Stereophile Amplification Product of the Year

[Galliardist]

Your room and speakers will matter 1000 times more than inaudible distortion.

Maybe, but can a room and speakers ever make up for a highly and audibly distorting amp?

 

[DonH56]

Feedback suppresses crossover distortion far below audibility (and practically below measurability).

 

[Galliardist]

I think you simply don't understand how harmonics work. Doubling is doubling. It's an effect found in old-school piano playing, where the bottom octave of the instrument is goosed by playing the next octave up or down. In classical music it obscures the composer's intention. Having the next octave up being played when it wasn't intended gives off a different sound, a different harmonic quality, than a composer intended. Of course, with most species of popular music it can be attributed to the luxury of improvisation. But as regards fidelity to original intent, it's an obvious and undesirable form of distortion.

Doubling is different to the harmonic output of a bass instrument, though. In classical music, such things are often done in editing a piece to be played by more instruments in a larger hall. Doubling was a feature of consort music very early on, and court music before that. And in piano music, the doubling is usually down (because the piano had fewer keys when older works were written): that does not represent a second harmonic in any way I can work out. Other clues can give away what is going on: these are dependent on the instrument concerned.

And what the composer intended, is not really the key thing for the audiophile process. Our role is as the listener, one step removed. We aren't producing "what the composer intended" but "how the performer interpreted it". This is a big difference, but one that early record reviewers just didn't get, often writing bad reviews of good music that didn't agree with whatever pocket score they had managed to get hold of and sort of read. Moreover, we can hardly think in terms of the sound that Mozart would have heard, when the standard modern interpretation is on a recently built Steinway piano that is completely different to any Mozart would have access to, played in a considerably larger hall in many cases.

Also, what's this "of popular music" stuff? Editing, rearranging, transcribing, reinterpretation, improvisation and ornamentation have always been features of art music and making it work in different places with different instruments and ensembles. The mainstream interpretations of most of the great composers are all subject to this. Does every orchestra have a heckelphone as specified for some works, the "correct" horn for the original score and orchestras that would have played the work, or do they work with the woodwind and brass instruments the players have? Do you use period instruments, or go with the new improved versions? What about improved or more modern playing technique? Do you use the full power of a larger orchestra? What can you learn from the musicologists and experts that produced both good and bad scores for more modern performance?

All of these things feed into making, say, the Karajan interpretation of Beethoven's symphonies quite different from what the composer would have heard. And if we heard Beethoven playing his own piano sonatas, they would be quite different to a modern performance, not least because, again, a lot of the meaning of the score and the playing conventions of the time and later, have been lost or discarded: once again, those early record reviewers caught a point in that process and decried conventions in playing as "wrong notes".

Sorry for the rant. but I have the opinion that a lot of the errors we make in reproduction come from misunderstanding our role in the process. We are the listener, just as we are when attending a live performance. We reproduce the performance: we get to decide if the performance is good, matches the composer's intentions or the tradition that gives us the great performances we have access to, and so on. We are not the musicians, and nor are our systems.

 

[Robin L]

Doubling is different to the harmonic output of a bass instrument, though. In classical music, such things are often done in editing a piece to be played by more instruments in a larger hall. Doubling was a feature of consort music very early on, and court music before that. And in piano music, the doubling is usually down (because the piano had fewer keys when older works were written): that does not represent a second harmonic in any way I can work out. Other clues can give away what is going on: these are dependent on the instrument concerned.

And what the composer intended, is not really the key thing for the audiophile process. Our role is as the listener, one step removed. We aren't producing "what the composer intended" but "how the performer interpreted it". This is a big difference, but one that early record reviewers just didn't get, often writing bad reviews of good music that didn't agree with whatever pocket score they had managed to get hold of and sort of read. Moreover, we can hardly think in terms of the sound that Mozart would have heard, when the standard modern interpretation is on a recently built Steinway piano that is completely different to any Mozart would have access to, played in a considerably larger hall in many cases.

Also, what's this "of popular music" stuff? Editing, rearranging, transcribing, reinterpretation, improvisation and ornamentation have always been features of art music and making it work in different places with different instruments and ensembles. The mainstream interpretations of most of the great composers are all subject to this. Does every orchestra have a heckelphone as specified for some works, the "correct" horn for the original score and orchestras that would have played the work, or do they work with the woodwind and brass instruments the players have? Do you use period instruments, or go with the new improved versions? What about improved or more modern playing technique? Do you use the full power of a larger orchestra? What can you learn from the musicologists and experts that produced both good and bad scores for more modern performance?

All of these things feed into making, say, the Karajan interpretation of Beethoven's symphonies quite different from what the composer would have heard. And if we heard Beethoven playing his own piano sonatas, they would be quite different to a modern performance, not least because, again, a lot of the meaning of the score and the playing conventions of the time and later, have been lost or discarded: once again, those early record reviewers caught a point in that process and decried conventions in playing as "wrong notes".

Sorry for the rant. but I have the opinion that a lot of the errors we make in reproduction come from misunderstanding our role in the process. We are the listener, just as we are when attending a live performance. We reproduce the performance: we get to decide if the performance is good, matches the composer's intentions or the tradition that gives us the great performances we have access to, and so on. We are not the musicians, and nor are our systems.

You will have to excuse me. I was the recording engineer for the San Francisco Early Music Society for the better part of ten years and listen to classical music mostly. I'm into HIPP, so to speak. Yes, the Karajan interpretation of Beethoven's symphonies [would be] quite different from what the composer would have heard on account of Beethoven being incapable of hearing most of them [starting around the time of the Eroica], among a host of other issues, like HvK's fondness for creamy smooth textures. And I have a funny feeling that Beethoven would slap Roger Norrington silly for being so musically sloppy so much of the time if he could hear Norrington's idea of how the symphonies should be performed. Anyway, I'm currently listening to Claudio Arrau's 1960's Beethoven piano sonata cycle. I'm sure that even if Beethoven could hear Claudio Arrau play he would disapprove. Though I don't.

I was pointing to doubling as the sort of distortion of the performer's intent that can obscure the original sound. It's not as great as re-orchestration but it heads in that general direction. I listen to a lot of historical recordings. The further back one goes historically, the thicker the sound becomes. And I have owned plenty of tube gear, so I have a pretty good handle on what tube distortion can do to the tonal qualities of recordings, even modern ones. I've only heard Audio Note gear once, at a demo at a Stereophile show. Very intriguing but as far from reality as 1950's technicolor.

And yes, I'm aware that popular musicians make a lot of choices that have nothing to do with the issues of amplifier distortion. I've got hours upon hours of Charlie Parker in the collection.

 

[rdenney]

A performer is an artist, and we may judge the artistic intentions and outcomes by whatever filter is important to us. We may think of their performance as a collaboration with the composer, we may demand absolute compliance with the composer’s written instructions, or we may further demand compliance with the composer’s context. That’s us judging the art, which is fine. There are artists willing to accommodate us—Norrington and Hogwood among others come to mind. No matter what, though, performance is NOT playback. Composers wrote their stuff down with a lot of performance practice assumed and not notated. They expected a contribution to the art from the performer.

So, what does that have to do with playback accuracy? Beats me.

On the subject of crossover distortion, distortion is distortion—the product you have left when the gozinta is removed from the gozoutta, corrected to unity gain. A distortionless amp produces a null flat line distortion product. If my Class D amp’s distortion and noise is below 105 dB under the maximum signal, then what do I care what is making that distortion? If the power is abundant enough that it never clips enough to create distortion products above some similarly low threshold, then clearly crossover distortion is sufficiently minimized. The distortion product is effectively a null flat line. Complaining about its causes becomes theoretical.

(Distortion magnitudes of -100 dB are one ten-thousandth of the distortion at rated-output if the subject amp.)

Nelson Pass did not become famous for making 1-watt single-ended amplifiers. He was famous for making large, low-distortion amps. Likewise Carver, Bonjuorno, Gudgel, Kessler, D’Agostino, etc. But large, low-distortion amps are a dime a dozen now. My $650 Buckeye is cleaner and more powerful than any Phase Linear, Spectro Acoustics, Adcom, B&K or SAE amp. And all but a few Krell, Sunfire, or Threshold amps. And Quad, Bryston, NAD, etc., etc. And any of the big Japanese manufacturers. And higher-end brands than that, of course.

So, how’s an amp guru supposed to make a buck when low-priced amps can be so close to the state of the art? They go up-market into esoterica, because that’s where the elitist clubs are with people who have that kind of scratch and that kind of credulousness. But they aren’t going to sell a 350-wpc amp with vanishingly low distortion in a cigar box to the heavy hitters. They have to make “art,” either with visual aesthetics, or magic audio, or both. Magic is just technology we don’t understand. But if we do come to understand it, we may discover that the magic wasn’t better tech after all, it was sleight of hand. Nobody pays esoterica prices to mere geek engineers with short hair and last decade’s glasses.

Clearly, the selection of this amp had much more to do with the reviewer’s self-image of where he fit into the audio world than it did with actual performance.

It’s different than another class of luxury product I’m familiar with: wristwatches. But the watch industry is more honest—nobody claims that expensive mechanical watches are more accurate (more true to time) than cheapie quartz watches. They claim accuracy in the context of their technology, maybe, but mostly they are selling image and both buyers and sellers acknowledge and embrace it.

A magazine like S’phile has to cater to the image, because only the enthusiasts will subscribe, and only the whales will pay for what their advertisers sell. The only thing I hold against them is the big lie, that these esoteric brands they praise are uniquely true to the music.

Rick “lifting veils” Denney

 

[Hauxon]

Maybe, but can a room and speakers ever make up for a highly and audibly distorting amp?

"..highly and audibly distorting amp" Is it? and yes true, speakers and room acoustics won't bring anything back, especially what's not missing.

 

[simbloke]

Nobody makes a Class B audio amp. They are all Class A, AB or D (maybe G or H but those are rare now).

Indeed.

The first thing I was taught about transistor amplifiers was that this is what class B would be like and how nobody ever does that.

Followed by class AB and how to bias a transistor, then negative feedback.

 

[fpitas]

To further stir the pot like to add few more notes.

Some seem worried about the amp being so under powered that it will be constantly clipping with powerful passages. Note that DeVORE Orangutan O/96 speakers were used be the in the Stereophile review. These speakers sensitivity is quite high 96db. Most of us know that you need roughly double the power to increase loudness by 3db. So if we calculate backwards we find out 8w into a 96db speaker will give you the same SPL as 64W into a reletively inefficient 87db speaker. If the Klipsch La Scala sensitivity is as advertised 105db it would deliver the same SPL as 512W into a 87db speaker! So an 8W amp or even only 3W can play relatively loud without straining or clipping if fed into a sensitive loudspeaker. This is of course makes your choice of speakers very narrow.

As discussed earlier second-order harmonic (tube) distortion is not audible to humans to the same extent as the high-order harmonics (of a transistor). Even if the measurements look shocking I assume distortion is simply inaudible with the AN amp and thus not a real factor when evaluating it for listening. When distortion can't be heard other factors that affect the signal matter more, like the quality of the components, number of components and simplicity of the audio path. Here's where the Audio Note Meishu probably makes it homerun.

Another thing to take into account is the single ended class A design. Not a tube thing but design thing. Single ended class A does not cut the waveform into two parts like class B. This is inefficient but still is the purest form of waveform amplification. That's why many of our transistor amps are so called class AB where the are biased to operate in class A with low power and switch to class B for more power. Class B simply is inferior, the embedded image describes how crossover deformation of the signal is the result of class B amplification. People have found ways to lessen the effect but using class A you avoid baking the signal during amplification.
View attachment 330461

There is simply much more to amplifier performance than measured distortion. I suggest reading the following article by amp guru Nelson Pass where he talks about lot of what we have discussed in this thread. And note, he is a transistor guy and his company makes transistor amps.

Single-Ended Class A by Nelson Pass

What everybody else said about how no one makes class B amplifiers for audio. But also, if you're worried about crossover distortion, the clear winner is class D.

 

[Hauxon]

What everybody else said about how no one makes class B amplifiers for audio. But also, if you're worried about crossover distortion, the clear winner is class D.

Our class AB amplifiers are indeed class B amplifiers disguised as class A. They operate only for the first few watts in class A mode before switching to class B. Some amps are biased higher up to 20 watts in class A but at the cost of heat dissipation and inefficiency. I have an amp like that, Cambridge Audio 851W. It’s good to have it for the winter. If your class AB amp doesn’t run hot it’s simply mostly a class B amp. To run your amp more in the class A mode you can have more sensitive speakers. …as with tubes.

But back to our hearing and distortion. Our brains are most sensitive to distortion in the frequencies of the human voice and mostly at mid volume. At low volume and very loud volume we don’t hear distortion as well. The same goes for extreme frequencies. So our class AB amps mostly switch to B to play low frequencies where we might not hear the crossover distortion.

For class D I can’t have an opinion I can stand by since I still haven’t heard a top tier class D amp and don’t understand the tech fully. There are however lots of components on these boards ..so I’m a little sceptical.

 

[DonR]

Indeed.

The first thing I was taught about transistor amplifiers was that this is what class B would be like and how nobody ever does that.

Followed by class AB and how to bias a transistor, then negative feedback.

Yep. Electronics 101 or 201... probably 101.

 

[DonR]

But back to our hearing and distortion. Our brains are most sensitive to distortion in the frequencies of the human voice and mostly at mid volume. At low volume and very loud volume we don’t hear distortion as well. The same goes for extreme frequencies. So our class AB amps mostly switch to B to play low frequencies where we might not hear the crossover distortion.

No, they are always biased right down to DC.

 

[mhardy6647]

What everybody else said about how no one makes class B amplifiers for audio. But also, if you're worried about crossover distortion, the clear winner is class D.

McIntosh did, as I recall, make a number of amplifiers that operate(d) in Class B. Here's an example, from the late 1970s. Of course, it's output transformer (autoformer, I presume) coupled, too!
https://www.worldradiohistory.com/Archive-All-Audio/Archive-Audio/70s/Audio-1977-09.pdf (p 70ff)

Maybe they've stopped -- but of course now they're starting to use 3rd party Class D power amplifier modules anyway.

 

[fpitas]

McIntosh did, as I recall, make a number of amplifiers that operate(d) in Class B

I have to think that was marketing-speak, for the following reason. At 0 current a transistor (or tube) has 0 transconductance. Now, to eliminate crossover distortion, you need to ensure that the output stage, including all devices, has (hopefully) no change in transconductance as the output passes through 0 volts. In normal design that means picking an appropriate idle bias and ensuring it remains optimum over time and temperature etc. Each device at 0 volta out has about half the nominal transconductance. Hopefully I made my point clear.

 

[DonR]

"The bias network for the output stage... assuring Class-B operation" ... hmmm.

 

[fpitas]

"The bias network for the output stage... assuring Class-B operation" ... hmmm.

In other news: often glossy press releases are not technically accurate

 

[mhardy6647]

In other news: often glossy press releases are not technically accurate

I am shocked. SHOCKED.

 

[TLEDDY]

“ Looks like a gilded outhouse in an era of indoor plumbing.”

OMG! ROFLMAO!!

Truly… almost soiled myself on this line! Thank you!

Tillman