The Matched Stereo Pairs Debate
Tuesday, September 13th, 2011 | by matthew mcglynn
It began with my contention that “99% of listeners can’t hear whether a stereo pair is ‘matched’ or not.”
I will admit that I have, on occasion, paid extra for so-called “matched pairs” of microphones. But I’ve also seen better engineers than me use stereo pairs of vintage mics that are far from matched — sure, they’re both U87s or U67s or C12s, but if you’d put them side-by-side you’d absolutely hear a difference. 50 years of wear, on top of sometimes questionable manufacturing tolerances — not to mention wholesale changes in specifications during the early lives of some of these mics — means they’re probably less similar to each other than are any pair of contemporary mics, matched or not. But they’re turning up on records with bigger budgets than the entire cost of my studio.
My comment was intentionally provocative, and it generated some great feedback. One of my favorite replies is from my pal and Microphone Show cohost:
Mathew Trogner (@MathewTrogner)
I agree. I’ve always justified not buying matched pairs by the reasoning that our ears are not matched pairs.
See, I was going to start arguing that between the unpredictable room reflections and the differing distances from a source to each of the two mics in a stereo array, the two mics in that array are going to hear things differently anyway. In fact, that’s entirely the point.
But Mathew cut through my pent-up vector diagrams and logarithmic decibel calculations with an irrefutable biological argument. Ironically, I know for a fact that my left ear hears differently from my right — there’s a notch on one side that probably maps pretty closely to the frequency response of my first pair of hi-hats.
It’s a fascinating question, no? If a listener’s hearing isn’t matched, what’s the point of matching the microphones?
More feedback from the crowd:
AJ Horsburgh (@imakec02)
So true. Especially when converted into mp3 and compressed.
Tim Gosden (@timgosden)
and since preamps, cables and converters aren’t matched either… who cares!
Jason Miller (@JasonMiller0607)
Couldn’t agree more. In fact, I prefer non-matched as it’s only going to “widen” the stereo image.
Adam Bennett (@iateyourmic)
Makes bugger-all difference. Seeing as stereo is the difference between two sides, arguably unmatched images are wider.
But not everybody agreed:
Randy Coppinger (@RandyCoppinger)
Allow me to offer a dissenting opinion. I’m big on symmetry, so I prefer matched pairs.
This led to my second-favorite response of the entire exchange:
Well, Randy… it doesn’t get more symmetrical than one mic panned in the middle.
After much back-and-forth, the two found common ground: first, that some manufacturers maintain higher manufacturing tolerances than others, allowing non-matched mics to serve for at least some stereo applications, and second, that the need for matched mics depends on how prominently the stereo part is featured in the mix.
This idea makes good sense to me. If you’re recording an ensemble with a single pair of microphones, maybe it’s more critical that the microphones are matched than if you’re recording drum overheads for a pop song.
What do you think? Do you have buyer’s remorse about paying “matching fees”? Were the mics as matched as you expected (they did come with frequency-response graphs, and you did stack them and hold them up to the light, right)?
Or have you been recording drum overheads with two different mics for years, without anyone ever asking or even noticing?
Related posts:
Posted in Microphones, Technique | 9 Comments »
Karl Jackson
September 14th, 2011 at 2:51 am
Precise stereo imaging, like you might want when recording a string orchestra, requires closely matched microphones, since frequency deviations between microphones causes frequency components that should come from the same location in the stereo image to be smeared. The recording won’t sound “wider” as one commenter suggests, but rather will lack focus and clarity. Which is perfectly fine for many applications.
Randy Copinger
September 14th, 2011 at 6:05 pm
If you are trying to create a “true stereo” recording, such as Karl suggests with a string orchestra, then you want a realistic distribution of direct sounds between the speakers and an even spread of ambience. The greater the difference in frequency response and transient response between the two microphones, the less able you are to achieve these goals. But for a music production that is complete fantasy (not realistic), anything goes including two wildly different mikes.
Don
September 17th, 2011 at 5:31 am
Let’s for the moment agree that matching is quantitative and not qualitative. In other words a manufacturer would never sell a matched pair because they ‘sound’ the same but because they measure the same. The obvious implication is measurement equals sound. In my 30+ years experience I’d say there are times in which measurement did not reflect what I was hearing and while not always the case, true nonetheless. With that as a starting point, it seems to me there are circumstances in which a matched pair is a better ‘bet’ than a non-matched pair. Much of my work is chamber baroque and choral. So a great stereo image up front is crucial to me, but so is a great quasi symmetrical space to capture in. So my summary postulations are as follows:
#1 If you’re going to buy matched pairs, it is senseless to own them if you use them less than 50% time in a matched situation
#2 Matched pairs to imply a lot of symmetry in placement and environment. Not perfect symmetry but enough to prevent the null in the matching
#3 I don’t see a point in matching LDC’s. I think this is my classical bias showing here, but a matched pair of front SDC omni’s makes more sense to me than going through the expense of matching any other mic. Anyone that reads DPA’s Microphone University for five minutes may see my point in why I’d say that.
Great topic – much good comments!
-Don
David Gennaro
October 27th, 2011 at 4:56 am
I know my ears aren’t even a matched pair… they have different frequency responses. I would like to own a matched pair of small condensers, seems more scientific . As if a control group in an experiment, technical trust, less variables.
Andrew
August 10th, 2012 at 5:00 am
“…our ears are not matched pairs” is a ridiculous argument against matched microphone pairs. Unless you’re listening on headphones, both ears hear both channels and will pick up tonal differences between them.
Diffy
October 6th, 2012 at 4:12 am
i have bought 2 Rode NT1-A mic instead of a matched pair , what are the disadvantages i will have,, if its get more widening i guess i can use my pan knob two channels ..
matthew mcglynn
October 6th, 2012 at 9:30 am
@Diffy, the biggest risk of using two non-matched NT1-A mics is that the capsules might be different. I know several different capsule models were used in the NT1 and NT1-A — but I’m not sure if that variation carried over beyond the introduction of the A revision. The good news is that you would probably be able to hear the difference between them. Just set them up side-by-side and speak into them, and/or strum an acoustic guitar, and/or play the drums. If there are gross frequency-response differences, you’ll hear them in the playback.
I’d expect the circuits to be pretty nearly equal in their response characteristics; any audible difference would, I believe, be coming from the capsules.
Andrew Bak
August 22nd, 2013 at 11:30 am
The only identical pair of mica I have is a pair that I use for drum overheads; with so much going on from the kit, I’m not sure I’d notice the difference. On a single source like an acoustic guitar? Maybe. But I would only be concerned if the difference between the two mics was so large that it would need changes in placement or corrective eq.
jorri
June 1st, 2016 at 6:39 am
Surely unmatched, or even different brands will have less phase cancellation when summed, and maybe a wider image. I think overheads particularly mic different instruments of the kit, especially when you aren’t close miking everything you can fine tune each to particular drums. I even EQ the different sides differently. Sure its not natural, but nor is close-miking, compression or any drum sound these days.