Tested: Where Does The Tone Come From In A Microphone?

[Steadyriot.]

Just watched Jim Lill's latest video on microphones and where their character comes from. Interesting stuff!

Only thing I'm a bit peeved about is that we don't know how flat his "hearing range" contraption is; thus how accurate his frequency graphs really are outside of the experiment. Doesn't matter in the comparisson between the mics but it still feels like a bit of an oversight. Would have loved a measurement mic in the lineup.

[øøøøøøø]

I admire this guy's interest in getting his hands dirty and making some tests. He really put the work in!

I'd caution against taking much conclusive away from this methodology, however. I hope he gets some feedback and does a "part 2" with a few additional things taken into account

[jorri]

I wasn't sure about his method again either. I think perhaps his mics move a lot, so at least he should have done a 'position test' frequency graph. What is the curve like if the mic is nudged 5mm? Some of his comparisons of the same mic model were visibly changing in between shots.
There are surely established ways of doing this since most mics come with a frequency graph in their manual.

How does harmonic distortion of a tube or transformer show on a frequency graph? I don't notice a huge frequency graph change when applying saturation plugins where audible difference is obvious: what you notice is the actual harmonics appearing in the graph. That should be the way to analyse saturation, and have seen it on the 'Snake Oil?' series a lot. I don't think this is drastically a change to an EQ spectrum when this happens, plus there would be a compressing effect and transient alteration, for this you need a more detailed curve that's not averaged. After all, the tube and transformer 'warmth' is not an EQ curve but a type of harmonic distortion.
How does it change when driven harder?
What purpose do these serve when running through tube and transformer preamps (is similar effect also achieved?)
What is the effect on transients and compressing qualities of a mic?
Should we judge a mic visually when the visual information is incomplete?

the variations of condenser capsules were the most interesting parts to me.

overly-flairy editing in that video so had to skip a lot.

[øøøøøøø]

The biggest oversight in *any* kind of mic shootout is testing only on-axis and only at close range.

In the real world, the biggest distinctions between superior and inferior mics tend to reveal themselves in things like pattern uniformity.

This is especially true with certain types of capsules… particularly anything CK12-related

[mediocreplayer]

In the real world, the biggest distinctions between superior and inferior mics tend to reveal themselves in things like pattern uniformity.

Can you explain further what this means? Is it whether the mic responds to off-axis sounds as it should (depending on its pattern)? Or is there more to it than that?

[øøøøøøø]

In the real world, the biggest distinctions between superior and inferior mics tend to reveal themselves in things like pattern uniformity.

Can you explain further what this means? Is it whether the mic responds to off-axis sounds as it should (depending on its pattern)? Or is there more to it than that?

It’d be easy to look at microphone documentation and assume it “does what it says on the tin.” Inconveniently, real life is much more complex.

You may know that graphs of frequency response behavior are averaged, or “smoothed,” such that very narrow-band anomalies don’t display (there’s no industry standard for exactly how this is done, by the way).

But graphs of polar response are even more…approximate.

One important consideration is that these polar patterns are 3-dimensional in the real world, and the 2D illustrations of various patterns don’t often convey a good intuitive sense of what’s actually going on. A cardioid isn’t so much “heart shaped” as it is “tomato-shaped,” with the rear null being where the stem would be (for one example).

But the bigger consideration (for purposes of your question) is that these patterns are generally NOT uniform at all frequencies.

A cardioid will morph into omni at low frequencies (the exact transition point depends on the design) and its pattern will become extremely narrow—and usually sprout a rear lobe—at very high frequencies (critical frequency here depends on diaphragm diameter). This is just one common scenario. In real-world mics things generally get a lot stranger.

If a microphone is damaged, poorly-manufactured or flawed in its design, more-idiosyncratic pattern anomalies can occur—but even a properly-working and well-conceived mic can have some very strange things going on off-axis. These are things that won’t really show up on typical documentation.

To truly comprehend this, it’s required to have at least a little understanding about how mics achieve their patterns.

A ribbon is “natively” figure-eight. A moving coil or single-membrane condenser is “natively” omni (if sealed at the rear). A moving coil can also be natively figure-8 if its rear is open (rare: think Beyerdynamic M380) and it’s designed properly.

Any other pattern is made by a series of engineering contrivances, and that’s where things start to get weird.

Mathematically, cardioid is achieved by a 1:1 sum of omni and fig 8, so the first cardioids (e.g. STC 4033, Western Electric 639b) actually combined a ribbon and a moving coil.

But this was cumbersome.

Later it was figured out that by leaking some sound to the rear of a moving coil or condenser capsule, you could cause *just enough* phase cancellation to create a cardioid… but the methods still in use for this are hilariously low-tech.

Thinking of this YouTuber’s favorite mic (the SM57), you’ll notice little side grilles. These are crucial to forming the mic’s pattern (if you were to gaff tape over them, the mic would become an omni)

There are also various mechanical and acoustic resonators inside an SM57 (including a big Helmholtz chamber behind the diaphragm). These resonators seek to counterbalance the diaphragm’s own natural resonance, and can also interact with the pattern-generating acoustical labyrinth.

This means the off-axis can… get weird. Peaky resonances can abound (think of taking a mid-band EQ, turning the “Q” all the way narrow, and cranking boost or cut…). Remember that these patterns are being achieved by phase-shift using building blocks that are resonant by definition and highly non-linear…

Similar things are achieved with condenser capsules, but through radically-different means. There, things get even stranger as the tolerances of backplate machining become critically-important (the AKG CK12 is so hard to get right that almost nobody will make a *true* clone).

Multi-pattern condensers usually rely on multiple diaphragms and careful engineering of backplate geometry (with a series of through-holes and blind holes to tune response and pattern).

There are always trade-offs when dealing with real-world physics.

This is already TL;DR but I’ve barely scratched the surface.

Summary: omni dynamics and single-pattern condensers are pretty reliable in their patterns (below a diameter-dependent frequency anyway). So are figure-8 ribbons. Anything else can get really weird, and the amount/acceptability of that weirdness is down to both design and also to manufacturing tolerances/precision.

[DrQuasar]

Thanks everyone for posting this stuff. I really enjoy reading it and I learn a lot. Very valuable.

The lesson that keeps popping up for me is that the secret to making a good record is to make more records. Really learn how to use the gear you have and maximize that. I really need to just sit down at my little home studio and record more and worry less about whether what I'm recording is "worth" recording. Do more "sketches" and stop worrying about whether everything is going to be a framed, final work.

How can I understand those granular details and how to use them if I don't encounter them while practicing? Such good information!

[øøøøøøø]

Practical experience is 99x more valuable than any received wisdom.

Maybe you can relate to this (I can):

After seeing someone else do it, you try a couple of MD421s on toms. You have the drummer hit the toms a few times and think “wow that sounds really good; my new favorite tom mics!”

Then you record something, and your attention is on how good the toms sound; you feel pleased

Then you go to mix, and you can’t get the cymbals to sound as good as last time

After lots of tail-chasing, *hopefully* you stop messing with the overheads and rooms and dig deeper. You finally figure out that the cymbal bleed into the rear of the MD421s is what’s really harming the cymbal sound (due to pattern non-uniformity that’s pretty unflattering higher up). So you gate or strip-silence, trying to minimize the fact that, while better overall, the cymbal sound now changes pretty drastically whenever the toms are hit

Maybe you live with it. Or maybe it’s too distracting and you mute the toms altogether (concluding, rightly, that the close mics are a sum-negative for the drum sound as a whole). The toms are represented well-enough in the overheads, you guess (but you don’t feel great about it).

Then you think, “hm, i guess the MD421 isn’t my new favorite tom mic after all” and you try something else next time

This is what learning to record looks like, and why shootouts that primarily compare on-axis behavior are… not the most instructive means of comparison

[marqueemoon]

I agree that it’s not very scientific.

Still fun though. The test of the differences in capsule construction was pretty interesting as was hearing mics with tailored frequency responses for bass instruments in quick succession.

[DeathJag]

Practical experience is 99x more valuable than any received wisdom.

I really love you all of your posts, thank you for sharing your knowledge and opinions. I can totally relate to your story about the Sennheiser mic picking up symbols with the rear lobe, only for me it was vocal mics picking up crazy cymbals during live recordings. Moving the vocals to a better angle and switching the vocal mics to Audix OM series seems to have solved it.

My question for you is, how do you deal with live recordings differently from when you're only recording drums? I got a little surprised at your declaration that close mics = negative gain, But I can see how that would be true for a nice room where the only sounds to be recorded are drums. (Kind of, sort of, not really!) How do you get nice overhead tom sound without having the cymbals totally drown them out? Perhaps the biggest factor is the fact that the cymbals are being slammed all the time versus an occasional lighter touch?

I am attempting to get this punk rock band off the ground and recording all of our practices, and we are super loud. I moved the drums about 20 feet away from the amps and put the vocal mics off to the side, pointing away from everything. The overheads do make the toms (and snare and kick) sound great, but the cymbals are about 1000% hotter so that part isn't useful. In fact I have dialed down those Coles ribbon mics to 2K, because 3K was a little too bright. (I remember you expressed interest in anything I discovered by tweaking the variable cloudlifters.) Bumping those down to 2K really mellowed out the harshness of the cymbals. With a quieter outfit I bet 3K would be better.

The close mics pick up the pure tone and bass of the toms, making them clear and full without phase issues. You'd barely hear them if I only used the overheads. They sound like cardboard on those!

[øøøøøøø]

My question for you is, how do you deal with live recordings differently from when you're only recording drums? I got a little surprised at your declaration that close mics = negative gain, But I can see how that would be true for a nice room where the only sounds to be recorded are drums. (Kind of, sort of, not really!) How do you get nice overhead tom sound without having the cymbals totally drown them out? Perhaps the biggest factor is the fact that the cymbals are being slammed all the time versus an occasional lighter touch?

I am attempting to get this punk rock band off the ground and recording all of our practices, and we are super loud. I moved the drums about 20 feet away from the amps and put the vocal mics off to the side, pointing away from everything. The overheads do make the toms (and snare and kick) sound great, but the cymbals are about 1000% hotter so that part isn't useful. In fact I have dialed down those Coles ribbon mics to 2K, because 3K was a little too bright. (I remember you expressed interest in anything I discovered by tweaking the variable cloudlifters.) Bumping those down to 2K really mellowed out the harshness of the cymbals. With a quieter outfit I bet 3K would be better.

The close mics pick up the pure tone and bass of the toms, making them clear and full without phase issues. You'd barely hear them if I only used the overheads. They sound like cardboard on those!

I fear I may have been misunderstood here—

I don’t mean to imply that close mics are a sum-negative in all or even most instances. They *can* be, situationally.

But the point I was trying to make is that there are a lot of variables, and it’s very easy to introduce additional variables without realizing it. It’s very possible to end up on the wrong side of the spiraling complexity that occurs when all of these unintended variables are interacting.

The scenario you describe here is actually a great example of this. You mention moving the drums as far from the amps as possible, but what this will tend to do with very loud sources is to make the bleed sound worse—it becomes more time-incoherent (it can create a slapback, in other words) and the room sound becomes a bigger part of the bleed (which is usually sub-optimal)

In a scenario where spill is inevitable, I usually have more success working *with* it, and trying to control it with soft gobos (rather than distance).

But absolutely do not take my word for this. Experiment and learn firsthand, because few things are true 100% of the time

If the cymbals are way louder than the drums in the overheads, it’s most likely that they’re being played that way. Which may be totally fine for the style, and you may want to lean on the close mics more. I close mic toms virtually every time I record drums, so please don’t think I was trying to say this is a bad thing to do

…only that, if leaning too heavily on received wisdom, it’s very easy to end up in a situation where some unforeseen variable ends up doing more harm than good, forcing your hand

[DeathJag]

I fear I may have been misunderstood here—

Ah! Copy that. The space I am stuck with, my basement, is long and narrow. I've moved the drums and amps all over the place and this seems to be the cleanest by quite a margin. The drums are hardly audible at all on the amp mics. There's also a concrete floor so this room could be easily considered far less than decent. They used to be closer together and there were so many problems mixing, which now makes perfect sense after reading what you just wrote!

Such good points about the off-axis response. One could even say that is more important than the on-axis response, because if it's picking up all the other shit you can't use it at all.

[marqueemoon]

Where I really notice off axis response and polar pattern stuff is as a singer.

Live, I like a really tight pattern for max feedback rejection, and I’m fine with eating the mic to get that. When recording I want to be a little more free to move around and work the mic while staying tonally consistent. Mics that have a huge proximity effect that comes on quickly or sound weird off axis are not great.

Picking an arbitrary fixed distance from a sound source and aiming a variety of mics straight at that source isn’t a comprehensive way to evaluate those mics because they’re designed for different purposes.

[øøøøøøø]

The drums are hardly audible at all on the amp mics.

This seems very believable! Guitar amps are loud and most people tend to stick not-very-sensitive moving coil mics right up on the speaker

I tend to be more concerned with guitar amp bleed getting into *other* mics— drum overheads by necessity are much further away from what they’re addressing and tend to pick up much more of what’s going on in the room

[jorri]

I fear I may have been misunderstood here—

Ah! Copy that. The space I am stuck with, my basement, is long and narrow. I've moved the drums and amps all over the place and this seems to be the cleanest by quite a margin. The drums are hardly audible at all on the amp mics. There's also a concrete floor so this room could be easily considered far less than decent. They used to be closer together and there were so many problems mixing, which now makes perfect sense after reading what you just wrote!

Such good points about the off-axis response. One could even say that is more important than the on-axis response, because if it's picking up all the other shit you can't use it at all.

Yep, distance isn't going to reduce bleed enough. In a ways it just makes things diffuse and blend together and make it worse for sure.

Talking of 'off-axis' response: Guitar cabinets and combos can also be treated as if they have a 'polar pattern', like a closed back being near to cardioid and open-back near to figure of 8, but they also will be very different for different frequencies, so the bass frequencies tend to be what's coming from the back. Figure-8 overheads are pretty useful in their null points.... if after doing this though, the room is just filled with the reflections of these instruments then nasty room sound might bleed rather than direct sound.

But there's no way around treating those mics which are further as if they are 'whole band' mics to an extent. In fact, doing this before i started with 'room mics' for the whole band, even though they primarily served as the drum's room mics. Then unlike any other situation, i would start checking phase relationships of guitar to drums and even considering delaying more distant mics (below audible slapback time) further by a small amount or flipping polarity, just listening for effect - as opposed to for example time-aligning everything. back to using ears though by the end of it if its unavoidable then use it rather than minimise it, there will be better placements for the tone of the bleed than others.