We’re allowed dissenting opinions here, right? Okay, then. In his May 2023 editorial on SoundStage! Access, Dennis Burger threw down a gauntlet, whether he realized it or not. The gist of Dennis’s editorial was how reasonably priced class-D amplification technology, including the Hypex Ncore NC2K and Purifi Eigentakt modules, is finding its way into more and more high-end components. This trend, which Dennis called “trickle-up tech,” has resulted in a growing number of overperforming products, including some of NAD’s recent Classic- and Masters-series amplifiers.
On many of the ideas Dennis floats, I am in total agreement. More money doesn’t necessarily buy better sound—check. Many high-end products are overbuilt for their intended purposes—check. There’s nothing wrong with lusting after those overbuilt components—check.
Dennis also fleshes out his reasoning as to why one might want to purchase an overbuilt component.
I can think of any number of valid answers to that question. Maybe you just like nice things. Maybe audio jewelry is what makes you happy. Maybe you’re smitten with the industrial design of one or another high-end audio manufacturer. Maybe you’re looking for components that zhush up the interior design of your listening room. Every single one of these answers is 100 percent valid, and don’t let anyone tell you otherwise.
That makes perfect sense to me. But elsewhere in his editorial, Dennis drops the bomb.
Think for a moment about what a power amplifier does and how it does it. The simplest explanation, of course, is that an amp increases the power of a signal to a sufficient level that it can drive a transducer. Along the way, the hope is that the signal is amplified with minimal noise, good frequency response, a high slew rate, and as little distortion as possible. And, of course, you want to make sure it’s a good match for your speakers.
Here’s the thing, though: once those measurable thresholds are met, and assuming the amp can deliver enough current to deal with the impedance swings of whatever speaker it’s attached to, you can’t tell the difference between two amps simply by listening to them. Nobody can.
Dennis’s statement is based on research carried out by David Clark back in the late 1990s, using ABX double-blind testing. These experiments have been repeated and written about in various publications and forums since then, and the results have been consistent: in an ABX test, the differences in amplification properties need to be pretty damn gross in order to determine which amplifier is X.
In case you’re not familiar with how an ABX test works, it goes like this. Two components are hooked up to the same speaker via a switch. The switch has three positions. The first is component A, and the second is B. The listener knows which is which—their identities need not be hidden.
But there’s a third position on the switch, labeled X. The identity of component X is definitely hidden. It’s either A or B, but the subjects of the test don’t know which. In a double-blind ABX test, neither do the operators of the test. The point is for the test to be literally double-blind so there’s no chance of accidentally giving away the identity of component X.
Further, the output levels of the two components are carefully matched, as even small volume differences can give the game away.
So music plays, and the subject can switch back and forth between A, B, and X, the goal being to identify whether X is component A or component B. The subjects can take as much or as little time as they want and can switch as quickly or as slowly as they want.
The end game here is that if listeners can correctly and repeatedly identify which amplifier is X, then there must be an audible difference between A and B. If the results are random, meaning that listeners fail to determine with statistical significance which component is X, then—so the thought process goes—there’s no audible difference between A and B.
Full disclosure: I’ve never participated in an ABX test. I have, though, participated in a blind listening test at Paradigm’s headquarters in Mississauga, Ontario, Canada. In this test, two different speakers were placed behind an acoustically transparent but visually opaque curtain. The levels were matched as closely as possible, but because speakers have differing frequency responses, this is only possible to a certain extent.
In this test, I was in control of the switcher, which had only two positions—there was no X in this test. While the music was playing, I swapped back and forth between the two speakers with about eight or nine seconds between each change.
That there were differences between these two speakers was easily discernable. But the goal wasn’t to determine if there were differences. Instead, the objective was to determine which speaker I preferred. While I could clearly hear that the two speakers were quite different, it took me quite a while to determine which one I preferred and why. The big reveal came at the end of the test, when the curtain was pulled back and I saw two wildly different-sized speakers, which bore no relationship to each other. The cognitive dissonance caught me quite flat-footed, as I had assumed, based on the sound, that the speakers were of equal size.
What does this have to do with amplifiers and ABX tests? Well, that A/B speaker test showed me that even when the differences are relatively huge, it’s still not easy to articulate and describe what you’re hearing.
Now, amplifiers . . . There’s no question that the differences between two well-designed amplifiers, driving loads within their design parameters, are exceedingly subtle. When I get a new amp in for review, I don’t even bother trying to make out what it’s contributing to my system. Instead, I let it settle in and bake for a while. Then I let it play, play, play. I listen for at least a month, and as time goes by, the thing will reveal itself to me. The tiny, subtle differences that I hear add up, and I can finally put fingers to keyboard to write about it.
I know that sounds precious. I can imagine Dennis reading this and rolling his eyes, convinced that I’m letting my sighted biases sway me, given that David Clark has already proven to the world that there really aren’t any differences between well-designed amplifiers.
It’s not that simple, though. Back to that ABX test. I know without a doubt that I’d fail such a test. It’s hard enough listening to two radically different speakers, without an X in there, trying to guess what they might be—what size they might be—let alone listening to three different options and trying to untangle which is which.
My feeling here is that an ABX test will reveal whether the subject is sufficiently organized, nimble, and perceptive to pass an ABX test. My feeling is that the inherent confusion imparted by an ABX test will serve to mask any actual differences.
Do I have any kind of solid statistical or practical proof of this masking hypothesis? I don’t. I’ll freely admit to that, so you can stop reading now.
But surely there has to be a better way. Like those blind A/B tests at Paradigm, couldn’t we just listen to two amplifiers and pick a preference? Couldn’t we just dispense with X? How would that work? I’m not sure—I’m not a practical research engineer. The optimum method, in my mind, would be to hook up two amps in my listening room and run them through a switcher, which would be locked with the key unavailable to me.
I could listen over the course of the review period and switch between my own Bryston and the newcomer. At the end, I’d have a pretty good idea whether any differences existed, or if I were unable to hear any differences, whether I’ve just been full of shit throughout my 23 years of writing for the SoundStage! Network.
At this point, I think I’m supposed to trot out several anecdotes, regaling you with instances of how radically different some amplifiers have sounded in my system, how my wife clearly heard the difference between two amplifiers from upstairs in the living room. I’m totally not going to do that, because good-quality amplifiers definitely do sound dramatically more similar than they do different. And because science beats anecdotes hands down, even if that science is, in my opinion, somewhat misguided. I think I’m clear in my reviews that, more often than not, the sonic attributes that define electronic components are extremely subtle.
It’s no secret that many—and maybe most—audiophiles are firmly entrenched in one of two camps. For the most part, you’re either a subjectivist or an objectivist. Either everything matters, and you’re not entitled to an opinion unless you’ve actually spent months listening to a component or tweak, or measurements are pretty much the end of the discussion.
I like to think I’m somewhere in the middle. It’s my hope that I’m rational enough to be able to dismiss ridiculously tweaky shit without the slightest twinge of guilt while also having the patience to listen to components that make some sort of rational sense.
That said, Dennis’s viewpoint has a ton of merit. It’s a given that he’s proofed himself against being bamboozled by dubious claims. His beat—affordable gear—is predicated on getting the most value for the money. Well-engineered products that actually do what they claim are the meat and potatoes of SoundStage! Access, and there’s no room for bullshit in the budget.
It seems to me that people on both sides of the subjectivist/objectivist fence are throwing the baby out with the bathwater. Measurements matter, and so does listening. That’s the basis on which SoundStage! is founded. It’s why we listen and why we measure.
It’s a real shame that Dennis doesn’t live closer to me. I’d love to have him over for a beer and a listen to this Hegel Music Systems H30A amplifier that’s sitting, troll-like, in my system right now. I’m pretty sure he’d love it, and we’d have lots to discuss. I don’t want David Clark to come, though. I have a feeling he’d be a bit of a buzzkill.
. . . Jason Thorpe