.  If the results of a later test are significantly different that would indicate that either the microphone responses have diverged since the base-line test, or that the test setup is inconsistent.

The same could also be used as a way to check how close the responses are of any two microphones. Without a clear 'known-good' base-line you wouldn't know to what degree they may have diverged from how close they once were to each other, but it would still be useful in comparison to the results of other well-matched pairs.

I've heard of recording engineers doing something similar in setting up for 'purist' two-channel stereo recording sessions as a way of matching levels as closely as possible through the entire recording chain, except they just listen for best differential null while adjusting gains, rather than of looking at a plot of the response of the differential.  A plot would show where the differences lie along the frequency spectrum.  The basic proceedure is as follows- Recording engineer sets up the pair to be as coincident as possible with no angle between them.  Have something make noise to generate signal.  Reverse polarity on one channel and sum them.  The result is the differential between the two.  The engineer adjusts gain of one channel relative to the other while listening for the greatest degree of cancellation of the signal.  Once that's done she flips polarity back to normal on the inverted channel, reconfigures the microphones in the desired stereo configuration, and retains the gain difference setting throughout the session.  That proceedure compensates for sensitivity differences between the two microphones as well as all gain differences through the rest of the recording chain up to the point of summation, but doesn't correct for frequency response differences.  We could do the same and plot the response of the differential to see the frequency difference. 

[Gutbucket]

[snip] what difference would it make if I have well matched capsule pairs when I'm recording imperfect sources?  Like for example when you've got a sound guy in the back of the room tweeking knobs changing the EQ or the pan by at least that same 1 db in your example above every few minutes throughout a show? 

That's a seperate problem.  This could be thought of as the potential to more perfectly record those kinds of imperfections.  Your point questioning the relevace of that capability is resonable, but the same argument can be and often is made about the quality of the microphones themselves, regardless of how finely matched they are.  Does one require Schopes instead of lesser quality microphones (well matched or not) to record crappy bar PAs?  Similar question in essense.

Or say you're recording an orchestra acoustically where there's an extra clarinet on the right, so maybe the clarinets sound 1db louder on the right than those on the left?  Similar thing, isn't it?

 

Not the same at all.  You wouldn't want to compensate for that.  Those differences are what makes it a stereo recording.  The capability for resolving those things accurately in the recording is the reason a well-matched pair is relevant.

[stevetoney]

[snip] what difference would it make if I have well matched capsule pairs when I'm recording imperfect sources?  Like for example when you've got a sound guy in the back of the room tweeking knobs changing the EQ or the pan by at least that same 1 db in your example above every few minutes throughout a show? 

That's a seperate problem.  This could be thought of as the potential to more perfectly record those kinds of imperfections.  Your point questioning the relevace of that capability is resonable, but the same argument can be and often is made about the quality of the microphones themselves, regardless of how finely matched they are.  Does one require Schopes instead of lesser quality microphones (well matched or not) to record crappy bar PAs?  Similar question in essense.

Or say you're recording an orchestra acoustically where there's an extra clarinet on the right, so maybe the clarinets sound 1db louder on the right than those on the left?  Similar thing, isn't it?

 

Not the same at all.  You wouldn't want to compensate for that.  Those differences are what makes it a stereo recording.  The capability for resolving those things accurately in the recording is the reason a well-matched pair is relevant.

I don't disagree with anything you've said above, but your response doesn't address the point I was trying to make.  I'm sure my examples weren't good examples.  I'll try again to summarize better my thought process...

Lets say we have two pairs of the same model and brand of high quality capsules on the same mic stand.  (Lower end mics don't do capsule matching, so this discussion really isn't relevant to any brands but those that offer capsule matching).  One of the pairs of capsules is matched.  The other pair either aren't matched or they've changed over time.  Lets also say that the unmatched pair are well matched on the lows, 1 db off on the mids, and again well matched on the highs.

Now let's say we're recording a live performance without any PAs.  You're in the sweet spot and you're getting a well balanced sound from that spot from the sound coming from on-stage.  On your recorder, let's eliminate the variable of differences in levels...all I want to focus on is the difference in frequency response matching, not differences in sensitivity matching.  So for this situation, the levels of all four channels are normalized.

OK, so the input conditions are defined for this theoretical 'experiment'.  The questions I'm seeking to answer are:

Questions
Q1: What do you hear on your recording?
Q2: What differences do you hear?
Q3: Do the differences you hear matter?
Q4: Are the differences you hear good, bad, or neutral?

Answers
A1: Since we're using the same brand of capsules, all four channels are going to sound the same in the holistic sense.  The only differences you hear are that the mids on the one of the channels of the unmatched pair will be 1db different in the mids from what the mids sound like on the matched pair. 

A2: The effect of what you hear on the recording is that the mids will sound closer or farther away from one side of the stage on the unmatched pair than they do on the matched pair.  Therefore, the imaging you hear on the recording has the mid instruments in a different spot in the stereo image than they were in reality (perhaps in the extreme the guitarist now sounds like he was standing to the left of the bass player even though he was actually standing to the right).  On the matched pair, the imaging is accurate to where the instruments were located on stage.

A3: With respect to whether the above differences matter...it's probably a mostly subjective answer when the only issue is locating instruments in the stereo image.  Do you want the imaging to accurately reflect the on-stage location realistically, or does it not matter to you.  In a controlled studio environment, it seems like there might be situations where it's less subjective.  For example, if I set up a stereo pair with the mics in close proximity to a piano, such that I've got the lows mostly set up in one channel and the highs in the other, I wouldn't want the imaging on the mid notes to shift left or right in the image since you'd want the sound to be consistent with the piano scale, which of course have lows to the left, mids in the center, and highs to the right.  However, as has been pointed out, usually stereo mic'ing isn't used in the studio.  For spot mic'ing, which seems to be how the majority of studio recording is done, the above discussion is not applicable.  In that case, stereo imaging is created solely in the mixing process. 

A4: Whether the above is good or bad, I think it's mostly subjective.  In the studio piano example, we'd probably all agree that it's bad.  In other settings, then it's probably just subjective to your preference.  If it matters to you that the instruments are located in the same place in the image as they were live, then yeah you care about matching.  If it doesn't, then what does matching matter to you?  For me and I think the majority on this board, nothing.  To support this position, while it's not unusual to hear people comment about soundstage and stereo image, I've NEVER ONCE heard anyone evaluate a recording negatively because an instrument sounded shifted one way or the other in the image.  In fact, unless you were actually in the audience you have no idea where the instruments are supposed to be located in the stereo image relative to where there were on stage.

Conclusions: 
C1: It's subjective to your specific recording goals whether matching is important.  Unless the logic I've posed above is flawed, frequency matching would not make a material difference to my recording needs since the image shift issue isn't one I care about on my live recordings. 

C2: Repeating my contention from an earlier post, I don't particularly think just because an unmatched pair sounds DIFFERENT from a matched pair, that makes the unmatched pair better or worse (other than in situations analogous to the studio piano) since you never hear an objective commenter complain about image shifting on a recording, and unless they were in the audience they'd never know where the instruments were actually located on-stage (relative to the mic position) anyway.