Static THD measurements alone won't tell you how an amplifier will sounds when driving a loudspeaker. It's far more complex.
56 responses Add your response
Specs are important to making sure you aren't getting ripped off. It's easy to rate a 35 Watt amplifier as a 100 Watt amplifier if you don't mind 10% distortion. Used to happen very frequently.
Accuracy is also not necessarily what you buy. You buy amplifiers that make you feel good, whether it's in how they reproduce music, hos sexy the tubes glow, or how difficult their name is to pronounce.
THD is a bad thing, but it's not the only bad thing. You can make almost any solid state amplifier have vanishingly low distortion if you raise the feedback enough.
Einstein said that not everything that matters is measured, and not everything that is measured matters. We are still very much in that place when it comes to electronics. There's a separate thread going on about Hafler. Measured great, boring and lacking in dynamics as you could possibly want. I don't know why. I'm sure with enough time and effort we could discover this, but no one has put in the time yet.
One criticism of amplifier measurements is also that we don't often measure linearity or have enough of a measure to explain how they would work with a particular speaker. That is, we can measure speakers, and we can measure amplifiers, but we don't have really good stats for predicting how they will work as a system, and how to correlate this to perceived behavior.
So, always listen for yourself, and spend the least amount of money you can to make yourself happy.
You answered your own question -- the specs steered you into spending money towards the advertised distortion spec. Marketing strikes again.
THD specs can be misleading if they are taken at one frequency at a very low power output. THD, in general, changes as frequency and power output changes. I don't see how anything can be a flat zero distortion and if there is I can't see how that would sound any better than 0.01% at full load -- which is way below the threshold of hearing.
Tube amps generate second order harmonic distortion which is one octave above the fundamental, so both frequencies cannot combine and interfere with the signal from the source. This is like having the same instrument playing two separate tones in unison an octave apart, resulting what appears to be a warmer, fuller sound. But it's still distortion even though it can sound great.
geoffkait..My point is they BOTH sound great in my two examples of amplifiers having the opposite end of THD specs.. Arnet you the guy ( asstrophysisist )
selling those wacky trinkets to fool the mind into thinking the sound is better? Yeah I remember buying that stupid clock with a round magnet on it. and you banging your finger on the phone to activate the sonic realm of nothingness.. lol. If not you disregard.
One of the best amplifiers Kondo Audio Note $150k Kagura has 5% THDI'm on the same track as erik_squires & when Almarg & Atmasphere see this thread I believe he will also be on this track. In fact all 3 of us have written very similar things in the past. Total Harmonic Distortion (THD) in an amplifier is a function how much global negative feedback you apply in the amplifier circuit. The more heavy handed you are with GNFB, the better the amplifier measures on the bench but the worse it sounds sonically because vanishing low THD creates other issues, for example, Transient Intermodulation Distortion (TIM).
So, what I can conclude for the Kondo AN Japan amp spec is that very little GNFB has been applied to this amplifier & to make this amp sound very its best one would need a heavily damped speaker that would put less stringent requirements on the amp itself to control the woofer's back EMF. I get this info by reading the Kagura spec.
SN/R 133 by Devialet and a lot less by Kagura.this is again an effect of the high GNFB applied in the Devialet. The higher distortion in the Kagura creates a higher overall noise floor thereby degrading the SNR (vs. the Devialet). Having written this, in home 2-ch systems, unless one has taken extreme measures to keep distortion from AC power, foot falls, feedback from equipment racks, rumble from TTs to a very low magnitude, the typical SNR are 90dB or less. So, having 133dB is great by the Devialet but most home 2-ch systems are unable to avail of this. In fact, the SNR of a typical turntable is in the 55-65dB range & most vinyl/analog Audiogon members are in heaven spinning vinyl (over spinning CDs).
If the specifications are not that important to the sound, why list them?listing specifications is a great way to ELIMINATE equipment from your short-list. It basically tells you what you should not bother investigating further.
With the Kondo AN Japan Kagura that might be a bad decision simply because of their stellar sonic reputation over the years. But Kondo AN Japan is an exception. There are other exceptions to this rul-of-thumb such as AirTight & FM Acoustics. There must be more...
By looking at specs of Devialet and NOT knowing the prices of Kagura and Devialet I would of gone with the Devialet just based on specs alone for the impression of it being a great sounding amplifier. Ive read other discussions on forum and cant quite get a handle on why BOTH amplifiers sound great."sound great" is a nebulous term. It's been said to "sound great" by the press. Do you personally know these people in the audio press? Do they share the same listening habits with you? Do they listen to the same music as you? Do they have a similar room as you? If you answered "no" to any of these questions then you do not know these people hence you cannot trust their word. Very similar to my telling you that Chevy Cruz is an excellent car & that you should buy one. Would you go out & buy one? Why not? For the same reason you would not bank on these 2 amps sounding great. You would need to get around some (friends, dealers in your vicinity, dealers when you travel for business, audio shows, audio clubs,etc) & see if you can hear one or both to make up your own mind. Meanwhile you short list these 2 amps as "investigate further".
jeffkait..My point is they BOTH sound great in my two examples of amplifiers having the opposite end of THD specs.. Arnet you the guy ( asstrophysisist )
selling those wacky trinkets to fool the mind into thinking the sound is better? Yeah I remember buying that fucking clock with a round magnet on it. and you banging your finger on the phone to activate the sonic bullshit.. lol. If not you disregard.
I'm getting a bad feeling...
derrickengineer OP25 posts07-30-2016 3:21pm My point is they BOTH sound great in my two examples of amplifiers having the opposite end of THD specs.. Arnet you the guy ( asstrophysisist )you seem to be an a$$hole! :(
keep you language civil here & if you cannot then don’t post here.
Plus, you seem to be very new (with 25 posts). If you dont know this already then, most of the time, we have a congenial attitude & mutual respect towards each other unlike other aggressive audio forums. No need for you come here & mouth of. If you absolutely must then go elsewhere. While i might not be the moderator, I resent the fact that you're upsetting the congenial atttitude of this forum....
if you are going to fix your attitude, disregard.....
Bombaywalla. you just put the whole audio review business out of business.lol. If it wasn t for the tons of audio mags I read as a kid growing up, now 54, I might have not been interested in high performance audio.. I appreciate their point of view and take it as research about an item , not the bible on that item reviewed. Two different reviewers might have two different opinions on the same audio piece of gear. Me as well. Some information is better than no information. lol.
Your right bombaywalla, this is a new account that replaced an older one on Agon.and have had little time to post and learn from others experience and knowledge. I dont get what you mean by investigate further,, and short list.. They were just examples that I saw today and thought Id ask about the two extreme THD specs on two great amps.My edit clock ran out..
Derrick, I think Bombaywalla was responding to the Geoffkait post.
Also, I am I am close to your age and remember, too, all the audio mags with lots of specs. I also chased my tail those days, but was lucky enough to live in NY, where I could listen to a number of unattainable sound systems, where upon I learned to listen with my ears, rather than look at the specs.
Bombaywalla posted some good information, but let's hope for Almarg and Atmasphere to chime in. They, too, can provide complete and well informed opinions that anyone with a brain, will know to respect.
I know he was re my comment to geoffkait. Well me feeling I was sucked into Mr. Kaitts mumbo jumbo and $ 250 rip off prompted my reply.to geoff. 10% THD sounding better than 5% yeah that speaks volumes...lol
My post was of the Sound quality vs specifications and my ears will make the final call on a piece of gear to buy. Hard to do when buying over the net.. Hence the great opinions of Audio reviewers, owners and forums are the key to making a sound purchase decision, pun intended. Cheers and thank you all for the great info, appreciated.
I know he was re my comment to geoffkait. Well me feeling I was sucked into Mr. Kaitts mumbo jumbo and $ 250 rip off prompted my reply.to geoff. 10% THD sounding better than 5% yeah that speaks volumes...lol
I went out for a mild mannered troll and got a sucker on the line.
Derrick, I agree, that buying over the net makes things harder, but, speaking for myself, and perhaps for you, too, having had time to 'listen with our ears', I think this puts us ahead of the rest. We know what we like and know when we hear it.
Perhaps the best part of buying on Agon is that we aren't buying at full price and if things don't go as planned, we can sell without loosing too much money
- And money, is the greatest barrier in this hobby. At least for me.
(Though I tend to gravitate to the 'best bang for the buck' equipment).
I dont get what you mean by investigate further,, and short list.. They were just examples that I saw today and tderrickengineer, investigating further & short-listing would be actions one would do if one was serious about buying. Somehow I got the impression you were in the market for amps hence your questions about these 2 amps.
looks like you are not in the market to buy but are merely asking questions to clarify.
Even then "investigate further" would mean try somehow to listen to one or both amps & judge for yourself if the reviewers words ring true for you or not.
And "short-listing" would mean that you make a note of them somewhere so that when/if you are in the market to buy you look more into these 2 amps. With sooooooo many amps in the market it’s hard to know where to start - one can be stumped on this matter. I’ve found that having a short-list of amps that have come to my attention while trolling forums, hearing at shows/friends, heavily discussed in forums makes the start for the search for a new amp much easier - it’s better than having a blank sheet as I find that I can get swayed by "flavor of the month" amp.
Hope this clarifies.....
Doesn't matter. Listen carefully to the amp using YOUR preamp and YOUR speakers. Your preamp needs to be able to drive the amp, the amp needs to be able to handle the load presented by your speakers. The amp needs to be able to provide the range of listening levels you will use.
Nothing else matters.
You can get minuscule distortion figures (.0001%) on a poorly designed amp with massive amount of GLOBAL feedback, and they sound rubbish.
The best is to get minuscule distortion under (.1%) without any or very little global feedback if possible. Some very good amp designers can get this with just LOCAL feedback around the input stage which leaves the output stage free of global feedback, this to me is the best way to go.
Amps that have 5%THD are not well designed, and are for the the crew that love their euphonic coloration's depending on the type of THD.
Gdnrbob & Bombaywalla, thanks!
I am in essential agreement with most of what has been said above. I would commend the following thread to the OP and anyone else who may be interested, in which I, Atmasphere, and others posted on the subject of amplifier specs in general and THD in particular:
One of the basic points I stated in that thread was precisely what Bombaywalla said above: Specs and measurements, including THD, can often be useful in RULING OUT potential candidates for purchase from consideration, in some cases by identifying specs that may be TOO good, and thereby may signify the likelihood of undesirable design tradeoffs such as excessive use of feedback.
Personally, regarding the specific amplifiers that were cited in the OP, if I were amplifier shopping I would probably not include either one on my short list, regardless of what the reviews may have said and even if cost were no object. An amp having a 5% THD spec, no matter how euphonic and pleasing it may sound on initial audition, would over the long term cause me to unduly focus during my listening sessions on how untrue the amp is being to the recording, rather than simply enjoying the music. And an amp having vanishingly small THD would over the long term cause me to unduly focus during my listening sessions on whether it might be introducing excessive brightness, sloppy transients, and other effects that can result from heavy-handed application of feedback, or from other techniques that may have been used to achieve those numbers.
As I see it there are more than enough well regarded amps to consider, in all power ranges, that have specs that are neither "good" nor "bad" to a degree that would raise concern. Given that, I would see little point in pursuing ones having specs like these.
geoffkait3,776 posts07-30-2016 4:58pm
I went out for a mild mannered troll and got a sucker on the line.
Well since you said it mr. kaits, is there a bit of quilt coming out from your magic clock thingy or by stating " got a sucker on the line" you know full well there's nothing magic about your magic clock ?
I would not rule out any amp based on specs. I would trust my ears. If something sounds good to me for 15 minutes, it will sound good to me for 15 hours as well.
There are those who think that with right speakers Kondo amplifiers are among the very best, especially for a particular kind of music. Ruling it out altogether would be unwise, I'd say. The biggest distortions we get are already there in the recordings, and the better your equipment the more you know it. And if you can't hear it - either your hearing is no good or you are in a state of denial.
How would you know that a manufacturer is not fudging the specs ? If an amp is not reviewed and tested you would never really know if the listed specs are accurate.
good point, but I doubt they would be quite so bold, you know, what with audio magazine reviewers who do measure amplifiers for THD among other things. It probably wouldn’t be too good for business if it turned out they fudged the specs. And if the amp mfgr intends to be a viable business he needs reviews, no? But more to the point, we (editorial we) found out almost 40 years ago that THD specs are meaningless. That was when certain amplifiers with exceptionally low THD sounded terrible and certain tube amplifiers with relatively high THD sounded sublime. So, this debate was actually concluded 40 years ago, regardless of it’s troll appeal over the years. Amps won’t sound the same in different rooms and different systems, anyway. There are too many variables.
geoffkait3,776 posts07-30-2016 4:58pm
"I went out for a mild mannered troll and got a sucker on the line."
to which drrekengineer replied,
"Well since you said it mr. kaits, is there a bit of quilt coming out from your magic clock thingy or by stating " got a sucker on the line" you know full well there's nothing magic about your magic clock?
That was almost a complete sentence. When you cast the bait out there you never know who's going to bite. - old audiophile expression
This is an interesting topic personally as I’ve had the opportunity to hear the Kondo Kagura at CES a couple of years ago. It was in the context of an all Kondo system to best of my knowledge. Verdict, superb natural sound is what I experienced. Pure,clean and transparent with full body tonality and harmonics.
Derrickengineer where did you see/verify the THD specification of 5%? That seems high and I wonder how that was measured. I believe that the Kagura has a switch to adjust NFB levels (output impedance ). I recognize that 2nd order distortion is congruent with nature and is readily tolerated by human hearing (unlike upper order odd distortion ) , Ralph (Atmasphere ) has clearly and thoroughly covered this subject very well on several occasions.
I’m skeptical about the 5% THD stated by the OP based on what I heard. There wasn’t a hint of an euphonic, colored , overly warm character at all. Stings,brass,woodwinds and piano were exceptionally realistic in presentation. Shortly after this listening session I visited the Pass Labs room,. This consisted of the XS 300 amplifiers driving the Marten Coltrane (Supremes I believe ). Without question the XS amplifiers will spec at much lower THD than the Kagura.
IMO the Pass and Marten combo was quite good but this system lacked the tactile flesh and blood realism compared to the Kondo suite. I was more aware of the fact of listening to a fine collection of electronic audio components. In contrast the Kondo seem to put me in the presence of live musicians. The two systems were very distinctly different. However the THD is determined there is certainly more to this story that is unexplained with numbers /measurements. Bottom line, the Kondo system was more emotionally engaging and convincing in that it simply pulled me deeper into the music.
My gut feeling is that although the Kagura has higher THD than the esteemed Pass XS 300 I suspect that the number is very likely less than 5% as reported by the OP. This is just my opinion after listening to these two ultra expensive systems virtually back to back. Inna is right, you just have to actually listen to components and decide based on what you hear. That’s why YMMV will always ring true. That Einstein quote posted by Eric is so on the mark. Audio components are built based on science and engineering principles no doubt whatsoever, yet there is a genuine component of "art" at play as well. We can clearly hear things that can't be fully explained at this juncture.
Charles, you raise a good question about the 5% THD number. I looked into it a little, and as you can see here, near the bottom of the page, that is what the manufacturer specifies. And that set of specs is quoted in various reviews.
As I’m sure you realize, the Kagura is a SET amplifier utilizing a parallel pair of 211 tubes. In looking further, I found datasheets for some incarnations of the 211 stating operating conditions in class A audio frequency applications that include an output power (per tube) of typically 12 watts **at 5% second harmonic distortion.** See the following:
I’m just speculating, of course, but perhaps the manufacturer wanted to release specs prior to concluding developmental testing and/or evaluation of a meaningful number of production units, and therefore simply chose a very conservative number that they could feel completely confident would be met. And perhaps they chose that number simply based on what appears in the 211 datasheet.
On the other hand, though, the 5% number in the datasheets is just second harmonic distortion, so presumably a corresponding spec for THD would be at least slightly worse. And the fact that the Kagura’s 50 watt power capability spec reflects 25 watts per tube, rather than the "12 watts typical" indicated in the datasheets, would also seem to suggest that the 5% THD spec may not be all that conservative.
Regarding your and Inna’s philosophy that "you just have to actually listen to components and decide based on what you hear," I of course agree (aside perhaps for the word "just"). However there is only so much equipment that any individual will have an opportunity to hear, i.e., that can be put on his or her short list. Given that, my own philosophy is why take chances by including on a short list components for which there is cause for concern, whether it be due to specs, published measurements, widespread controversy about its sonics, excessive criticality of system matching or setup, manufacturer reputation for customer service and support, or any other such factor. Especially if the short list indeed needs to be "short" in a particular case, or if an audition needs to be relatively limited in scope or duration.
Simply put, my preference is to play it safe when it comes to deciding what to short list. But that is just my own preference, and I don’t consider it to be any more valid or less valid than anyone else’s approach.
I find it interesting that most of the gear in my system measures poorly. I have an all tube system I also use a ladder dac that measures worse than any delta/sigma dac. Yet my system sounds more real, more musical then almost any other better measuring system I have ever heard. Measurements tell you nothing about how a component will sound
I can not disagree with your well presented and rational reply. To be perfectly clear I do in fact believe that specifications and measurements have relevance and are useful. Particularly for screening purposes and initial "short list " selection. You have to begin somewhere and as you note you cannot audition every component that may interest you.
Really my prime contention is these parameters simply aren’t reliable in determining the sound quality of an audio product. Specifications can identify and likely eliminate gross incapatible matches, no question.
However if an opportunity exists to hear the product of interest then this is by far the best way to proceed.
The 5 % THD of the Kagura would rule it out for perhaps many yet the reality of listening to it reveals stunning performance. Given what I heard makes me curious to see a verified THD measurement. If it is truly 5% then it would lead me to question the importance of THD in the overall scheme scheme of things.
There is a thread on this site about amplifier choices for the Wilson Sasha. The OP was considering Pass Labs and Cary 211 tubed amplifier. There were numerous well reasoned arguments to choose the Pass rather than the Cary. The OP ultimately bought the Cary and is ecstatic with it driving his Wilsons. In fact he describes it as "fantastic and magical ".
This type of result isn’t predictable via relying on specs alone thus my science plus art comment above. What ahendler posted has been confirmed by more than a few experienced listeners.
Specifications are of value but other variables are undeniably at play.
To all - most likely the 5% for the THD of that amplifier is a typo or whatever! maybe language issue! who knows. The generic great sounding tube amps back when this whole THD was big like 30 or 35 years ago came in with what was then thought to be relatively HIGH THD circa 0.5% THD or perhaps 0.05% THD, which IS relatively high compared to the Japanese amps and other SOLID STATE AMPS that were touted as having vanishingly low distortion were coming in circa 0.0005% or maybe even less in some cases. I.e., it was a tube amp vs solid state amp argument primarily. So, I kind of doubt the 5% number. It doesn’t make sense. And you know what Judge Judy says - if it doesn’t make sense it’s not true.
Some forms of distortion are often associated with a distinctive sound that may have unique appeal and sound. Performance and sonic appeal are not the same thing.
Think of digital photography. There are many ways to make a photo that is accurate in terms of its representation look better or worse. In either case it’s a form of distortion in that the original is less accurately presented.
What one likes or not is almost always a subjective determination. Accuracy alone may not be the full ticket. But accuracy is objective and a good place to start always I think. In my case I think I strive to end up there as well. I want to hear the variations in recordings not have my gear give them a particular flavoring. Especially since I listen to all genres of music. Old and new. So much variety in each piece of art to hear. No special sauce needed.
Revealing the nuances and different character of individual recordings is exactly what the Kondo Kagura excelled at doing so impressively. No "special sauce" required. This along with its clarity, openess and naturalness is what made it stand out relative to the Pass-Marten pairing. This combo did a very good job of sorting out recording differences also The Kondo room/system just did it better based on my listening encounter.
Something to keep in mind is that "harmonic distortion" is a way of measuring a distortion from an amplifier, you don’t play music, for instance "at 5% distortion". The amplifier is driven by a single frequency at a time to some rms voltage (or power) output level (typically), under load of some kind, and the level of harmonics relative to the drive level are measured. Music isn’t at any constant rms voltage level, and not at one frequency at a time, so the conditions of the test seldom occur at all, or if they do, they do only for a very brief instant. Usually the harmonic distortion is given at the upper end of the amplifier’s output power capability, as that is where it is nearly always highest, and is given to indicate where the amp might be expected to be running out of juice.
With music, you might hit that peak output power for maybe a millisecond or so, and unless there is a lot of it hitting or exceeding that peak, you won’t be experiencing the effect that mght occur up there. The average power with hifi music is way lower than the peak, so if anything, the relevant distortion at the lower power would be what might be relevant.
The percent distortion spec, as usually given, should be read for the POWER level that produces it (i.e., to indicate what the manufacturer assumes is the maximum power the amp might do). NOT for the distortion level that the manufacturer is calling its limit. An amp with a soft clip can go near that maximum without distressing you because it takes longer to get there, and won’t suddenly go way higher in that region. An amp with lots of feedback (and a low distortion spec to match, used to indicate the power) will usually go from low distortion to high distortion with only a small increase in the power produced. The spec may be 0.0000x% where they call it, but it will probably be 5% or worse with the power level just a smidge higher when the amp clips.
Not making a comment on which is the better amp, just about the "distortion spec". Unless it is given at levels way lower than the amp can do, read it is as a POWER spec.
Your explanation makes a very compelling and logical case for the seemingly overstated importance /relevance of the THD specification. It certainly is consistent with the not infrequent observation of relatively poor measuring components that sound exceptionally well. Stereophile reviews are a good example. A reviewer says a component sounds terrific, yet subsequent test bench results are average or even poor. Conversely there are examples of excellent measuring components that sound bad. Again, this is why Einstein’s quote regarding what can and can’t be measured is timeless in its profundity.
It is fair to say however that some specifications and measurements are more relevant than others. On the other hand THD may be less meaningful and predictive than is generally belived. Good post bwaslo! An isolated measurement under test conditions may have little correlation with actual music signal performance quality .
From the website of things that Einstein never said,
8. "Not everything that can be counted counts, and not everything that counts can be counted."
It’s a pleasant point to be made, but Einstein never said this. Quote Investigator points to a 1963 paper by sociologist William Bruce Cameron: "It would be nice if all of the data which sociologists require could be enumerated because then we could run them through IBM machines and draw charts as the economists do. However, not everything that can be counted counts, and not everything that counts can be counted."
One thing Einstein did say is, "Education is what's left after you forgot everything you learned in school."
Bwaslo 8-1-2016 1:36 am edtThis is an excellent point, among other excellent points in Bwaslo's post. In looking further at the descriptions of various Kondo amplifiers I see that **all** of the SET and parallel SET models listed at their site (Kagura, Ongaku, Souga), and also the Gakuon II which is not listed as a current model, are indicated as having 5% THD at their specified maximum power ratings. Presumably if their max power ratings had been specified less aggressively (i.e., somewhat lower), the specified THD would have been significantly lower.
Interestingly, though, their Overture integrated amp, which is a push-pull EL34 design, has its max power spec based on 1% THD. Just guessing, but perhaps their thinking is that differences in how THD is distributed among the various harmonics when the two kinds of designs (DHT SET vs. push-pull pentode) are operated near their limits would result in say 2% THD in the push-pull EL34 design being subjectively more objectionable than 5% in the DHT SET designs.
On the other hand, though, as Atmasphere has pointed out on numerous occasions, for best sonics a SET should not be called upon to provide more than a small fraction of its rated power, as in contrast to push-pull designs the distortion produced by a SET becomes vanishingly small at low power levels.
All things considered, for an application requiring just a small fraction of the Kagura's rated power I'll retract my previous statement about ruling it out from consideration on the basis of the 5% number. Although I would feel more comfortable in doing so if comprehensive measurements (such as John Atkinson provides in Stereophile) were available.
I'm not surprised that you recognized the significant and excellent points contributed by bwalso. I must say that I agree with your further comments. These SET amplifiers if used in their sweet spot of operation have exceptionally low distortion. I heard the Kagura driving efficient speakers at moderate or typical volume levels (definitely not lowered volumes). Most likely well within the amplifier's very low distortion range. This would to a large extent explain the wonderfully clear and natural presentation I described. bwaslo got it right with his post.
Interestingly, though, their Overture integrated amp, which is a push-pull EL34 design, has its max power spec based on 1% THD. Just guessing, but perhaps their thinking is that differences in how THD is distributed among the various harmonics when the two kinds of designs
Well, marketing does get into it, too. A SET amp, particularly one with no or low global feedback, won't have too much better distortion just below the power used to spec the 5%, while distortion of another amp likely will quickly get lower at lower powers. So, quoting at a lower power won't markedly reduce the distortion figure for the SET, but can reduce it a LOT for feedback class AB amps, and of course that looks better to a customer having only a data sheet to look at.
At the 20dB or so power leve below clipping either amp is used on average at, though, the distortion produced is likely to be low enough to not matter in either case (Geddes did some research showing that for lower order distortions - which pretty much all amps do before clipping - people are just not very able to hear it; other factors are much more audible).
At very low levels, though, a badly biased class AB amp (or one for which the bias doesn't track thermal well) can have quite a bit of distortion relative to the output level due to crossover effects. And crossover distortion (not speaker crossover -- class AB switchover distortion) is very high order and can be heard. Class A amplifiers won't do that at low level, (And, by the way, neither will class D amps.... its a class AB thing).
Thanks once again, Bwaslo. And thanks Charles & Derrick for the nice comments. For the record, the Kagura uses 3 db of feedback, which while of course not being an insignificant amount is much less than what is employed in many other designs, tube-based as well as solid state. From its description:
Kagura’s pure silver coiled output transformer handles output impedances of 4, 8 and 16Ω. By selecting positions at the speaker impedance switch and change relevant jumper plates, speaker impedance switchover is done with NFB level always set to our designed level –3dB.Best regards,