Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[klaus]

If, before auditioning a microphone, I examine some data to find that the microphone advertises

• an extended frequency response that doesn't deviate more than 0.5 dB

That information sets me up to believe, before I've even heard it, that the mic will perform well.

Come again? Neumann, a leader in high quality condenser mics, usually specifies ±2dB across the transmission range, and I am unaware of anyone doing any better. The bulk of that tolerance, by the way, can be attributed to capsule variations, while Neumann's mic amps, downstream from the capsule, are typically within ±0.5dB.[/list]

[soapfoot]

I take your point

[Timtape]

Brad, yours is a long post. I choose to deal with it one point at a time, the first being:

I'm glad you brought up confirmation bias, because it's salient here. If, before auditioning a microphone, I examine some data to find that the microphone advertises

• an extended frequency response that doesn't deviate more than 0.5 dB
• very low THD+N figures

That information sets me up to believe, before I've even heard it, that the mic will perform well...

OK.  Same solution. Blind listening where nobody  knows what mic they are listening to, and therefore what  '"specs" they are listening to, whether claimed or actual.

If we really want to  circumvent confirmation bias...we can.

[soapfoot]

Brad, yours is a long post. I choose to deal with it one point at a time, the first being:

OK.  Same solution. Blind listening where nobody  knows what mic they are listening to, and therefore what  '"specs" they are listening to, whether claimed or actual.

If we really want to  circumvent confirmation bias...we can.

Blind listening is great.

However, what's near-impossible is blind working. And I'm continually amazed at how often the thing that wins the "shootout" doesn't win in context; isn't selected when actual music is being made and recorded.

It took me some years to learn this for myself-- but now I seldom-ever do "shootouts" anymore, blind or sighted. I've just found that, for me, they're not all that predictive of what I'll find useful when doing actual work

[Jim Williams]

If we really want to  circumvent confirmation bias...we can.

Not really at this point of audio equipment development. Everything else in the system also imparts its own sonic signature as well. That skews the tests every time.

When you get to the point when even a power cable changes the sonic picture you are also evaluating the "rest of the story".

[klaus]

Jim Williams raises a relevant point. We could not begin to explore the original premise (the possibility for an accurate mic) unless we identify and factor each active and passive component's contribution to the sound arriving back at our ears.

What is left, then, is a rather clumsy approximation to "accurate": Which total chain sounds closest to what we liked in the original experience?

That brings us right back to my original skepticism: we cannot even begin to search for, let alone, identify, an accurate mic until we can extract and isolate the mic from the rest of the recording/reproduction chain.

[Timtape]

Blind listening is great.

However, what's near-impossible is blind working. And I'm continually amazed at how often the thing that wins the "shootout" doesn't win in context; isn't selected when actual music is being made and recorded.

Sure in a work situation, being blind to the gear we use can be very difficult.  But you mentioned the making and recording of music  changing which mic you choose. Can you give some practical examples?

[Timtape]

If we really want to circumvent confirmation bias... we can.

Not really at this point of audio equipment development. Everything else in the system also imparts its own sonic signature as well. That skews the tests every time.

When you get to the point when even a power cable changes the sonic picture you are also evaluating the "rest of the story".

Again, confirmation bias is  a cognitive issue, not an  equipment issue.  Again, if you're not familiar with the concept, please read up on it in the relevent sources. Here's one: https://en.wikipedia.org/wiki/Confirmation_bias

[soapfoot]

Sure in a work situation, being blind to the gear we use can be very difficult.  But you mentioned the making and recording of music  changing which mic you choose. Can you give some practical examples?

Sure.

The humble Shure SM57 is popular on guitar amp. But even in that application, it's not always something I find exceptionally gorgeous in "solo." It's certainly not anything anyone would ever describe as "accurate" (whatever that could possibly mean). It's got weird resonant peaks in its frequency response, a presence peak, a roll-off beginning around 200 cycles, its off-axis coloration is pronounced and not-too-helpful, always, etc.

However, in the context of a distorted electric guitar in a production I often prefer a good 57 to more expensive microphones which sound subjectively preferable in "solo", and which are objectively better in almost every commonly-measured way. It's a very popular microphone in this application, so either 2-3 generations of professionals are "easily led," or there's something about it that just works exceptionally well in that application, leading to the consensus.

Or for another example-- an RCA 44BX in good condition doesn't have a bandwidth extending much beyond 15k cycles (and is far-from-flat at least an octave below that). But on a trumpet, I prefer it to most other microphones, generally speaking-- even to expensive vintage German and Austrian condensers costing 5-10x more, and certainly more than modern offerings from the likes of DPA, etc which measure far better in every way.

My personal preferences are not overly meaningful. But to the extent that there's some degree of consensus among working professionals, that also constitutes data-- data that's every bit as meaningful as anything an Audio Precision Analyzer could measure.

[Timtape]

 ...in the context of a distorted electric guitar in a production I often prefer a good 57 to more expensive microphones which sound subjectively preferable in "solo",

As we know, compared to condensers, dynamic mics are almost immune to distorting under high SPLs. If the condenser mic overloads badly, it will probably sound horrible, adding unpleasant distortions that aren't coming from the amp cab.
So in this context, the SM57 which isnt overloading under the high SPL, is actually truer to the sound of the amp cab. It's more accurate to the source.

Re: the RCA 44BX on a trumpet. The trumpet  can have incredibly piercing highs especially when played full force and - unlike the player - we're standing in front of it! It's natural to sometimes want to tame some of those highs in a recording as it can sound more pleasing. Using a mic with  somewhat rolled off highs can be one way to achieve  that.
In this case we're talking about deliberate (but often perfectly valid) inaccuracy to the source.

[soapfoot]

1) I know the difference between an overloading microphone and a subjective personal preference.

2) I've never had a problem with frequencies above 15k on trumpet on any microphone. If anything is problematic, it's usually a strident (upper) midrange (well within the capabilities of a 44BX).

I disagree about "deliberate inaccuracy."

I select microphones, typically, that sound most to me like the source sounds in the room. I maintain that there's a decidedly imperfect correlation between that characteristic and a microphone's measurable parameters. Otherwise there would exist such a thing as the "perfect microphone," and I've yet to encounter it.

[Timtape]

I select microphones, typically, that sound most to me like the source sounds in the room. I maintain that there's a decidedly imperfect correlation between that characteristic and a microphone's measurable parameters. Otherwise there would exist such a thing as the "perfect microphone," and I've yet to encounter it.

How a source "sounds in a room" can vary enormously, not least on the characteristics of that room,  the placement of the source in the room, and our placement in the room, in relation to the source. There is no one sound - unless maybe the room is an anechoic chamber. So which sound is preferable? It's a choice of course.

Placement of the mic in relation to the source in the room is equally important, often far more important that the mic's spectral characteristics. Mic placement makes a huge difference, and again is a choice. 

Related to this, and recalling Klaus's candy wrapper example, I've been surprised that this subject hasn't been referred to in the thread:

https://en.wikipedia.org/wiki/Dummy_head_recording

http://www.neumann.com/?lang=en&id=current_microphones&cid=ku100_description

[Jim Williams]

Thr trumpet has a large amount of upper harmonics present, all the way to 50k hz.

Seems no one ever complains about the sound of a trumpet until we stick our less than accurate stuff in front of it.

Most microphones cannot deal with all those upper harmonics so they either distort them or in the case of a ribbon mic, low pass them.

[klaus]

https://en.wikipedia.org/wiki/Dummy_head_recording

http://www.neumann.com/?lang=en&id=current_microphones&cid=ku100_description

Your reference to dummy head recording is problematic:
You need to listen over headphones to re-create the binaural experience, and even with headphones, the results are, musically speaking, poor, despite the three-dimensional impression they give.

I never figured out why dummy head recordings lack musical warmth. Could be the plastic material of the simulated ear canals, could be the size and placement of the membranes...

An interesting direction for more a ergonomic type of recording that was never optimized to overcome its shortcomings. Probably too few takers to invest a lot of energy and money to try to perfect it.

[Timtape]

Thr trumpet has a large amount of upper harmonics present, all the way to 50k hz. Seems no one ever complains about the sound of a trumpet until we stick our less than accurate stuff in front of it.

Forget microphones and gear  for a moment, and consider a more commonplace example. Some peoples' sibilants are  loud compared to their consonants and can annoy many people listening. A friend of my late mother used to speak her sibilants VERY loudly. My Uncle found it very uncomfortable and  humourously nicknamed her "sss".

In the old days, sibilants, cymbals etc caused major problems for disc cutting and for other low fidelity media, causing often massive audible distortion.  The recording gear, including mics steadily got much better and can now handle pretty much anything.

But an annoying sound is still an annoying sound! So no surprise recordings are often produced with EQ, de essing and other artificial manipulation of otherwise natural sounds not just because the gear might struggle to capture or reproduce it - though that can  be true with cheaper gear - but also because it is a more pleasant sounding balance to most listeners. That's why strict fidelity in recordings, while a good general rule, is not always desirable for the listener, just as a very sibilant person is not always appreciated in real life. 

Most microphones cannot deal with all those upper harmonics so they either distort them or in the case of a ribbon mic, low pass them.

We cannot hear ultrasonics. That is what the term means.  The only reason we could know trumpets have harmonics to 50kHz is because a specialised mic capable of that band faithfully captured them! And a specialised analyzer plotted them in a form we can see but not hear.  So if we cant hear them, why capture them? In practice there are very good reasons not to capture them. But that's another discussion perhaps.