The Unanswered Question: Do Amplifiers Really Sound Different?

Everybody argues about amplifiers. Whether or not “they all sound alike,” or if different components (tubes vs. transistors; transistors vs. FETs, etc.), operational classes, layouts, or moon-phases really make the differences some would have us believe. Everybody argues, then, but nobody does anything about it.

Nobody, that is, but this guy. Knocking about the Internet when I should’ve been working I stumbled upon his site, and on the “Tech” tab I found his Preamp Tube Comparison. It made for fascinating reading. And listening.

Let me back it up for a moment. This guy, who, appears to be a semi-retired Serious Professional Double-E, is in this instance talking about valves. That is, tubes. In tube amps. And not hi-fi tube amps, either, but in a guitar-amp circuit based on the venerable Fender “tweed” Champ. (For reference, that’s the sound of Clapton with Derek & The Dominos.) Yes, I fully appreciate that a guitar amp, being a musical instrument and thus music production component, as distinct from a music repro-duction one, is very different from a modern high-fidelity audio amplifier, be it tube- or transistor-based.

But the common link that binds them is this: every amplifier ever made, sooner or later, will be driven into its non-linear operating modes. That is, it will distort: clip, sag, compress, “sweeten,” “warm,” whatever you want to call it. Non-linear is non-linear: the output differs from the input by some quantifiable measure.

Now, a lot of audio folk believe fervently that amplifiers all sound different: that a Class A discrete-transistor 100 watts reproducing “Louie Louie” at car-radio levels is immediately distinguishable from a cheap receiver’s IC-power-module 100 watts doing the same thing. Fine; I ain’t arguing. Instead, I’ll reiterate what I’ve been saying for twenty…, err, thirty, err—what’s in a number?—years. Which is that amplifiers sometimes do in fact sound different, but mostly or perhaps entirely because different technologies, devices, layouts, and topologies behave differently in their non-linearities and as they enter and exit non-linearity. And every amplifier, from 1 watt to 1 kilowatt will, sooner or later and however briefly, visit non-linear-land.

For the vast majority of real-world audio amplifiers, it’s sooner rather than later, and far sooner than nearly everyone thinks. Music’s peak-to-average ratio is vast, and real-world loudspeakers connected via actual wires to real speaker terminals impose loads that send most amps, even very powerful ones, non-linear, for brief, sub-millisecond instants, on nearly every dynamic event, even at modest master-volume settings. These excursions are not measurable by our conventional, steady-state-biased tests, but our ears are integrators of remarkable power.

In the case of a guitar amplifier—especially a 7 watts-downhill-with-a-tailwind Champ—non-linearity visits, and indeed pretty much takes up permanent-resident status, from the moment you turn the thing on. In fact, that’s the whole point: the way the circuit non-linearizes, both statically and dynamically, is what gives the Champ its signature charm. Indeed, guitar-amp designs that didn’t go non-linear easily enough (late-‘70’s Ampegs), or that did so in less- -euphonic ways (Kustom, Standel, and indeed nearly all early solid-state models) quickly failed commercially.

Which brings me back to the above-referenced web page. Since I don’t expect all three of my readers necessarily to drill all the way down, here’s the exec-sum. This enterprising fellow Barton—for that’s his name, apparently—collected a baker’s dozen of 12AX7 dual-triode preamp tubes and ran a wide range of common, and not-so-common tests, including some under real-world conditions with the tubes in his Champ-clone test fixture driven well into non-linear operation.

The numbers and charts are hugely interesting in and of themselves, especially to tube geeks, but my real focus here is on the audio samples from his listening tests. Yes, these used pure tones, and yes, tones have no harmonic content and no dynamic shape, but what I found most ear-opening of all was the stitched-together composite, assembling a snippet from each tube, found near the bottom of the page. It’s the last clickable audio file on the page: I’m not linking it in hopes that you’ll read your way down.

Not only does the amount of audible harmonic distortion change with each tube, but so does the character—the harmonic content—clearly audible in changing timbres.

Does this prove anything in our world? I dunno; you all tell me.

Rob Sabin's picture
You can clearly hear what seems like a reduction in distortion and increasing clarity of the tone as it moves from tube to tube. Very interesting page!