I haven't read any answers here but can say that harmonics are a result of free electrons and free electrons can stay in vacuum tube for a longer time and that's why they create more harmonics and regular mics.

That might have been true in the 60s when even love was free.

But in this new monetized century, just like beer, lunch and advise, electrons are no longer free. Today we pay for electrons. Even Texan power advocates had to abandon their free electrons campaign this past winter.

Also... in a tube, the behavior of the non-uniform density electron cloud moving forward to the anode happens far too fast to be relevant as a cause of audio distortion. But, could be an issue when recording microwaves at a radar studio.

All active devices are non-linear in practice. A significant difference is that when tubes are pushed into their non-linear range, the resulting distortion is predominantly 2nd harmonic - as nature intended. Transistors generate (nasty sounding) odd order artifacts - as cost-cutting accountants intended.

In a properly designed and built condenser mic circuit neither a tube nor transistor would likely be pushed into a non-linear operating point - other than @ chessparov2.0 doing his Ian Gillian impressions at full deafcon level. But we have pads and padded rooms to treat that.

And you were doing so well. No circuit is made up of just a tube. And the surrounding circuit can be many things from a simple Class A amplifier to a cathode-follower, a Class C amp, and so on. When we add another tube or section of a tube we can get into Class AB, and B. The same circuit topologies can be applied to transistors, though component values will obviously change. Each amplifier class operates the gain device differently and with a different type of distortion characteristic. For example, a cathode-follower applies negative feedback and has very low distortion. A Class C amp (used only in RF amplifiers) is richly loaded with distortion, but it doesn't matter because it's working into some very tightly tuned circuits and is used for its efficiency, not linearity, which is quite poor.

But it's the circuit, not the active device, that determines the distortion characteristic. Transistors have a linear operating range too, and can be biased to operate in Class A just as well, with very similar distortion characteristics. The claim is just false.
Oh really? Ever look into the type of distortion generated by the human hearing mechanism? There's nothing natural about distortion. Of any kind.
Or not. The lowest distortion circuits in the world today are built with transistors. And it has nothing to do with cost-cutting, it's using the right device to get the job done.

The tube/ss argument always goes this way. And in the early days of SS amps, there was merit to it. Some of those were pretty nasty. But that was 60 years ago, and now we have things like 24 bit DACs with immeasurably low distortion. Last I looked there wasn't a tube among them, until someone put one in the output buffer, thereby ruining the noise floor.

That's one area where tubes never win: noise. They're hot, there's thermal noise. SS runs cooler, less thermal shaking of electrons. It's just physics.
Pads that avoid operating a condenser mic into it's (more) nonlinear region would be those capacitive ones found between the capsule and amplifier, otherwise, off it goes into distortion, just like any over-driven amplifier.

To me tubes are capable of better sonics than solid state. But not as accurate. Better for me is measured in how much my ears open up.

Solid State is measurable to have better transient response.
Tube takes the energy that can't go through cleanly, and moves it around in a good way.

I think even dr Spock would’ve gotten the accountants and “as nature intended” jokes ...

I've run into far too many people who honestly believe those things to assume this time they were jokes.

"Fascinating".
"Spock, are you out of your Vulcan mind?"

Chris
P.S. Mando's mention of Ian, is most appropriate
on Easter. (Cue "Jesus Christ Superstar")

Great post (the science)

Great post (and the art)

Where I live, every day is Apr 1st. Probably best I don't get out much.

There is a huge difference between measurable and perceivable. The primary mantra of most accomplished professionals in our business is;
"IF IT SOUNDS GOOD---IT IS GOOD" !
Hugh

Several years ago an engineer was telling me how the Main engineer made a comment about the mic pre used on the vocal, "That Altec sure sounds good"
(1567a Tube mixer) the 2nd engineer said, "Thats Not the Altec but the JM-130"
A solid state pre..

Which goes to show it’s often what you think something should sound like that’s more important than what it actually sounds like

To further complicate matters, since sound is entirely constructed by the brain and is subject to changing human perception, nothing “actually” sounds like anything but what you think it sounds like. ... on any given day, month or year, and subject to weather, your mood and the joint you smoked.

Probably why that ART MP Pro smoked the Great River.
Also Solar Flares, can be a factor.
A Max Factor.

Which reminds me of a story...
Chris

This guy?



Or this guy???

Hey, it's April 5th. This can stop now.

Assuming you aren't kidding...

Of course there's a huge difference between measurable and perceivable. They aren't even the same thing, apples vs camels. Measurements can predict the perceived, if very careful analysis is done (big problem there), and perceived can be influenced by all kinds of non-audio factors. Perceived impressions are erratic, not predictable, not even repeatable with controls.

Take all that and ignore it, because in the end, "If it sounds good, it is good" always wins regardless of what the truth is. Variations apply everywhere outside of audio.

The “virtual logical” one

fascinating

Back when digital audio was new, I gave a studio tour to a local audiophile club. One guy ragged on how awful digital was, brittle, bright, harsh...you know the drill. I played him a few studio demo bits, one from 15ips Dolby A tape, another form an LP, one from a 16/44 master, and identified all of them before playing. He hated the 16/44 master, liked anything "analog". Then I lied. I put up another 16/44 master, and told him it was 15ips Dolby A. He loved it, said something like, "That's exactly what I'm talking about! Listen to how wonderful it is, nothing like digital!". Then I dropped my bomb and told him I spoofed him, and it was a 16/44 master. The room went up for grabs, and he was significantly P-Oed. But I made my point, nobody could tell analog from 16/44, and that was in the early 1980s when ADCs and DACs actually were pretty bad!

So, it's not only what sounds good, it's what people think they like, for whatever reason they like it. And has little to do with what is actually good or bad.

I figured as much but his title wasn't Dr.

It was Commander (or Lt Commander) which opened things up for my joke.

Actually you missed my point, many have used the word "Warm" to describe the JM-130 mic pre, He thought it WAS a tube pre...but was not..

Ah, good catch- I must’ve had the second one ingrained in my unconscious mind from childhood (yes, I’m that old).

No, I got it exactly.

People attribute sonic descriptions to things without knowing what they're listening to.

Oh noooes! You could be old enough to have been exposed to non-linear thermionic radiation - and may have even liked it.

Check with your semi-conducting odditory authorities - you might qualify for remedial confirmation-bias therapy.

I can't be around a 4038, due to my magnetic personality...
Chris

If you ask somebody like David Bock, they aren't going to say that they threw in a tube to add specific coloration, but rather the tube circuit was the superior design. The focus on tube vs solid state is myopic because you are comparing microphones as a whole. Even the U67 vs U87 thing, U67s have an added high roll off circuit for sibilance, and simply swapping out the tube (and transformer) doesn't a U87 make. I've also heard plenty of tube equipment that you might assume was solid state.

I'm not waist deep in the science, but I've been told that tubes are good at amplifying low level signals. I would guess that is the reason that even hifi top end mics Bock, Brauner, and others like the c800g use them.

...I thought it was your phantom powers.

What did the ghost say, when haunting the Old West saloon?...
I'm just here for the boos.
Chris

Of course I liked it - that and deep breathing of insecticides, the Beatles and captain kangaroo define my generation.
No need to remedy my confirmation bias, otherwise I wouldn’t enjoy half the expensive gear I own!

Here's how it works.

1. No gain stage is capable of passing audio with zero distortion.

2. A sine wave is the most benign signal there is, and the distortion is harder to detect and hear, but it's still there.

3. Real-world audio, such as a human voice, is going to distort more. The most subtle distortions occurs in the form of added spurious harmonics.

4. In the human-voice example, these harmonics are easiest to hear on open vowel sounds.

5. The proportions of these various added harmonics are, in a general way, different in tubes than in solid-state devices.

No opinions here, no subjectivity, that's how it works, and you now know what to listen for. Now go try it for yourself.

True in the absolute, but once distortion gets down to a certain level, it is no longer a factor. For example, a non-exotic opamp operated below clipping can produce distortion figures below 0.0005%. And in SS preamps, some form of opamp is typically used for the gain stage.
You have it exactly backwards. Distortion is most easily heard with sine waves because there are no other harmonics in the signal. Anything else, speech included, already contains massive amounts of harmonics that mask those produced by a distorting amplifier, and thus the distortion is much harder to differentiate.

Sorry, not correct. Firstly, a tube or a SS device is never used by itself, there's at least a bias circuit and some form of coupling. The bias on the grid or base places the input signal in the linear portion of the gain curve. Both tubes and transistors have that linear range. The linearity is variable depending on the specific type, but it's always there. The trick is to put the signal within it. There is no special difference in how a basic Class A amplifier works, tubes vs transistors. A mic preamp often includes an input step-up transformer (to match a mics impedance to that of the gain device) a tube preamp gain stage is usually several Class A amplifiers, because no one tube section has enough gain by itself. A discrete SS amp topology is pretty much the same. That topology, regardless of the active device type, produces primarily odd-order harmonics. The variations are now many and at what level. There will also be some even-order harmonics as well, but if input biasing is done correctly, they will be very low level.

And here's where we finally see the differences. There are very few discrete SS preamps. Typically the gain stage is an opamp or instrumentation amp comprised of several. Opamps operate with negative feedback, linearizing them and reducing distortion to a vanishingly low level. So even with an input transformer and an opamp, the chances of distortion of that circuit are are extremely low, and limited mostly to the profile of the transformer.

As to how this is audible, every complex signal masks its own distortion products unless they reach a level above which masking no longer works. The level varies with frequency and the original spectrum of the signal.

But, as you should be able to see, the distortion type and spectrum caused by an amplifier is not strictly and only related to the gain element, it's the entire circuit topology, and it's possible to make a discrete SS preamp that profiles just like a tube preamp.

The main differences between tube preamps and discreet SS preamps are in the type of transformer and how the output stage works. Tube grids have extremely high impedances. The input transformer is there to try to match a mic source impedance of 150 ohms (with a bridged load) to a tube grid in the meg-ohm range. Optimally, the match is closer to 100K. That's a high-ratio transformer, and those are a bit more difficult to make. They tend to be lower bandwidth, and need to be loaded with a complex impedance so they don't ring on a transient. A transistor is a current input device, so it's a very low impedance input. The input transformer for a discrete SS amp would not need to be a high-ratio device. The input transformer for an opamp based preamp would depend on the opamp. The classic no-frills 5534 works well with at 1:10 step up. A 990 opamp works best with a 1:2 stepup. There are noise advantages and disadvantages here too relating to input impedance, as well as differences in bandwidth and transient response.

Well, we differ on that. I have provided a bit of circuit theory to support my conclusion. As to the "go try it for yourself" idea, that one is heavily loaded with expectation bias. If we're going to go down the "no opinions, no subjectivity" road, which is the "objective" road, we need to stick to science and physics, not "go test it" biased subjectivity. And that process is hard for people to do on their own.