Two microphones as matched pair, what properties are matched?

[kingkorg]

3. sensitivity can be corrected easily by gain knobs or in the DAW.

Sorry, but no. So many applications where you don't have the same reference source sound in order to match the levels. Like binaural nature outdoor recording. The end user is not always an engineer to know how to match levels, or have a reference to match it to. They buy matched mics in order to avoid to even think about it. Sometines there's just one stereo gain control. Depending on preamp and mic topology, different gain settings could lead to different THD and sound. You are again omitting noise, which is kind of weird having in mind the focus of your low noise mic project.

I'll point again to the huge ASMR binaural community where matching all of the aspects i mentioned including noise are the key. I don't think any of those artists have a slightest idea about how to match the signals.

 

[MicUlli]

Sorry, but no. So many applications where you don't have the same reference source sound in order to match the levels. Like binaural nature outdoor recording. The end user is not always an engineer to know how to match levels, or have a reference to match it to. They buy matched mics in order to avoid to even think about it. Sometines there's just one stereo gain control. Depending on preamp and mic topology, different gain settings could lead to different THD and sound. You are again omitting noise, which is kind of weird having in mind the focus of your low noise mic project.

I'll point again to the huge ASMR binaural community where matching all of the aspects i mentioned including noise are the key. I don't think any of those artists have a slightest idea about how to match the signals.

Hmm, i thought matching L/R levels of stereo signal is one of the trivial things in recording...
If different gain settings of a stereo preamp (we talk about <6 dB) produces audible sound change i would change the preamp...
How do you want to match noise performance? My approach is to make electronics noise contribution as small as possible...

But this discussion brings me back to the beginning:
How much deviation (in dB over frequency) should be allowed for a matched pair? Thanks for the useful comments so far, especially @Grim Reaper and @terry setter because they quoted real numbers / measurement procedures..

 

[kingkorg]

How do you want to match noise performance? My approach is to make electronics noise contribution as small as possible...

This makes sense, but if i go again for the ASMR card, those people are really pushing the limits, and use substantial compression on the top.

Røde went for basically source follower unity gain, which eliminates a lot of the variation within circuit. Than they used VLN FET which has tight prodution tolerance to begin with. This pretty much leaves the capsule as the only variation.

KM type of capsule is a nightmare to use for this purpose due to fiddly construction which depends highly on mechanical assembly process, regardless of machining tolerances.

In modern day and age mic application has new meaning. YT and similar content creators have made for some unforseen and often overlooked aspects to become crucial. It is also more likely for certain mics to be heard AND SEEN by billions of people for decades to come. No one could anticipate sales figures of Sm7b/Cloudlifter combo.

No one in the industry cares what mic Paul McCartney or Lady Gaga will use on next album. They are obsessed with the next Joe Rogan sitting somewhere in a basement playing the next big computer game.

 

[Rossi]

@MicUlli: Those plots look weird. If measurement conditions for both were identical, I'm pretty sure the mic with the red plot is out of spec and should be serviced (i.e. needs a new capsule). Even for non-matched KM 184 the difference is way too big, and there seems to be quite some bass loss.

Here's a plot of my own pair of KM 184. I bought them as the standard matched pair set.



Regarding the difference between the KM 184 and the KM 84, I have a feeling people are exaggerating. I often read that the KM 184 is a bright mic. As the plot shows, that's not the case. I worked as a reviewer for close to 20 years and during that time I measured quite a number of small diaphragm mics, and just about all of them had a (much) bigger treble boost.

I don't own a KM 84, but I have a vintage KM 86. Here's a comparison plot of the KM 184 vs KM 86 in cardioid mode (please excuse the slightly different scaling; those plots were not done a the same time):

 

[MicUlli]

@MicUlli: Those plots look weird. If measurement conditions for both were identical, I'm pretty sure the mic with the red plot is out of spec and should be serviced (i.e. needs a new capsule). Even for non-matched KM 184 the difference is way too big, and there seems to be quite some bass loss.

Here's a plot of my own pair of KM 184. I bought them as the standard matched pair set.

View attachment 109630

Looking upon these frequency responses you can be quite happy with your KM184

 

[Marc HENRY]

Neumann says their micrphones have such a narrow tolerence, that if you have bought a microphone and two years later you buy a second microphone of the same type, you have a 'matched pair

In my repair shop i have sometimes a dozen of u87 at hand. With my Stax headphone and home made preamp, you cannot find a pair. Even When sensitivity curve are 0.1db close, tone apears always different. Same story with my own u47 clones, even with Thiersch Blue. What is a pair of mic ? It depends of the résolution of your listening equipment but i do not Care a lot about this. My two ears do not give me the same sound neither

 

[Marc HENRY]

I had recently the hard task to make three trios with a lot of ten M50. I noticed something for what i have no explaination. How IS it possible that two KK50 of the same type can show a différence of 8db sensitivity...with exactly the same response curve profile ?
When there is such a sensitivity différence, even with same response profile, the listening sensation is not at all the same.
The worst pair in that lot was two M50 with exact same looking external and internal, consecutive serial number...

 

[kingkorg]

I noticed something for what i have no explaination.

Aging? Many times people describe mojo of certain mics not noticing some of the components are way out of specs, and even operating in non linear regions. It can be as simple as a resistor in capsule polarization voltage divider.

 

[user 37518]

It is possible and relatively simple to accurately match both phase and magnitude of two (or more microphones) with DSP. This would have to involve at least AD conversion inside the mics (or in a accompanying box) or AD and DA if someone wants to use their own preamps.

Another alternative is for manufacturers to provide a hybrid solution, in which they measure the response of a pair of microphones in an anechoic chamber or another suitable arrangement like a mic calibrator, and provide a transfer function file for users to include in their DAW by means of a plugin or similar. This would produce excellent matching, and the costs of all this might be even lower than the costs involved with the more stringent and low-tolerance manufacturing processes required to produce small variations between different capsules.

My best guess as to why this is not popular is because many users would probably frown upon a plugin+mic solution. Something similar to what some Studio Monitor manufacturers did years ago by selling the Monitors plus the software to correct for the room (and the monitor) deficiencies. As far as I know, these type of products are not as popular these days, especially among the high-end crowd.

 

[Khron]

Something similar to what some Studio Monitor manufacturers did years ago by selling the Monitors plus the software to correct for the room (and the monitor) deficiencies. As far as I know, these type of products are not as popular these days, especially among the high-end crowd.

Sonarworks seem to be doing pretty well though...

 

[abbey road d enfer]

It is possible and relatively simple to accurately match both phase and magnitude of two (or more microphones) with DSP. This would have to involve at least AD conversion inside the mics (or in a accompanying box) or AD and DA if someone wants to use their own preamps.

Another alternative is for manufacturers to provide a hybrid solution, in which they measure the response of a pair of microphones in an anechoic chamber or another suitable arrangement like a mic calibrator, and provide a transfer function file for users to include in their DAW by means of a plugin or similar. This would produce excellent matching, and the costs of all this might be even lower than the costs involved with the more stringent and low-tolerance manufacturing processes required to produce small variations between different capsules.

These are valid proposals as long as the directional characteristics are identical, which has yet to be proved.

My best guess as to why this is not popular is because many users would probably frown upon a plugin+mic solution.

+1. Actually many people frown upon the idea of using EQ to compensate the extra brightness of some mics. The typical solution they resort to is actually inadequate, since they apply a 1st-order filter to correct a 2nd-order bump. They think that doing it in the mic's head amp is benign compared to doing it externally...

Something similar to what some Studio Monitor manufacturers did years ago by selling the Monitors plus the software to correct for the room (and the monitor) deficiencies.

I'm a user of such a system (JBL4326+4312). It's great but does not dispense from proper acoustic treatment of the room.

 

[kingkorg]

It is possible and relatively simple to accurately match both phase and magnitude of two (or more microphones) with DSP. This would have to involve at least AD conversion inside the mics (or in a accompanying box) or AD and DA if someone wants to use their own preamps.

Another alternative is for manufacturers to provide a hybrid solution, in which they measure the response of a pair of microphones in an anechoic chamber or another suitable arrangement like a mic calibrator, and provide a transfer function file for users to include in their DAW by means of a plugin or similar. This would produce excellent matching, and the costs of all this might be even lower than the costs involved with the more stringent and low-tolerance manufacturing processes required to produce small variations between different capsules.

My best guess as to why this is not popular is because many users would probably frown upon a plugin+mic solution. Something similar to what some Studio Monitor manufacturers did years ago by selling the Monitors plus the software to correct for the room (and the monitor) deficiencies. As far as I know, these type of products are not as popular these days, especially among the high-end crowd.

I've been thinking about this, however there's latency involved with DSP which would severely limit and create issues in multi mic situations combining them with analog mics. Even 1ms is way too much.

 

[Gus]

+1. Actually many people frown upon the idea of using EQ to compensate the extra brightness of some mics. The typical solution they resort to is actually inadequate, since they apply a 1st-order filter to correct a 2nd-order bump. They think that doing it in the mic's head amp is benign compared to doing it externally...

This is a good point.
I think a possible reason for this is phantom and things like USB powered interfaces and not having good no or low latency EQ in the path.
When you add the more "correct" EQ(with more active parts inside the microphone) you can run out of phantom power.
You can use passive EQ parts often with an inductor but that costs more money.

OR this could be the copy copy copy "design" of some microphones

I don't understand current tube microphone designs they have an external supply and they are often still made the same way as in the past

I do like how the HPF in a U67 is implemented I have added this to solid state microphones. IIRC wavebourn shared a circuit with this in a solid state microphone.

 

[Tubetec]

Theres was the other thread lately relating to compensating for DSP processing in a mixdown situation . In a tracking or live situation even small delays can have a perceptual impact on the monitoring in the performers cans or in ear monitors , most especially for vocalists , where you end up with a direct path via the skull and then a delay that wanders in the monitors depending on the digital processing invloved ,its a variable you simply dont want because it can throw the performer off.
Its like additional slop in the gears . I remember back in the day when computer recording first came into being , many very enthusiastic people set up studios feeding monitor mixes done digitally to the talent ,blithely unaware it was having an impact , fully cock sure they were at the cutting edge of technology .

Likewise in some respects , in the rat race at the bleeding cutting edge of whats technically possible ,something was missed , what happens on the very short term peaks in program material , for all their good op amps distort in a most unmusical fashion , again these distortions can have a negative impact on the performance itself .
I sometimes wonder , do many designers actually have a clue about the job of sound engineering , or is it just join the dots and colour by numbers .

 

[user 37518]

I've been thinking about this, however there's latency involved with DSP which would severely limit and create issues in multi mic situations combining them with analog mics. Even 1ms is way too much.

It can be done during post-processing, that is, record as is and then just apply the transfer function correction to the recorded files. Something like Protools could delay-compensate for this with no major problems. Also, if the processing is done OTB, the 1 ms delay can be compensated by moving the mics closer 1 foot. I mean, this is something one would do anyway while choosing the sweet spot. Granted that this wouldn't be so easily done with things like a snare drum which is closely mic'ed, but then again, matched pairs are not commonly used for these purposes.

 

[user 37518]

These are valid proposals as long as the directional characteristics are identical, which has yet to be proved.

That is a really good point. I guess that it would still do more good than harm, particularly in those situations in which the mics are either very close together (XY and similar techniques) or if they are relatively close to the instrument so that the on-axis response becomes dominant. Either way, people will still don't like it; most are reluctant to the idea of having a digitally-assisted solution, many would probably believe that the mic looses its naturality. When, as you mentioned, most mics have internal filters and/or eq's to tailor the frequency response, so it is not like they are listening to 100% the pure sound coming from the capsule anyway, particularly if it involves a mic from the 60s with dried caps,, a half-dead tube, and a capsule with 6 decades worth of spit and breath humidity.

 

[abbey road d enfer]

It can be done during post-processing, that is, record as is and then just apply the transfer function correction to the recorded files. Something like Protools could delay-compensate for this with no major problems.

Indeed! However purity believers frown upon the idea of EQ. In the micbuilders forum, the most recurrent question is how to linearize sibilance without resorting to EQ?
I guess you know that "EQ produces phase-shift" and "Phase-shift is nasty". These are the most common "pearls of wisdom" coming from mic gurus.

 

[abbey road d enfer]

a capsule with 6 decades worth of spit and breath humidity.

When it's Sinatra's or Elvis" spit, it's called mojo.

 

[kingkorg]

It can be done during post-processing, that is, record as is and then just apply the transfer function correction to the recorded files. Something like Protools could delay-compensate for this with no major problems. Also, if the processing is done OTB, the 1 ms delay can be compensated by moving the mics closer 1 foot. I mean, this is something one would do anyway while choosing the sweet spot. Granted that this wouldn't be so easily done with things like a snare drum which is closely mic'ed, but then again, matched pairs are not commonly used for these purposes.

Lots of potential issues for live analog applications. Marketing around this in case of commercial product would be a nightmare. On the other hand i was thinking about gutting one of those cheap Sonic IR guitar pedals and putting them in a u47 body.

 

[user 37518]

Lots of potential issues for live analog applications. Marketing around this in case of commercial product would be a nightmare. On the other hand i was thinking about gutting one of those cheap Sonic IR guitar pedals and putting them in a u47 body.

I don't agree with your comment, the delays are not very significant since it is not a digital process that involves a lot of processing. If the manufacturer provides the transfer function file, it can be applied with convolution in the time domain, this is extremely fast and, if really wanted, a solution such as a dedicated DSP or FPGA could do it in close to 1 sample time (22.7 usec for 44.1 kHz or 5.2 usec for 192 kHz), this has already been done in applications such as active control of sound. Moreover, even if a sluggish digital system is used which produces a larger delay such as 1 ms, I still think it is also not a problem; consider that nowadays the norm for live sound are digital consoles, and they seem to be doing just fine.

 

[Cqwet Dbdfte]

I don’t believe Neumann or AKG ‘matched’ anything back in the day ... (pre-1980s ...?)

Users would select mics that were sufficiently close for certain applications.

It’s hype, really.

Yep. Spill over from "audiophile" nitwits, who like the term "matched" because they lack technical knowledge, and it can't hurt.
In a recording situation mikes are used for different signals, not the same ones, else more mikes would not be needed. Differing input signals would swamp out the mikes (same type) minute differences.
A more challenging aspect would be a Blumlein or a far field binaural application, but close miking not so much.
My $0.02.

 

[RuudNL]

I have worked in many professional studios, and they even did not have matched pairs.
Just "microphones"...

 

[Cqwet Dbdfte]

I would attribute "matched" pairs to tube amplifier aficiandos whos poorly designed multi-tube amplifiers lacked individual bias settings for each tube and no other provision for equal current sharing, which has spilled over into marketing mojo, and ignorance.
Even if tubes of the same type have slightly different gains (mu) , the resulting effect translated into ( log GAIN * 20) would be so small as to be inaudible.
If there are other effects, which MAY be possible, borrowing from the catalog of audiofiles, one would enter the minefield of psycho-acoustics, with very uncertain and non-uniform results.

 

[ccaudle]

In a recording situation mikes are used for different signals, not the same ones

Orchestra and large jazz band recording typically uses stereo microphone configurations, and it is very convenient to have microphones which are close in sensitivity and frequency response for those cases.