A friend called me up a few weeks ago and asked if the DAC we both own had received an automatic firmware update he hadn't heard about; something had changed, it wasn't good, and he couldn't figure out what it was. His system's sound was suddenly pale and unfocused.
The next morning, he called me again. He'd remembered that, earlier the day before, he'd swapped out a meter or so of the standard Ethernet cable connecting his network switch to his DAC with an expensive, audiophile one.
This is not a review of an Ethernet cable or my friend's ears or the science of subtle cognitive bias. It's a review of Pass Laboratories' XP-22 solid-state preamplifier ($9500). The relevance of this cable storyto which I'll return in a momentis this: Lately, I've reviewed a lot of preamplifiers, and I've often found them to improve the sound of the system I've heard them in. Other reviewers have found the same thing. From a scientific perspective, this doesn't make sense. Adding a preamplifieryet another active component, hence more noise and distortionshould not improve the sound of an audio system, but degrade it. Any improvement seems almost as unlikely as a network cable dramatically altering a system's sound.
By the time my friend figured out what was going onor what he thinks was going onhis system was sounding "glorious," he said: better than ever. If his ears and brain are to be trusted, not only did that network cable dramatically change his system's sound, it also "broke in" over those first few hours of use.
The scientist in me finds this story incrediblein the sense of lacking credibility. My friend, though, is a solid guya scholar, a technical professional. Was expectation bias at work? Maybe. But consider that his expectations were at first so low that he hadn't even remembered he'd swapped out the cable. He obviously wasn't listening for a big change. But that's what he heard. I knowcrazy, right?
Poets, art historians, and musicians may find such stories easier to swallow than do former physicists (like me) and engineers. To those of us who like to think, rightly or wrongly, that we have some insight into what goes on inside an audio system at a mechanistic/electronic level, such experiences are mind-blowing. They can cause us to mistrust our own experiences, let alone the experiences of others.
Skepticism about our own perceptions is healthy and can save us money, but too much skepticism is bad for subjective reviewing: If it's true that we hear what we expect to hearand to some extent it is truethen it's equally true that we can, in marginal cases, fail to hear what we expect not to hear. Furthermore, to continue to doubt yourself when faced with convincing if statistically inconclusive evidence strikes me as not especially good for happiness and mental health. Explaining away your own perceptions starts to seem less rational than embracing them.
Pass Laboratories XP-22 That a preamplifier can improve the sound of a system relative to the same system without a preamplifier is mysterious, but there's nothing intrinsically mysterious about a preamp. Its functions are straightforward, as are its means of performing those functions: some relays or switches to choose a source, a variable resistor or some other technology to adjust the volume. Shift the signal from side to side with a balance control. All of these things can be achieved via straightforward electronic principles.
The Pass Laboratories XP-22 is a two-box preamplifier, one box containing its dual-mono power supplies, the other its signal-path electronics. These cases are connected via a robust umbilical described by Pass Labs as "aviation-grade." On the rear panel are two balanced (XLR) and three unbalanced (RCA) inputs, plus a tape loop (RCA). There's a home-theater bypass on input 5. The two outputsone balanced (XLR), the other unbalanced (RCA)can be used simultaneously. The left/right balance can be adjusted using the provided remote-control handset, a chunky metal slab that duplicates all front-panel controls, and then some.
My 55-year-old, presbyopic eyes can read the big blue characters on the XP-22's display from my listening chair, without glasses. Volume is displayed for the left and right channels separately. That display can be dimmed or turned off via the remote control or a button on the front panel. Then, when you change a settingsay, the volume or the inputthe display comes back on for a few seconds before turning off again. I like this feature, though it's not, I think, unique to this preamp or Pass Labs.
The XP-22 is handsome, with a robust, sturdy look, but those looking to make a colorful fashion statement should turn elsewhere: Pass Labs' preamplifier designer, Wayne Colburn, said that the XP-22 is availableold joke alert"in any color you like as long as it's instrument gray."
The XP-22 will replace the XP-20 in the Pass Labs lineup; the older model ($8600), remains in the line for now. Colburn told me that the XP-22's circuits are actually closer to those in Pass Labs' pricey, top-model line preamp, the Xs ($38,000), than to those in the XP-20. Like the luxury Xs (get it?), the XP-22, Colburn said, uses cascoded JFETs in a patented "Super-Symmetric" topology, with output MOSFETs (footnote 1) heavily biased into a constant class-A by an optically isolated controller"ie, an optocoupler. "We use an optically isolated transistor in the output bias circuit for a couple reasons," Colburn said, "the first being that the transistor is isolated from any input control signal, so there are no artifacts from such a connection, and also because the LED which controls that transistor has a consistent voltage characteristic which is also free of thermal drift."
The XP-22's circuits are more heavily biased into class-A than its ancestor the XP-20. "The higher bias in the output FETs lowers the distortion a bit, and some additional tweaking managed a few dB less noise," Colburn told me. A single-stage volume control, featuring an Avago-sourced encoder with good bearings and a nice feel, lowers the parts count. (The XP-20 has a two-stage controller.) Colburn said it's the same volume control used in the XP-30, Pass Labs' costlier three-box preamplifier ($16,500), which John Atkinson reviewed in the April 2013 issue.
All of these changes render the XP-22 "quieter and more dynamic than the XP-20," Colburn said. "We experience these improvements as offering a better sense of space and improved dynamics." Still, Colburn didn't try to oversell the XP-22's upgrades. "The sonic differences between the older and new designs are fairly subtle," he told me. "We are simply making the product incrementally better."
Tweakers be warned: The XP-22's owner's manual includes this note: "We can't guarantee that your audiophile grade fuse, won't blow at a different in-rush current threshold, than your stock commercial fuse. Use of other than approved fuses, may invalidate your product warranty and result in product damage."
Personally, I don't know much about perception bias or any of that stuff. But, I do know a bit about engineering.
My own experience at my day job, which has nothing to do with audio, is that a lot of these little things that seemingly make no sense from a scientific perspective really do make sense once you dig deeply enough. There are lots of engineering design "rules" that are accepted as being perfectly fine but really are only ok, at best. When you examine very closely, you can not only find flaws through analysis but through actual measurement. The problem seems to be that nobody wants to go to that length, unless a problem pops up. (Think of the O-rings on the Space Shuttle.)
Military snipers and guys who have to aim artillery weapons learned long ago that they have to take the rotation of the Earth into account, because the planet turns a small amount during the flight time of projectiles. People just tossing a football in the backyard don't need to worry about such things, and don't even notice, since there's other factors that dominate their throwing success. But, the Earth still turns, whether you notice it or not.
It's much easier to place the blame of a fault on somebody's perception. You have to wonder how many snipers and the like were chastised for being bad shots before the idea of Coriolis effect caught on.
But, to be really fair, people are moody, the air temperature changes as does the humidity and barometric pressure, and a zillion other things. So, personal observation may not be completely precise. That doesn't mean that it's dismissible. Just tossing off observations without taking everything into account isn't very scientific.
I'd also offer that almost every measurement I've ever seen for a piece of audio gear has been performed with the device under test in as much isolation from external influences as possible. That makes sense, except that these devices are never used that way. There are always interactions with other gear, but these get ignored. The only real test of a system is the listening test which is less controlled for the reasons stated earlier.
So, you have a funny dichotomy. The "objective" tests are performed not as the equipment is used. The "subjective" tests are performed as a system, but the results are less easily measured. Why is it that there are few or no attempts at rationalizing this?
"So, you have a funny dichotomy. The "objective" tests are performed not as the equipment is used. The "subjective" tests are performed as a system, but the results are less easily measured. Why is it that there are few or no attempts at rationalizing this?"
IMO, it's because the "pure subjectivists" don't believe in blind tests. So don't bother taking them and continues to make claims of massive differences! And then they find excuses why blind tests should not be trusted!
These days, it has gotten to the point where on some forums, talk of needing to do blind testing and the dreaded acronym "DBT" has become banned.
Personally, I have no use for the results of blind testing. Not that it's good or bad - that's not my complaint. It's that it doesn't help me get anywhere or make any progress. You've been very clear on your position in all this, as you routinely publish commentary on the topic. I'm not going to be critical of that nor say it's wrong. Or, right. Just that that's not a direction I wish to pursue or will travel.
I'm hardly a "pure subjectivist" by education, training, experience, career path, or in what I do in my own time. I am simply more interested in figuring out why something might sound better or worse. And, then doing something about it, at least when and where I can.
"Engineering" or "technical" testing would have been a better choice than "objective", because that was what I was really getting at. (However, I don't think that "objective" is the same as "objectivist" in this context) The limited measurements that are usually performed and the way they are performed are hardly representative of actual system performance. I mean that from a strictly technical point of view. It's just ironic that "subjective" reviews are the only ones that take system considerations into account.
We often hear the term "voicing" as part of the design process when manufacturing hifi products. For many designers, this is the final step (see the recent interview with Juergen Reiss of MBL on darko.audio).
This interests me, and is relevant here (I hope), for two reasons - it speaks to the fact that hifi products have a 'voice', and the only way to know what a given thing sounds like is to listen to it playing music in a system. Again, this coming from the people/engineers who actually make things.
If we were to apply the false (and forced) dichotomy of "Objective" v "Subjective" to this process, are we to believe that the people who make the gear don't know what they're doing? That the 'voice' designers speak of is nothing but a myth that can be easily dispelled by a few DBTs?
Of course, the suggestion is ridiculous. As ridiculous as believing that someone can determine how we will respond to a given component, speaker, or cable when playing music in a system based on measuring it in isolation.
...a designer striving to hear from their gear, a sound that's in their head. At least this is the way I've had it explained to me by a number of designers.
I am familiar with some of the work Sean Olive has done at Harman and I'd imagine he would represent the "If I make it accurate." approach.
I also know an amp designer who doesn't listen at all. It is not a part of his process. The fact that I've found his amplifiers to be among my favorites is neither here nor there (but I do find it interesting).
What fascinates me most is this difference in approach. I am certainly in no position, nor will I ever be, to pick one as being best. Or even most correct.
Which gets us to what we, as listeners, are after and again there's no single answer that holds more weight than another. In my opinion, of course.
We do voicing in instrument building and I take it to mean the same thing in speaker design (with appropriate exchange of mechanical versus electrical analogues, art notwithstanding). For instance, we may string up a newly finished violin or guitar and take a little here and there off the bass bar or struts, respectively, after listening to notes in every register. Even before final assembly we may measure braced free plate spectrums or compliance and take 0.01mm-0.1mm of top thickness off.
... "pleasant" versus "accurate" debate that David Hafler identified over thirty years ago? https://www.stereophile.com/content/manufacturers-comment
Do you want your audio equipment to pass signals through it with as little change as possible, or do you want it to be "voiced" by adding helpings of what JA1 referred to as second harmonic "sauce", or whatever other modifications might make the resultant sound quality somehow more appealing to you?
To answer your question, I want a hifi that allows me to listen to music as an activity unto itself for as long as I want.
True Michael. There is of course many preferences that any listener will need to decide for themselves around "voicing"; especially of speakers.
But typically this is not what audiophiles argue about. It's whether for the money the box resonances demonstrated on measurements were really meant to be there. Or if a subjective reviewer says a new DAC is "awesome" yet it's clear that the noise floor is poor - was that "voicing" and intended? How about probably the most contentious of issues - amazing $10,000 cables that if one were to put in a test bench would show no difference compared to any other decent cable?
Beyond technical discussions and measurements, blind testing IMO is a useful technique both to keep oneself humble and "honest" (appreciating our own limitations as listeners), and also as a corporate exercise to determine whether something makes a difference for the majority of listeners. We might not bother with this for speaker "voicing", but we sure can to determine if supposedly neutral $10,000 cables, low noise $25,000 amplifiers and less jittery DACs have audible value when companies claim their more expensive devices are technically "better".
Where we differ, and this is admittedly simply a question of approach and preference, is I do not feel the need to keep myself "humble" when picking out hifi gear and "honesty" doesn't even enter the picture. To clarify, I listen to music on the hifi for pleasure.
On the subject of cables and measurement, I would again point to Jurgen Reiss. I just listened to Jurgen's presentation in Munich on his approach to speaker design. He explained that he ships his speakers with three different wires for connecting the tweeter. Why? Because each wire, which are made of different materials, changes the overall sound of his speaker. The fun part is he has not been able to measure this difference.
"Value" is to my mind a very loaded term and best left up to the person doing the spending to determine.
After regaining consciousness in the aftermath of reading your post, a little light bulb turned on for me.
Back in the Jurassic days of audio and web publications, a guy who was as far from being a "pure subjectivist" as you can find authored an article as to why the typical measurements for cables used in audio systems were incomplete.
I'm certain this information has been lost or plain ignored over the decades, as I suggested earlier.
His approach was also one that I think is worth considering. Take your observations and investigate as best you can what might be behind what you observed.
His demonstrated approach - dunno what he was really thinking and I won't speak for him - appeared simple. He heard something. Then he set out to see why that might be. Then, if he could, do something about it.
Then what results do you "personally" find useful!? Is it really up to the "person" to decide? Does a person pick and choose what test was used to determine if the Earth is spherical or flat?
What better way than to literally have human subjects listen to playback of a full system and report whether something is audible while controlling for the inevitable psychological biases than blind testing?
No matter what technical or objective testing one does, you could always complain that the measurements used are "limited"!
The issue is that in audiophilia, there is an attempt to disregard the most powerful and relevant direct test there is! To continue doing this leads us round and again never knowing if subjective comments made have any relevance or truth at all.
How people process sound, especially when it comes to music which is intended to strike emotional portions of the thought process, is very much individual. There's been lots of studies showing that it is part genetic and partly it is learned. I don't even pretend to imagine to be highly knowledgable in this area, so I'll accept the word of the people who devote their careers to studying this.
If this was not true, then why do all the perceptual testing for codecs like mp3? The testing was performed to find a common ground of what was acceptable for the application. Emphasis on the word acceptable. If everybody was alike, or at least statistically so, why not just use the results from a single person's testing?
Except for one thing. I have all sorts of psychological biases when I sit down to listen to music. Yup, I said it. Sorry, opera and rap music do not do it for me. Nothing against fans of each genre, but it's not what I'm interested in listening to. I am equally sure that I am flawed in what aspects of a music reproduction spark my interest. My wife has different priorities in what parts of the music reproduction are desirable or unacceptable for her. Thankfully, we are pretty much able to find something that works for the two of us.
Listening to music is a hobby. So, finding the ultimate "truth" is not really that important, if it is even possible. Especially so when so much of the original performance, if it even was truly performed by humans in the same room at the same time, has been doctored along the way to produce something that met somebody's idea of "right" or commercially desirable. As a simple example, go listen to however many versions of Tommy by The Who that you can download or purchase on CD. Then go listen to the short excerpt from that album that Mikey Fremer posted on YouTube in a video about record wear. Which one is The Truth? Did Pete Townshend not know what he wanted to convey the first time around? The record companies are in the business of selling product, as the saying goes. Anything else is a means to that end.
Huh? So, some statistical test performed under conditions that are not like how you actually listen to music are ok, but my wanting better engineering testing and examination is a bad idea? Wait - I guess accepting incomplete testing from a psychological standpoint is consistent with accepting incomplete engineering testing. My bad!
My own view is that the blind test thing always comes out when somebody wants to show that somebody else's listening observation is at odds with their own personal dogma. Not always, but you can count on it just like you can count on Hitler eventually coming up in a political discussion. (Godwin's law)
Do audiophiles jump to erroneous conclusions? Yes! Proof of that is how they'll buy something new soon enough.
Do manufacturers throw out bogus explanations in their promotional materials? Yes! Unfortunately, it seems to work for them.
It all may be annoying, but very much a first world problem. (My own very personal complaint is that very earnest people are looking for information on products that might allow them to listen to music in a more enjoyable way for them. They might find a review like the one above, which offers one man's opinion of how the product worked for him. But, they also will find faux-scientific explanations as well as people telling them that what they might experience with their own ears is clearly wrong and not what they like. Just great...)
"I agree, the dichotomy should not be there...IMO, it's because the "pure subjectivists" don't believe in blind tests. So don't bother taking them and continues to make claims of massive differences!"
If I'm not mistaken, I believe John Atkinson himself (up until recently, Stereophile's editor, if you'll recall) recently weighted in on this subject, feeling that strictly controlled tests simply can't reflect the way we usually listen to and experience music. And I think very few people make claims of "massive" differences. Lab measurements provide most of what we need to know regarding objective measures of performance.
Although I wish JA would stop painting blind testing in a negative light with his experience of amplifiers from the 1970's :-). I have faith that in time he will.
... of removing extraneous factors, such as level mismatches, knowing what the equipment costs, looks like and its brand name, so that the listener can concentrate purely on trying to perceive whatever differences in sound quality may, or may not, exist?
... only one variable at a time should be considered. If aesthetics are your top priority, then use that factor to cull visually unappealing products from your shopping list. Then move on to an unsighted comparison of sound quality.
... using a controlled experiment to test a given hypothesis - such as does one piece of audio equipment sound any different, let alone better or worse, than another.
You still spend quite a bit of time posting comments on the site of the organization that made you a redundancy. Wouldn't your efforts be better directed toward creating new content for your own site?
Regarding time etc., while I appreciate your concern, there's no need to worry - I take a measured approach.
...not exist for another listener with different source material in a different environment and sound system.
I have taken part in more than 100 blind tests since my first in 1977, as organizer, proctor, or subject. My opinions on their efficacy or lack thereof are based on that experience. See www.stereophile.com/content/simple-everything-appears-simple.
... structure a "blind" - or single variable - test of comparative sound quality in such a way as to eliminate your objections to that testing method?
It is time consuming and costly to apply DBT properly. Look into research methodology and testing- it's not something you can apply in your basement or living room! Of course, you can do single blind testing in a "casual, informal" way for the kind of results that you are looking for... You might actually get informative results, but those results may not have any bearing whatsoever outside of that specific test situation.
As mentioned before, different listener, different source material, different listening environment, different sound system, etc. invalidates all previous test results.
Have any of you, those who keep asking Stereophile to do DBT's, ever participate in or perhaps know how to design a proper DBT or SBT?
Sounds like Ortofan remembers something about the "scientific method" from grade school, but does anyone else have any experience with research methodology or test design from college or university?
Ever measure or simulate what happens as you adjust the level of a volume control on most preamps? In many of the designs (half? more?), the overall bandwidth is lowest at the -6 dB setting of the volume control. That's because the output impedance of most volume controls is maximum at that spot. In concert with the input stage capacitance, the RC low pass characteristic gives lowest bandwidth at that point. Of course, there's ways to minimize that effect.
In addition, many input stages become less stable at various volume control settings and various source impedances, including the parasitic capacitance and inductance associating with the connecting cable. (The explanation for that one involves real math and is way too lengthy for discussing here. There's articles from professional journals that can be accessed on line. Or, just go measure or simulate the circuitry, if you want proof.)
Point is, as I mentioned above, there is often a good explanation for behaviors you might observe by listening carefully, but are not normally considered in design or testing. That is, unless you specifically look for it. (It's very easy to *say* that everything should be looked for...)
One other thing thing: Full, complete disclosure - I have never seen, never mind heard, any piece of Pass Labs equipment. I might recognize Nelson Pass at an audio meeting of some kind, but only because I've seen pictures of him. I wouldn't know Wayne Coburn at all.
Using a Wadia 781i which should sound fine without a preamplifier, I found that the sound is much better with a preamplifier (Audio Research Reference 5 SE) between the Wadia and my pair of Pass X 600 power amplifiers!
Is it just me? Or does an extremely well-designed, well-implemented machine like pretty much anything Nelson Pass has ever built actually have some intrinsic value, both in terms of quality of audio experience and just that je-ne-sais-quoi that comes from a fantastic piece of gear? And that Pass Labs makes something like this available at less than the price of a used Honda Civic whereas "certain other purveyors of audiophile exotica" add another zero to the price?
Is that why I once was able to recognize, within three seconds, the voice of an old friend I hadn't heard from in 10 years, over a poor quality long-distance connection, without knowing in advance he was calling?
As long as we have adequate time to become accustomed to, and to recognize and characterize the qualities of a particular source, we can actually make pretty good comparisons with other sources based on memory.
I would never trust my audio memory for more than a few minutes. If I really had an audio memory I would not have had to take lecture notes. Trying to write down in words what my musical listening experiences at a any given moment are with certain gear could be meaningless hours or days later. Compare a glass of wine from today to one from a month a go? Hardly.
We do know differences from the immediate changing of gear, cables, carts, etc. , which makes sense to me as I have done it time and time again. To me, it must be in the near same moment or all bets are off for my memory.
Yes, many people feel that aural memory is too short-lived to be useful. But I think that's based on a misunderstanding of how memory works. I may not be able to make "precise" comparisons of glasses of wine (to the extent that such subjective appraisals are "precise"), but I can remember whether what I drank a month ago was good, and whether what I'm having now is better or worse, and in many respects, I can also judge how it differs--assuming I'm someone enough into wine to do that in the first place. Same with audio. As an example, I have a particular CD ("Sweet Jones" by 9Lazy9), and on track 4 there are cymbal taps following the introduction. Since that CD is very familiar, and I have listened to it many, many times, I have a very clear impression of just how prominent and clear those taps were on my earlier speakers (Dynaudio Excite X12 and Special 40) relative to the ones I listen to now (Buchardt Audio S300 MKII). As long as there IS a clear difference--and in this case there is--and we've had time to home in on what it is, I believe memory is pretty trustable. That's what I'm saying when I talk about having enough to time truly become familiar with something. No one remembers every word of a lecture, just as no one remembers every note of the music they've just heard. That's not the issue, and not what is meant by aural memory (unless you're claiming that we must have perfect eidetic memory before our memory can even be considered competent). But if you become familiar with the voice of your lecturer over say a semester, then you would recognize, even with your eyes closed, if a different lecturer had taken his or her place. (The issue is the "sound" of his or her voice, not what was said.)
When I replaced my Kimber KS2020 coaxial digital cable with a new Audioquest Diamond cable, I did not hope or expect that there would be break-in. I hoped it would sound good straight out of the box. So I was dismayed to find there was unpleasant brightness and glare and I assumed I had bought an inferior cable. But since I knew about break-in effects I gave it some time, and within a few hours the treble became less forward and the bass solidified. It's now a beautiful sounding cable, providing better clarity than the old Kimber.
Euphonic distortion Better match to your amp than the output of your DAC Doesn't matter as long as you like it.
May be Mr.Austin would be interested in reviewing the McIntosh C52 pre-amp ($7,000)? ........ C52 has built-in DAC and headphone output ........ C52 also has a bypass-able 8 band parametric EQ .......
Subjectivist: Open to experience, yet even though experience is why reproduced audio exists, often vilified for experiencing it.
Objectivist: Typically caught in the circular argument that You Can't Hear That, which denies objectivity but may explain why they hear so little. See also Audio Protectionist.
Audio Science: That present state that authoritatively speaks for all future knowledge while forgetting so much of its past.
Scientist: Presenter of such knowledge. Practically unable to seek all of it, entirely suited to present fragments of it, and uninterested in the rest, at least until it happens, in which case is first to (re)claim it.
The Blind Test: Uniquely equipped to destroy most sensory relaxation and insight, and with it the difference between any two generic colas in three-second testing of folks who don't generally drink colas.
Artist: The Unfortunate who hasn't the credibility of a scientist, the integrity of the Objectivist, the voice of the Subjective listener, or the tester's sidewalk card table, but who has the responsibility to make it sound right anyway.
6061 Aluminum Manufacturing
Audiophile: Even though experience is why reproduced audio exists, the folks deeply misled by theirs while it consistently forms strong collective associations, conclusions, and eventually, hypothesis.
The Audio Press: Writers who, one hopes, keep these dynamics in mind and don't strip reason and art from experience.
I'm not saying you're wrong - you probably are not - but who would want to be around that kind of hobby? Not me...
Elevator Hoist, Tower Crane, Passenger Hoist Parts, Hoist Roller - Jingyang,https://www.jingyangroup.com/