Help

A friendly man at the audio meeting kindly offered that I borrow his spare Threshold SA/2.  He said something about my Krell FPB 300 being complicated, and therefore not as musically pure as the Threshold.  Oh, and his SA/2 blew away the Krell integrated at a previous meeting, he said.

Given the many things I've said to other people over the years, I suppose I deserve this.  It starts long ago and far away...

Things have really turned around from the days, way back in 1978, when I was working at an internationally famous (and perhaps not all for the best reasons) audio dealer and tube equipment modification factory called Audio Dimensions (me and several other rogue technicians from diverse elite backgrounds, led by the Tu-Be or not Tu-Be wizard himself, for a few years anyway, "Ike").

Threshold was one of the key brands we carried.  Ike was a serious tube freak who had been experimenting radically on tube gear since the 1950's, and he might have preferred to carry Audio Research, but there was already one Audio Research dealer in San Diego, and since we made what you might loosely consider competing products, often liberally borrowing ideas and circuits and especially parts values from Audio Research gear, this wasn't going to happen.  We actually carried mostly solid state amplifiers, except for in-house, and the direct tube driven Acoustats.

Anyway, Threshold had almost zero respect among the tube freak technical staff at the store, and although Ike and sales staff would be very careful what they said around many customers, the mask would come off when the front door closed.  We carried the so-called Class A models 400A and 4000A, and pretty much believed their fake Class A was worse than plain old Class AB.  A but fuzzy on the grip.  Mucking with the bias was not the sort of thing real linear designers do.  "Trash-hold" was our nickname for Threshold.  I arrived too late to see a model 800, except at a customers home one very memorable day (he had several, and drove 24" woofers along with Bozaks having 6x 12" woofers).  The 800 seemed to get a bit more reverence, for some reason.  It certainly looked cooler.

Whatever merit our collective opinions had, I've come to see them much more dimly through the lens of time, for various reasons.  We had a number of well known brands like GAS, and a few lesser knowns such as Mike Moffat's first company, and lots of good speakers like Acoustat and Beveridge, but what for a significant part of my short tenure was the favorite amp among the staff, and sometimes the owner too, was the rather new Nikko Alpha III, a MOSFET amp, one of the very first.  I think the owner was thrilled by the prospect of solid state devices operating like and allegedly sounding like tubes.  Well sometime about 10 years later I managed to buy a Nikko Alpha III, and I'm not sure it was the wonder we always thought it was.  I always grew tired of listening to it for some reason, I think it had high IM from inadequate driver circuitry, however no measurements I performed in the early 1990's confirmed this (I didn't have decent test equipment until around 2010), so I don't really know, perhaps there was some other problem in my system, I went through fixing many many problems, often years after they had started, finally I decided that no matter what I did the amplifier sounded bad.

But anyways, after quitting my job there, I didn't much follow the history of Threshold products, since they had never interested me in the first place.  One guy who became one of my best friends a couple years later got suckered into buying the Threshold 4000A that was at the bottom of the store rack for many years.  The deal involved him trading in a pair of Marantz Model 9's, and his KLH Nine's, for LS3/5A's in every room major room of the house, 5 or so, and the 4000A, because a 4000A could drive up to 5 pairs of LS3/5A's, no problem, Ike must have said.  Not much later the buyer found it necessary to volume control autoformers for each one, that's where somehow I met him through his installer who started asking me questions about Threshold amps, since I had once worked at the store, then introduced me to the guy with the 4000A, who became one of my best friends for several years, and I miss him now.

When I heard about his deal with Ike, I had the greatest sense of loss.  He had traded in three of the most rare, most well respected and legendary products for a pile of trash just for convenience.  My opinion then and still is that every room needs a high end system.  The Model 9, when it crossed the bench at the back of ADI, had produced the nicest looking square waves of any tube amp that had ever crossed the bench.  I don't recall hearing it though.  I'd wished I had.

I thought better of the LS3/5A's at first, they had been the little darling of the store, until I myself owned a pair, and couldn't stop modifying them into something completely different.

But I never thought much of any Threshold product, except obviously the 4000A could provide a pretty good amount of current, at least enough for a serious 4 ohm load--though the guy mostly kept the other rooms turned down, I think no more than 4 at once making about a 4 ohm load...

Nelson Pass was not the only designer I now greatly admire that I thought was a total charlatan then.  Certainly I'd include Bob Carver, I mean what had he done except create cheap high power amplifiers following a old and bad design (quasi complementary, which I've now come to believe isn't that bad). And he had these bogus non-linear preamp circuits that created greater peaks, less apparent noise.  All that is phoney baloney (and I still pretty much think so).  The best sounding nearly always is to be as linear as possible.  However, frequency response adjustment is a linear thing, and fine when done correctly.

Now, btw, I think I have very much respect for both Pass and Carver, about the same as John Curl--whom I've always respected.  They are all very ingenious circuit designers who have been very successful...no doubt a result of many positive factors.  However, one should avoid hero worship.  Just because people are good at one thing--ingeniously designing high performance audio circuits in this case--does not mean any of them are more than amateur audio scientists.  Engineering is one thing, design another, and science yet another.  Carver may be the most in line with audio science as it exists, though curiously he now makes tube amps (i.e., what he can sell, since transistor fads didn't follow the science very well).  Curl and Pass are both of the sort of "everything matters" school of audio design.  I think that's fine in some ways, I want my amplifiers to be Better than is actually necessary.  But both of them are essentially refuseniks when it comes to doing blind testing, particularly of the "prove you can hear it" variety like ABX.  And, basically, you can't do objective science without blind testing.  You can only prove things, if you to want to consider that proof, to yourself.  Or you can fool suckers, but that's not proof at all.

Carver, on the other hand, promoted a kind of blind testing marketing stunt, very much in line with actual audio science and engineering.  Proving he could make his amplifier sound identical to any other by adjusting the frequency response under load.  That's brilliant marketing with a solidly scientific slant.  However, of course it didn't play well in the increasing tweak funded audio industry, where spending more and more for purportedly necessary extreme technology increasingly became the norm.  Carver was far out of the mainstream audio nut mainstream.  But he had his fans, others who didn't like the combination of BS and high prices in other gear.  Last year I was quite surprised at how nice an old Carver amplifier sounded.

*****

Actually, I had no knowledge whatsoever of the Threshold Stasis models until the early 2000's when I was refreshing my Nakamichi cassette player collection, and learned about the Nakamich variety first.  Long before I even knew there was an SA/2 I knew that Pass had left Threshold and designed a series of MOSFET amplifier for Adcom.  I had thought that had happened in the mid 1980s.  A friend and I had been very interested in trying one of the Adcom amps, as they are rather high bias MOSFET amplifiers, and my friend knew that MOSFETS can be perfectly linear under the highest bias, unlike any other amplifying device.  However he later calculated that as the Adcom amplifier use many MOSFETS, the current in each one doesn't come clost to high enough for complete linearity.

At least a decade ago I had learned about Aleph amplifiers, and lusted for them, and Pass Labs amps, and lusted for them, and First Watt amps, and lusted pretty much only for the F5, which is in my mind the few with a High Fidelity design.

Since the time I bought my Krell FBP in 2008, I have lusted very mightily for a big really Class A Pass Labs XA 160.5 or above.  As much as the Krell is class A in momentary biasing, not so much in thermal stability.  The XA amplifier are, I would admit, even more Class A than my Krell.

But een in the perspective of an XA 160.5 or 200.5 monoblock pair which might replace it, the Krell FPB 300 is an economizing amplifier, which throttles power consumption back to a "mere" 300 watts when not playing anything or very hard.  When I'm actually playing music, the FPB 300 runs at the 2nd plateau which consumes about 600W for both channels, but it can ramp up to 1200W (or higher) for awhile (the fourth plateau).  A pair of Pass Labs XA 200.5 monoblocks constantly consumes 2000W continously when running.  The 160.5's would be 1500W.  That would be a bit much for my 14x16 living room, actually, even with my top of the line Carrier central air system.

Just in the past few months, I saw an SA/2 and was somewhat intrigued.  It's a nice looking amp.  However, I didn't think it was in the current power class I need for Acoustat 1+1's.

Now I've seen the mighty SA/12e and read about the SA/1 and SA/4.  And they sound...interesting.  The SA/12e looks like it might actually deliver something like full rated power in Class A.  OK, that's the Threshold I want (or maybe the SA/1e, which some say is better).  I don't want some friend telling me I deserve less.  Though I see now that the prices of the final top dog Threshold amplifiers is higher than the Krells...high enough really to be out of reach.  I'd sooner get an XA 160.5.

If High End audio is about anything, it's about long term lust, and I can tell you from experience you should never settle for second best, except for pennies on the dollar, and then be prepared to lose a lot more pennies.

In this regards I completely disagree with the likes of Floyd Curl (whose book I'm now reading, if nothing else, he is a very experienced audio engineering scientist who knows a lot of things valuable to me, but he's also too much a promoter of multichannel to be objective about stereo).  He wants to reduce me to a practical Consumer of equipment engineered to Reproduce Music, end of story.

I am not a Consumer!  I am a Player!  And I am a Magician!  And I'm a Kid Who Plays With Big Toys and Lusts for More!

Perhaps that's actually what the science of Audio Equipment Engineering ought to be about.  The Psychology of Lust and Stage Magic (fooling people, especially yourself).

The lesson of life is...you ultimately get those big toys you were thinking about.  Then the question becomes keeping them and using them better than nothing at all.

*****

The Threshold 400A is a 100 watt per channel Class AB amp, with a clever circuit that keeps the transistors on both sides always turned on through a kind of sliding bias.  This does not have the transfer function perfection of true Class A.  I was right in thinking it's not Class A, but it might not be a bad idea, either.  I don't see how it would generate "trash."  Of course we hardly ever did blind testing at Audio Dimensions, if at all only once and it wasn't double blind.  A single blind experiment is a magic show.

But I remember when, long after I was employed in audio, Dan D'Agostino came to an audio show in San Diego.  It was the mid 1980's  Now here was a true Class A amp, the KMA-50.  I didn't like the fan, though, and I told Dan I'd like to see an amp with no fan, 300 watts of Class A power when demanded, and it's ok if it runs a bit hot.  I told him not to slide the bias, that was fake, but hold it for awhile (see earlier post)...

Unlike Pass, D'Agostino was clearly going for the true Class A, and I respected that.  The KMA-50 used so much power it  needed a fan, and yet only produced 50 watts (well, maybe more into lower impedances, I didn't think  much about that then).  So I respected him more than Pass and Carver.  I knew little about the various Levinson Class A amps until much later.  I had also had known about the Electro Research A75 by reading Stereophile about it.

Ever since the Aleph era, it was almost seeming like Pass was on a mission to erase the idea that he ever made "Fake Class A" through sliding bias like the 400A and 4000A.

Anyway, I also started reading about "Papa" at the Rocky Mountain shows, and his workshops, and boy I thought I was missing a lot.  But I did read many of the white papers.  And I've seen Pass a lot at DIYAudio, which I read a lot.

So, like many people I think Nelson Pass is the very definition of cool.  But should I swap in my 19 year old amp for a 34 year old amp that Pass designed, just because Pass designed it more simple to make it sound more pure?  After I complained the 100W rated SA/2 might not have enough power or current capacity for my incredibly inefficient Acoustat speakers, the audiophile bragged that Nelson Pass always delivered high current.

Well, I didn't want to argue, but I well knew that NOT to be the case.  Nelson Pass has been anything but consistent over the years in his designs.  He loves to play around and try different things, not just high current amplifiers. Curiously, the XA 160 was a surprising weakling in Stereophile tests in 2003.  John Atkinson wrote:

The Pass Labs' measured performance left me scratching my head.  Its massive construction, high heat output, and high price suggest a limitless delivery of watts--yet assuming neither review sample was malfunctioning, the XA160 offers only moderate power.  Its current limiting of the clipping points is also something that historically has been regarded as a bad thing in a solid state power amplifier.  But like its highish otput impedance, this is something that is more typical of tube designs, which might well go some way to explaining why MF felt the XA160 to sound like a tube amplifier.

There is now a comment in the online version that this issue represented a production error of some kind, but a retest was never done on that exact model.

A few years later, Nelson Pass remarked:

Some manufacturers recommend the arc-welding capacity of their amplifiers -- we specialize in sounding as good as possible into reasonably ordinary speakers.  This is of course an arbitary decision on our part, but we are content with it.  There are more powerful amplifiers on the market, and we are happy to see them get their share of business as well.

The limited current issue was addressed and fixed starting with the .5 series of Pass Labs amplifiers.  That's precisely why my major lust for Pass Labs amplifiers  begins with the .5 series, such as the XA160.5, rather than their predecessors.   Not that a .8 isn't better, but the good high current starts with the .5 models.  And if Acoustat capable high current is the thing needed, you can just write off all the Aleph series amplifiers, all the First Watt amplifiers, and so on.  As I said, Nelson Pass likes to play around with different things, not just Curl or Krell high current.  However, I'd love to have a First Watt F5 powering my super tweeters, and I intend to build one someday.

Anyway, a few weeks later, this guy sent me a link to an add for Stasis 2.  He said that was a 200 watt amplifier which might meet my needs better.

I did a little searching on that model, and found someone in 2012 who complained of having bought one of those and spending a ton of money refurbing it, and they could have had a Pass Labs amplifier instead.

Well, that would have come to my mind also...

But anyway, what about the SA/2?

As it turns out, someone online has already done a related comparison, the flagship Krell (from 1998) vs the flagship Threshold of a few years earlier, the Threshold SA12/e.  The Krell is characterized as being "a little better" than the Threshold overall, with the difference being greater transparency, better extremes, and rock solid bass.

My FPB 300 is the smaller version of the FPB 600 preferred by that author, made at the same time, and has the same circuits just scaled up in the number of transistors and transformers.  The SA12/e is the ultimate Threshold, the biggest and most powerful Class A monoblock ever built by Threshold.

Which Threshold amplifier are their best?  Reading this thread, it appears the most respected Threshold amps ever made are: The SA/1 monoblocks with e mod in balanced mode, the SA/4e stereo amp in balanced mode, and the SA/12E in balanced mode.  BTW, over at DIYAudio.com, when Nelson Pass is asked about these amplifiers, he says it has been 25 years since he even listened to these amplifiers, and his sense of what sounds good in amplifiers has changed since then, but he remembered the SA/12E as being an improvement over the SA/1.  He also said that they sound "similar."

The Threshold amplifiers are all bipolar.  For some reason, since he left Threshold, Nelson Pass has sold mostly MOSFET amplifiers.  If amplifiers sound different, we might expect them to be a little different.  Or maybe it's just those "simple" circuits...

*****

Anyway, despite age, factory claims, or what any particular reviewers thought at any particular time (and of course, just like collectors, sellers, and fanboys, reviewers can have conflicts of interest) or even what their designer thinks of their earlier work now, it's particular amp could sound better than another, right?  Maybe there's some special magic in some particular model, you just gotta hear it, if you don't hear it for yourself, you are missing out.

Maybe not.  There's been a long line of audio engineers and reviewers that the amplification problem has been solved, to the point that "all high fidelity amplifiers sound the same."

Julian Hirsch, the longtime equipment reviewer for Stereo Review, declared that all good amplifiers sound the same in 1975.  But given his previous history of occasional hyperbolic claims, nobody should necessarily have taken that seriously.

But by the late 1970's, the ranks of people making this essential claim had grown quite large, and to include a number of pretty respectible audio engineers and reviewers.  And by the early 1980's there was an extensive body of double blind testing done by audio amateurs, engineers, and provacateurs, all essentially showing the inaudibility of amplifier differences, and more solidly the non-existence of unknown factors which could make amplifiers sound different.  All this happened, but was pretty much kept out of view of most consuming audiophiles by the growing High End audio magazines, which promoted the opposite ideas, that every amplifier has a peculiar sound, and if you don't get the right one, your life is ruined (and that right one will most certainly change in the next issue).

And in the late 1980's, this turned around into being more about digital, digititis and so on.  But I've got enough things to talk about in this essay.

In fact this amplifiers-sound-the-same stuff grew out of the original "objective" (though not called that until decades later) way of looking at audio design.  When electronic audio reproduction began, it barely worked.  But by the late 1940's, low distortion and high power amplifiers had been created through novel circuits and the use of the new-at-the-time negative feedback.  Getting the harmonic distortion below the approximate perceptual level of 1% was a huge achievement.  It's becomes easy to hear the effect of harmonic distortion pretty above 3%, and the intermodulation distortion that is caused by the same thing (nonlinearity) is quite objectionable sounding.  But by the 1940's, amplifiers could be made which were not only not objectionable sounding, they were starting to have no effect on sound at all.  By the beginning of the 1970's, it had become possible to push distortion even lower than that, ultimately to level like 0.03% which there wasn't any good evidence anyone could hear.  These numbers had a profound effect on the thinking of Julian Hirsch, who had seen a trajectory of great improvement, with what had originally seemed impossible now commonplace.

But meanwhile, the "subjectivist" reviewer like J Gordon Holt and Harry Pearson arrived on the scene, essentially claiming the numbers don't matter, listen for yourself, and their finding was that  many of the new super low distortion amplifier did not sound as well as the best of the previous generation of tube amplifiers, and also the new generation of tube amplifiers that was appearing from companies such as Audio Research.

A rift was created, grew over time, and now dominates audio.  Subjectivist reviewing is now the majoritarian way it is done.

There are many many stories behind all this.  Stories told differently depending on who is telling them.   A friend of mine was involved in the Stereophile amplifier test of 1979, in which blind testing did find at least two top shelf amplifiers whose difference could be heard under double blind conditions.  As it turns out, one was tube and the other solid state, the the resulting effect of differences in output impedance is now well understood by all audio objectivists.  At the time, many subjectivists, as I was at the time, glorified in the new findings, which disputed the findings of a Stereo Review test a year earlier.  But while some were temporarily shocked, it didn't last long.

One of the bigger things talked about is how ordinary audio engineers were wrong and had "missed" TIM, which was caught by some of the brighter lights, such as Matti Otalla, who made some very impressive low TIM amplifier for the Citation line of Karman Hardon.

Many many have peddled this story, but it has been overtaken by a similar but related story of how "feedback was used to cover up all ills" but "that doesn't work," TIM being part of the technical footnoes.

This has become a major mantra in subjectivist audio.  But as applied to well engineered audio equipment, it is entirely wrong.

For one thing, it was proven by audio scientists not long later after the discovery of TIM by a few engineers, that TIM is just the same old IM, but cause by a newer kind of sloppy engineering made more likely by transistors.  The same underlying phenomenon could just as well be measured by better IM tests.  The old IM tests were created for tube equipment which had lower bandwidth and wasn't so subject to front end overload.  Transistor amplifiers needed a tougher test for high frequency IM.  And so, the 19+20 kHz two tone IM was invented and is pretty much the standard now (though, ironically some say the old IM test should be used also).  And that was the end of the chapter on TIM in the objective audio community.  An actual test for TIM has never been standardized, because it isn't felt to be necessary.  This was not a failure, but a victory for objective audio science and design.

As far as feedback being the bane of all good design, that is either such a nebulous argument it cannot possibly be debunked, or just obviously wrong, as nearly all of the most widely loved amplifiers of the last 70 years have used feedback, and many have used it not entirely correctly either (but still gotten away with using it wrongly).

It is obvious feedback can be used wrongly, but audio engineering has guidelines and techniques, which if followed, result in the correct use of feedback.  It's not cutting edge science.  In the 1950's Peter Baxandal worked out the optimal amounts of feedback in certain simple cases using mathematics.  The conclusions are still a bit radical, and not always followed.  The upshot is that either very little feedback should be used, or very much.  It's the in-between area that can result in increasing amounts of distortion.

Many famously good amplifiers have gone after the large feedback side.  Large feedback amplifiers have included the Halcro and Wolcott, both highly praised by subjectivist reviewers.  And one of my favorites, the Electro Research Eagle 2, loved by J Gordon Holt.

I'm not saying feedback can be used wrong--of course it can!  However, it's not clear which "early transistor amplifiers using too much feedback and therefore having TIM" are the ones that are actually supposed to sound horrible.  Whom did they sound horrible to?  Was double blind testing done???

One early transistor amplifier using lots of feedback, the Sony 3200F, was the long standing favorite of my super audio nut brother in law.  Since tweako audio was both his business and his passion, he went to all the audio dealers, all the shows, everyone's house who would let him in to hear his tweeks, and sometimes his amplifier, a 3200F.  He told endless stories about how people preferred his 3200F to many of the most highly touted amplifiers.  It remained his favorite until 2014, when he switched to a particular Emotiva amplifier.  I believe Emotiva is a great brand of reasonably priced no bullshit high performance audio products--nobody really needs something better than an Emotiva--so I'd agree with him on that, but there's no reason to believe Emotiva or one particular Emotiva is unique.  Anyway, wrt the Sony 3200F,  I don't doubt it was a high fidelity amplifier either, at least as designed and when brand new...it might be a bit worn out in 2014.  It got a rave review in Audio in 1968, and had perfect objective performance up and down the line.  In 1968.  Are all these "bad sounding" high feedback amplifiers all made before 1968???

One amplifier, however, which seemed to slip off the tongue a lot among subjectivists in the 1970's (I know, I was there) was the Crown DC 300A.  Now this amplifier was widely held by subjectivists to be a bad sounding amp because of feedback.  But honestly I've never heard it myself, and I know that many many people praised it as a good sounding amplifier.  It was in the Crown lineup for a long time, and Crown was a very successful professional audio gear manufacturer.  I suspect it may not be a bad sounding amplifier at all, it was just a convenient target of something most consumer audiophiles have never heard.

Another amplifier supposedly in the "measures well, sounds bad" category is the Audionics PZ-3.  The PZ-3 was so-named because it had 0.03% distortion, a long held goal of what would be inarguably "good enough."  Notably according to Lynn Olsen, a very experienced audio designer and blogger, the PZ-3 was bad sounding, and his friend designed a low TIM replacement, the CC-3, which became a famous and top seller.

The PZ-3 however had an objective problem.  The undersized driver transistors would blow out.  The PZ-3 cost Audionics a lot of money in warranty repairs.  And the low TIM (which was a big buzzword at the time) CC-3 was a hot seller, which having the buzzword of the hour saved the company.

But it seems not everyone thinks the PZ-3 was actually a terrible sounding amplifier.  Perhaps it sounded better on some speakers than others because of impedances.  I was thinking of this when I saw a PZ-3 for sale on eBay, said to be in excellent working condition.  There was no CC-3 for sale.  Sadly I missed my chance to buy a PZ-3 and find out for myself.  And see if I could confirm it blind.

Though I have a good candidate for sounds-bad testing of both measurement and blind listening kinds: the Nikko Alpha III.

Once again, I don't see any evidence that blind tests were performed to prove the PZ-3 was even an inferior sounding amplifier.  And even if it did sound different, it could have been because of high IM, or it could have been because of it's dynamic output impedance.  It's not due to some factor which audio engineers have never understood or invented a measurement for, that's for sure.


By the mid 1980's I began hearing about double blind testing all over again.  I decided to try this myself.  I was sure I had the sound of capacitors nailed, having read Richard Marsh's seminal article on capacitor types, and even hearing a lot about it previously.  I constructed a box in which I could select from a worst-case case electrolytic capacitor (actually, 5 in series into high impedance), and a best case film capacitor.  The box was designed with an inner switch which could be set by my sister after doing a coin toss.

It became immediately apparent to me I could not tell the difference between these two kinds of capacitors.  Everything I listened to and for sounded the same as I turned the outer switch.

This started to change my thinking a little bit, but not my actions very much.  By that time my main audio system used two different tube amplifiers and a big bipolar amp for triamplification.  But my primary tube amplifier, a Citation II, was giving me endless trouble and I wanted to change.  I almost bought an Electro Research Eagle 2.

Now many years later I know the 1980's were rife with objectivist challenges.  For many many years David Clarke had an "amplifier challenge" in which he bet significant money that under properly controlled blind tests, people could not hear a difference between amplifiers...amplifier of their own choosing.   Mr Clarke traveled widely doing tests blind test shows for many audiences.  Nobody won the bet.

Part of the challenge involved matching levels very carefully, to 0.1dB or better.  If the amplifiers varied more than a few tenths of a degree in frequency response, recognized to be audible by audio objectivists, he would equalize one of the amplifiers.  To satisfy listeners, he would equalize the "lesser" amp to sound like the better one.

This kind of level matching is virtually never done by non-objectivist audiophiles when doing equipment comparisons.

Not too differently, Bob Carver challenged magazine reviewers, notably at Stereophile, that he could make his newest transistor design sound indistinguishable from their favorite tube amplifiers.  It appears that nobody ever proved Bob Carver wrong.

One of the key issues is not just matching the "frequency response" specification, as it is measured into an 8 ohm resistor, but the actual frequency response of the amplifier under the load of the intended loudspeaker.  These are very different phenomena.  Loudspeaker loads are very complex, and when an amplifier has a sizeable output impedance, as most tube amplifiers do, the amplifier response will be significantly if not greatly altered by it.

I myself had observed this in 1974, when I replaced my Dynaco SCA35 amplifier with a Marantz 2270.  At first I was greatly distressed by the loss in low frequency response.  After a month or so, however, I decided this was the new normal.  The SCA35, I decided, had bloated bass, and the 2270 had tight bass.  Their actual frequency response specifications would not have led you to think they could have an audible difference.  But they had very different damping factor, and that was very important with the low efficiency Advent speakers I was using.

Now in this case, I heard the difference, and decided new is better, perhaps not unlike a lot of people switching to transistorized amplifiers.  Meanwhile, a very very small number continued to think tube amplifier were better.  Most people were entirely unaware of this tube backlash, even into the 1990's, even now.  I caught the backlash side as soon as 1978, when I worked at a tube amp chop shop.

Even now there are long discussions between well known audio subjectivists and objectivists at DIYAudio, and perhaps a few other places, though never in any Audio Club in which I've been a member.  Most audio clubs, with the exception of the famous Boston Audio Society, and a similar Michigan group, are on the audio subjectivist side.  This may have something to do with sponsorship by major high end retailers, or it may be leanings of those who are interested in such societies.

What do I think about all this?  Though it's hard, hard to write down the full argument (I've already tried and failed many times in this blog), I think the audio objectivists are essentially correct.

I'm think I'm pretty much believe what NwAvGy says in this long and thoroughly documented article with many links.  And likewise with a large number of others: Ethan Winer--who has written Audio Expert, a book I just bought, the posters at Hydrogen Audio (a website which does not even permit subjective evaluations without statistically valid blind tests to confirm them), Peter Aczel (though his good vs evil rhetoric bothers me sometimes)--I've read a majority of the Audio Critic magazines and his posts over the years, and so on.  Here's an online issue of The Audio Critic which starts right out with letters, the first being a telling of how the late audio salesman and internet personality Steve Zipser was unable to determine which amplifier was a cheap yahama receiver and which was his personal $15,000 super amp in two separate sessions of invited blind testing.  That's very typical in the annals of blind tests, when subjectivists would even dare do them.  Then the magazine procedes to a small discussion of audiophile myths, which is greatly expanded in a later issue.  Then Peter Aczel's famous/infamous identification of the good guys ("white hats") and bad guys ("black hats") in audio.  That white hat/black hat thing is where Aczel crosses the line a bit, in my view.  Not just the identifications, I'm OK with that, but the ugly things he says about people like John Atkinson are OT.  He's a cranky old crank, though basically he's right, I think.  (I touched words with him a few years back myself.  He's the kind of person who, much like his antithesis James Bongiorno, is impossible for me to agree with.  He'll find fault with everything I say, even as I'm trying my best to agree.)

I know matters may not be quite as simple as a particular article or even book is able to present.  My favorite audio website, DIYAudio, has a wide array of both audio objectivists and subjectivists.  John Curl and Nelson Pass are among the most celebrated of the subjectivists who believe there is still a large "unknown" in what makes amplifiers sound different.  Scott Wurcer and moderator SY are among the more celebrated objectivists.  It's impossible for me to capture even the spirit of the arguments in this small post.  The major forum for objectivists and subjectivists to argue with each other is the nearly endless Blowtorch blog, nominally about the thinking behind John Curl's Blowtorch preamp of many years back.  This rambling discussion has gone on for more than a decade now and represent about ten thousand pages just in "Part 2."  I've read hundreds of pages myself.

There have also been smaller discussions specifically on what does or does not cause amplifiers to sound different.  But these small discussions may be dominated by one faction or another, and for whatever reason I find reading The Blowtorch Blog Part 2 linked above to be the most interesting.

It is of course false that "all" amplifiers sound the same.  But that's not really the question.  The question is whether we know what makes amplifiers sound different.  And the answer to that is a simple yes.  And High Fidelity amplifiers, which we know how to make, are indistinguishable, and lesser amplifier can most often be equalized (or the higher fidelity amplifier equalized) to sound identical.  No one has ever proven otherwise.

An amplifier, and the signal sent through it, are in principle fairly simple things.

High fidelity amplifiers, defined in a kind of circular way, are indistinguishable from each other now.  However, when I say "high fidelity" this does not necessarily include all amplifiers.  In fact, it may include rather few amplifiers is the use of compensatory equalization is not permitted.  Many many amplifiers of today, and especially including many tube and other high end amplifiers, could sound different because of different frequency response under load.  So, if we just discard everything that doesn't meet the high fidelity standard of roughly +/- 0.1dB from 20 Hz to 20 kHz, we aren't left with very many amplifiers actually.

The Krell FPB 300 however is pretty much in the high fidelity group.  And I suspect the SA/2 is also. This probably wouldn't be true of an Aleph amplifier, or some of the First Watt designs.

If legions of people haven't been able to prove they can hear differences not qualified by high fidelity standards, I don't think I'm going to be able to either, nor anybody I know.  In fact I've run 3 blind tests on my most serious audiophile friend over 3 decades.  It didn't seem to change his thinking after he lost every one.

In the end, it's not really about particular amplifiers.  It's about audio engineering.  It's whether or not scientists have a handle on what audio equipment needs to do, as audio equipment.  Basically the audio science says you need primarily the flat (or identical) frequency response (under load of course), and after that, fairly low Intermodulation Distortion (IM).  Harmonic distortion as such is not detected very well, you may need over a percent or so to hear it on real music...but IM is far more noticeable and objectionable, and should be below 0.1% if not 0.01%.  Beyond that, there is no X factor, unknown to engineering and science, that "explains" why equipment sounds different.  In effect, you don't actually need to "listen for yourself."  That's what the snake oil salesman always says in the hopes that some will have their expectation bias follow his suggestion.

Peter Aczel has been one of the most forceful critics of audiphile subjectivism, and he discusses the X factor idea.  That's really what it comes down to, and in 37 years since 1980, no X factor has ever been discovered.  There have been books written, such as by Ethan Winer (I just got that one) debunking audiophile subjectivist myths.  There are major websites, such as HydrogenAudio, dedicated to the objectivist view, to the extent that purely subjective claims cannot be made without methodologically valid blind testing.

*****

So, knowing all this, what does one do?  Does one buy the cheapest possible amplifier?  Well if that's all you can afford.  But many preaching the objectivist audio creed buy very nice amplifiers.  Boulders even.  Because it's nice to have something much Better than actually needed.  If you can afford it you can flaunt it.

What they don't buy, are non-high-fidelity amplifiers, such as most SET's, which have predictably audible distortion and impedance interactions.

After my last tweako amplifier...actually the Nikko Alpha III I finally got in 1989, I moved on to what I thought plain vanilla objectivist amplifiers, the Parasound HCA's, only later discovering Peter Aczel was no particular fan.  But I don't think he was opposed to them completely, despite complaining about construction quality not being the spec he was used to (he liked megabuck Boulder amps).  They measure well enough to be fully High Fidelity, as well as having the extra High Current capability that has been Curl's trademark.  No slewing amps here!

But I didn't really think the HCA-1000A was good for the 200W recommended Acoustat 1+1's, and indeed when I first got them, and running full range, and equalizing the bass response heavily, and playing loudly, I got the HCA-1000A to shut down.

So, I began to think about amplifiers, and this is always the way to start...speaker first, and then what kind of amplifier does the speaker need.

If I had even known about the Threshold SA/1 and SA/12e and SA/4 they might have been on my list.

I knew about Pass Labs (too expensive), Aragon 8008 BB (about the most massive power supply ever, 2kW of torodial transformers), a few others, and Krell, all noted for high current.  Parasound just sounded too pedestrian, the JC series had not come out yet, the 2200's had a mixed history and looked a bit junky.

I really figured I'd get the Aragon but I kept losing bids on them.  Finally there was a reasonably priced FPB 300.  Oops!

The FPB 300 was like a dream come true.  If I hadn't actually told D'Agostino that I'd like an amp like that in 1985, it was the kind of thing I had long been thinking of, since having finally turned away from tube amps which had always been too high maintenance, and which I didn't really think necessary.

It had everything I wanted, not just the high current and power made possible by 3kW transformer, huge number of transistors and all.  It has the full regulated power to the output stage, another longstanding dream, a big one.  It has the real full balanced operation.  It has high power bandwidth: 300kHz.  I used to think it had a no feedback output stage, which I figured would be most perfect for electrostats.  In reality, there is no global feedback, but the drivers and outputs are locked together with feedback.  That's actually still fine and good, and reacts to speaker push-back as well as anything.  I'd bet it has relatively low settling time with regards to pushback from the speaker, which is the biggest factor in electrostats (they reflect most power back to the amplifier).

The Strickland designed amplifiers do something similar because the output side of the outputs has no feedback, feedback is taken from the input to the outputs.

It has always sounded great to me.  When I first needed to get it repaired in 2010, I needed a repacement and sure enough the Aragon 8008 BB was for sale again and I snapped it up.  It has been a lust worthy and capable amp, but not as good (in my long sighted usage experiences) as the Krell.

Now there's a long line of people who say the Krell sounds "slow" somehow, or "dark" or whatever.  There's no reason to believe this whatsoever.  None of these people have probably done blind testing, blind testing is extremely rare and especially among subjectivist audiophiles.  They're driven by their expectations, notions like "complexity" or "industry", perhaps the dark colored aluminum itself.  And perhaps their love of things that glow in a particular way.

The Krell has wide bandwidth--to 300kHz or beyond--more than most amplifier Nelson Pass has ever designed.  It has a no loop feedback design, limiting the speaker pushback to very fast settling output stage.  It has regulated power, class A operation, low distortion (spec is 0.03% !) and noise.  This is a high fidelity amplifier, and got rave reviews by leading subjectivist reviewers when it came out, as the #1 worlds best at the time.

If in double blind comparison the Krell actually was found to sound "darker" than some other amplifier, it would be most likely because the other amplifier was editorializing, adding brightness, probably IM, that would grow tiring over time.

Otherwise, there would be some aspect of it's behavior, probably easily measurable, which would expose this characteristic.  I know of nothing of the kind.  It is somewhat complicated, yes, but the resulting IM distortion is conservatively below 0.03% even by the most stringent tests.  And it is the IM, not the simplicity or complexity, that is the important factor.

Just because lots and lots and lots of people believe something, doesn't make it true.  Lots of poor evidence doesn't make it much better.  What's needed is at least a small amount of very good evidence, such as the kind that comes from careful measurements or double blind tests.

Anything else, is pretty much a waste of time.  It creates false "findings" which are really superstitions, that can be long lasting and very harmful.

When it comes to power amplifiers, high fidelity power amplifiers that is, the problem of amplification is so well understood, and so well handled, that pretty much you do not have to listen if you have trustworthy objective information.  You do not have to "listen for yourself" and in fact, unless done with all the controls, is an ivitation for trickery.

Peter Aczel said you do not have to listen to Power Amplifiers, in 1990.

Now speakers are a different kettle of fish.  There is no such analysis or perfection of loudspeakers.

Also, there's such a thing as Infidelity: Non High Fidelity amplifiers.  Non-High Fidelity amplifiers have high levels of distortion, limited or unflat frequency response, or high output impedance.  These are cases of not being faithful to the source.

But with some speakers, in some room, some people might like the effects.  Of particular use, and not necessarily bad, is high output impedance.  It's not actually fully settled science how the speaker and the amplifier ideally should interact.  It's quite possible a higher impedance is better under some circumstances.

And so, a pioneer in transistor amplifiers, Bob Carver has gone to making tube amplifiers.  And there is some argument for it.

But if it's a different frequency response that's created, and that's likely the only good that could come out of a high output impedance at the speaker interface, that effect can equally well and more efficiently created with digital equalization.  No noise nor distortion need be added if both inputs and outputs to the EQ process are digital.

I'm there.  Famous reviewer at The Absolute Sound and longtime friend of Harry Pearson, Robert E Green, has long been there.

Nowadays, given digital sources, it's trivial to stick in any kind of digital equalization you want.  My preferred vehicle is the very interactive Behringer DEQ 2496.

Tweako audiophiles would prefer to do their adjustments through largely ineffective and very costly tweaks, more largely for tweaking the imagination.


*****

But the Krell is still perhaps not a very good objective choice, largely because it is more than I need.  There's no real reason I "need" Class A operation.  It's a waste of energy, it leads to a need for more maintenance than other amplifiers.  I personally like the ideas in it though.  Perhaps I'll discover a reason why it is, actually, needed.

That's my penultimate argument with audio objectivists.  We know a lot about the requirements of audio equipment (especially amplifiers and digital) but there, indeed, may be more.  I think one of the bigger open ended things is frequency response.  My personal belief is that we need 40kHz response, as well as deep infrasonic response.  So I continue to work on these things.  I can't prove the need for them, but I think there's some possibility they may sometimes have an effect.

This argument does get off the track in my own peculiar way, with my own theory of information, and so on.  None of the things I want conflict with ordinary High Fidelity requirements however.  I just go beyond it in a slightly different way than wanting more of the good stuff (linearity) than necessary.

My position breaks from the objective ones precisely at the point that audio is engineering in pursuit of reliably conveying musical information from the musical producer to the consumer.

My idea of the goal is not reliable conveyance, but peak experience.  Audio should be capable of delivering, on occasion, peak experiences.  Audio science shows what reliably transmits the information, but where is the science of magic magic?  I think, partly, it's in the improbabilities.  We can predict to a great deal of certainty what amplifiers are going to sound different in a fair DBT.  But it is never perfect certainty.  There could always be some peak difference that isn't yet accounted for.  We shouldn't fool ourselves into believing these things are going to be ordinarily or easily hard.  They are in fact going to be nearly impossible to hear on a reliable basis.

Peak experiences are not something you can blind test for.  Peak experiences cannot be predicted.  They result from unpredictability more than predictability.

So it is, that people might like the always different (not "high fidelity") vinyl reproduction.  Because of audible noise, listening to a record twice never presents exactly the same information.  Perhaps this is a virtue.

Anyway, this possibility that you are on to something that others have missed...this is part of the thrill.  It can't, and probably you won't even allow it to be confirmed or disconfirmed, because that would likely destroy the suspended disbelief.

The findings of objective audio science should be regarded more as liberation than enslavement to a particular technology.  There is no great need to go to zero feedback, or costly capacitors, or whatever the current fad is.  You can slip past all the fads and carnival barkers, and just do your own thing.  For me, that's electrostatic speakers, and high bias high current power amps.  For you?

Meanwhile, you don't have to bother with testing amplifiers, cables, or whatever.  Or, you can choose whatever cable you like, probably having no deleterious consequences (cheaper cables) except to your finances.  It's usually the more expensive cables that have actual audible differences from not actually being "high fidelity."

It's that freedom I treasure, and the freedom to make things as complicated as they need to be to do the things I want, and not be limited by the superstitions of others.

But as much as you can follow your own path within the fairly wide corridors of high fidelity, and still not be screwing things up, or otherwise, the last thing to believe is that you, just you, or your bud, have really found the answer that somehow decades of audio scientists and engineers have overlooked.  The last thing to do is take over every gathering with your own soapbox.  It's fun to believe you are on to something, but don't make it a crusade, which will only prove to others how wrong you are.  It may be true, but only with the lowest imaginable probability.  Getting others to accept a new truth now is only making a sale.  It is always better to let the come out on it's own, rather than pull it out in a magic show.  Unless you are selling something.

So the last thing I want is patronizing advice, unsolicited help, arbitrary or sacrificial rule enforcement ("all audiophiles must use audiophile cables", "you must listen for yourself"), killjoys ("you must spend more on music than gear"), or anyone trying to sell me something I didn't ask for (I can smell a sales pitch a mile off).

Not that I don't deserve it, a lot of it, at least most of it.

*****

My audio nut brother in law has often said that you don't admit to being an audiophile without being open to criticism.  That's the name of the game.  If you present something to listen to, be it an album, or an audio system, or even make a statement about audio reproduction, you should be open to and expect criticism.  It seems to me over the years of being in audio societies going back to 1978, people are less prone to criticize now.  Even my brother in law rarely criticizes individual audio nut friend systems anymore, if he visits them at all.

Actually, in the audio parties I've had, I haven't gotten very much criticism (except from those you might expect it from) and I might wish I had gotten more (from the people who were silent).

This is clearly a dicey area.  But I'd note that there does seem to be the situation that the people who talk are the ones least interesting to listen to sometimes.  Especially when they're selling something, have a need to prove something, etc.

Criticism is more appreciated than advice, but the latter is more commonly given, it seems.

One thing that is actually a bad thing, is audiophillia nervosa.  When people can't listen, because they think, without good reason usually, that it isn't sounding good.

Or if they feel the need to spend more than they should, constantly, etc.

Regular audio is a need but high end audio is a luxury, a rich mostly-men's game of some kind.  Probably a social substitute for unfufilled sexual needs.  It should not be the kind of additiction that leads one down the tubes.  It's excusable as a kind of bling.

And since it is really just bling anyway, it might as well be blingful, in the ways that resonate with my personality.

Anyway if criticism were to lead to audiophilia nervosa (I'm not sure it does, but "advice" might) that would be a bad thing.

Most of these superstion things, the "need" for special cables and amplifiers, and all the tweaks, etc., do not lead to bad sound.  And they should not be sold as such.  If sold at all, it should not be as a cure to an ill, but to elevate beyond.

And so one should always respect one's peers, or attempt to sound that way.  They are using their NOS dacs, their underpowered SET amplifiers, whatever, that's they're way, it's not like they need conversion therapy (which doesn't work anyway, as I understand it).

Leadership should generally be by example, demonstration of new and better, rather than cutting anyone down.  Or any thing except as part of the argument at hand.

Some people have the sense of coolness which is a lack of uncoolness.  That can be very tricky.  When cooties are found here, they are found over there as well, there's no end to cooties.  You call my Krell complicated...I'll show your Pass Labs is plenty complicated also.

My own sense, I feel fortunate in believing, is more in the sense of coolness being cool things, of which there might be a few duds, but still enough cool to win the day.

I think that's the way Nelson Pass himself feels about things.  He's just playing around, he invites others to join in his game, but he's not telling anyone else they're out of line.

He's praised simplicity, and he's pointed out that simple circuits generate relatively lower order harmonics, which are less audibly offensive.  The higher order harmonics can be horribly offensive.

However, a well designed amplifier with correctly applied feedback (he uses feedback too, and does measurements), even if it is complex, with reduce higher harmonics to past the point of audibility, to below the noise floor.

The "simple" amplifiers many people love, SET's, often generate distortion harmonics up to 3% and trailing downward through all the harmonics you can count just by slightly lower amounts.  Isn't it more a factor that there is *so much* distortion?  Yes, provably the distortion in a high distortion amplifier is audible in blind testing.  Does having higher proportions of the even harmonics somehow make the other better?  Actually, the asymmetric non-linearities behind even order distortions are disproportionately capable of generating IM, which is the worst of all.  The only good distortion is no distortion.

There is no credible evidence vanishingly small amounts of high order distortion produced by more complex circuits is audible.  Pretty much when you get THD below 0.03%, measured under all conditions of interest, there is no audible "distortion," regardless of the order of the distortion or the complexity of the amplifier.

Nelson Pass does not do blind testing, nor John Curl, nor their followers.

They purport to hear the result of complexity, but it does not seem possible that they actually do, in well designed but complicated competing amplifiers like the Krell.  And you will find plenty of complexity in Threshold and Pass Labs amps too.  It's Nelson's best efforts, the SA/1 and the XA's are very fine amplifiers, and I'd love to have one appropriate to my needs.  But there is no reason to believe they are unique in regards to their actual performance.  Other companies are able to obtain lower orders of measured distortion, etc., without resorting to Class A operation or any other of Nelson's clever circuit designs.  There is no objective evidence, either from measured performance or statistically valid blind testing, that they are actually better amplifiers in any way, than other High Fidelity amplifiers of the same power.

And there is certainly no proof.  I don't find either the white papers suggested by Curl or Pass to be fully convincing.  Interesting yes.  Convincing, no.  Clever circuits do not enable insufficient use of feedback to achieve "better" performance.  Correct design of all aspects, both the amplifying and the feedback, is required.  And once you get to a certain goodness, it's quite like more good doesn't help much if at all.  Numbers matter, and the lower the better.  What's only possible is the selective reporting of information, for example rising distortion at high frequencies that intermodulates to lower ones.  It's always possible for specs to be misleading and incomplete.  They are always incomplete.  Measurement can be incomplete, but in the area of amplification all aspects of performance are tractable, and pretty well understood.

So I will not bother with these "simplicity" superstitions.  I've already created too many superstitions and money pits of my own.  I should have spending more time in measuring and adjusting speaker equalization and the like.  I might have been fine with a less costly to maintain amplifier, such as one from ATI or Bryston.

Not to mention, finding better ways to get the music I want playing.

*****

WRT the purported amplifier test, it was as invalid as any sloppy test could be.

Not only was there not serious attempt to match levels, during the interregnum of changing amps I had noticed the preamp balance was off, so I fixed it.  I did not know a vote was going to be taken.

It was a fully sighted, in group test, not level matched, inevitably steered by starting positives from the fans.  And certainly the 2A as a separate heatsink adorned amplifier looked more solid than the highly ventilated and curvey Krell integrated--a very different aesthetic.

I take these sorts of things as jokes.  I voted because I fixed the level.  I'm sorry if it played out that it was a club members amplifier (I had thought it to be the stores), if he felt badly about his amplifier, as a result.

I'd rather not take a very forceful position with other people.  I try to accept them, as far as I can.  I will answer such questions, generally, as to whether A or B is a better sounding amplifier, despite not having any good basis of comparison directly.  I could always add "but it was just because I never hear anything the same twice, or I'm biased toward this kind of thing in a sighted test", etc., but I don't.  Why should I bother, except to be a snob?  A sloppy test deserves no better than a sloppy answer.  Nobody should be fooled.

So in the subjectivist Rome, I just follow along as the other Romans do.  No one should require me to do otherwise.  Nobody should require me to be the asshole who always says no.  If they want to be it, If they want to stand up for real science or whatever, I'm fine with that too, to a point.

But, the thing I think at least now I try not to do, necessarily, is sell my tonic.  I might make a comment that my way is better, but not that you, especially you, need it now.

I find myself increasingly put off by people who are consistently trying to sell some idea.

Though perhaps it's still more interesting than when nobody says anything.