All power amps “good ones” sound the same.

[user 37518]

Speaker technology and acoustics have also both advanced over the decades.

Amps not so much, yes there is the new Class D hype with GaN transistors and such but that doesn't mean lower distortion, in many cases its quite the opposite, what the new amps definitely have is higher efficiency.

 

[Gold]

When seeking answers it's important to ask the right questions.

Yes. Otherwise you get answers like the conclusion of the AES paper quoted above. I would summarize it as “most amplifiers are good enough for most people”. Bravo. Brilliant work.

 

[user 37518]

I just realized that there is no study, objective trial, or coherent argument that will convince you otherwise, I'm beating a dead horse with these discussions.

 

[Gold]

I just realized that there is no study, objective trial, or coherent argument that will convince you otherwise, I'm beating a dead horse with these discussions.

Convince me of what? I haven’t taken issue with anything in either of the AES papers. I certainly agree that most amplifiers are good enough for most people.

 

[scott2000]

Out of how many?

should've said I got 5 on Paul... 

 

[Gold]

I’m not a numbers guy but it seems in the first AES paper there was variation noted with the two amplifiers but it didn’t rise to the level of statistically significant.

Is it so hard to imagine that by cherry picking equipment you could swing those numbers over the threshold?

 

[Gold]

A study that would be useful in settling this question would be to do multiple studies with multiple speaker and amplifier combinations.

I can write the conclusion. Only the study needs to be done.

The audibility of speaker and amplifier combinations falls along a bell curve.

 

[Gold]

The second AES paper is an ode to mediocrity. The middle of the bell curve. That’s fine for commercial interests but doesn’t exactly advance the state of the art.

 

[Winston OBoogie]

Looks like I missed an amp party while on sabbatical (chill fest) at my girlfriend's place.  Hope y'all had a good shindig

How are we defining "good ones"?

For me:

There are quite a few examples of up-to-date good circuit designs in Bob Cordell's amplifier book.  Fewer, but nevertheless, fine examples in Doug Self's book on the same subject.

I've borrowed bits from both of those books -  as well as from the writings.musings of other folks such as John Curl, Nelson Pass et al., Peter Baxandall who took Cherry's output inclusive miller comp. and made it transitional output inclusive miller comp etc.,  our own late Brad Wood, and many others. 

I've built several small power amps to drive my headphones with these bits and bobs that I’ve learned and played around with.

Since I know my headphones quite well, and don't have much of anything else in the signal path besides a DA converter (a quite average one btw), I'll say "good ones" (power amplifiers) do NOT sound the same.

It may be that in a studio with all the usual distractions, the creeping volume, numerous devices in the monitor path, they'd all sound similar enough that I'd feel differently.

Anyway, someone who is a good mastering engineer and knows his system well (speakers, room, doo-dats..) would probably be a better judge than me.

That's my story anyway.

 

[5v333]

wouldnt it be hard to judge an amplifier from listening since the distortion of the speakers dominates?

and by saying so it would also be easier to judge an amp with really good speakers/drivers than with crappy speakers/drivers.

 

[JohnRoberts]

wouldnt it be hard to judge an amplifier from listening since the distortion of the speakers dominates?

Yes, typical loudspeakers make orders of magnitude more distortion than "most" amplifiers, that said amplifiers can make several different kinds of distortion that will not be masked by typical loudspeaker distortion and can be audible. One obvious such distortion type is crossover distortion from starved, under biased class A/B output stage. Crossover distortion is most apparent at low levels on high frequency content. Another audible distortion is saturation/clipping, but this is harder to hear. Occasional clipping on brief transients is hard to hear as distinct distortion, while severe clipping that corrupts the lower frequency audio envelope is easily audible. A third type of distortion not concealed by loudspeaker distortion is current limiting. This can be hardest to hear as it only occurs during difficult musical passages, that depends on loudspeaker load, and amplifier protection circuitry.

As I've shared before the qualification for amps sounding the same is only when they are operated completely in their linear region. i.e. not clipped, not current limiting, and not faulty (crossover distortion is easily managed). 

and by saying so it would also be easier to judge an amp with really good speakers/drivers than with crappy speakers/drivers.

I would say it is never easy to hear flaws in a modern amplifier operated properly. What you can hear will be dominated by a) program material, b) loudspeakers, c) room acoustics, d) and finally expectation bias.
-----
As I shared before decades ago I experienced wildly different magazine reviews for my very accurate phono preamp from reviewers listening through their "reference" systems. Since every single link in those "reference" audio chains was weaker (less accurate) than my preamp, they were effectively listening to their systems. The reviewer who heaped praise on my preamp obviously had a superior system, compared to the reviewer who thought violins sounded like sawing on wires. :-( 

Likewise modern amplifiers are more accurate than the rest of most playback systems.

This is a very old debate, apparently not settled yet in the brewery. 

JR

 

[gridcurrent]

wouldnt it be hard to judge an amplifier from listening since the distortion of the speakers dominates?

and by saying so it would also be easier to judge an amp with really good speakers/drivers than with crappy speakers/drivers.

no dog in this hunt, but offering some opinions and observations.
I've built A/B boxes which is a story unto itself as they introduce a measure of degradation.
Into a consumer loudspeaker system, at moderate levels, McIntosh MC-75's were very hard to pick from a then new Marantz 510m (massively more powerful).
Into a high efficiency system, most certainly less "linear", the comparison was not night and day; rather night and next week.
Not to pick on users of modern monitoring systems fitted with direct radiating drivers, but it appears differences between amplifiers powering such are less obvious.

 

[Winston OBoogie]

I will second what John said for the most part
But sooner or later, the amplifier has to leave the confines of the lab bench and go out into the big wide world and be used in situations and systems beyond the designer’s control.

I’ll assume most of these good amps have about the same dB voltage gain , and for a given power output into 8 ohms nominal, the supply rails for each are within reasonable comparison.

But are all these “good” amplifiers biased the same?  Do they all have the same nominal unity loop gain frequency? Is there loop gain enhancement?  Identical 19+20kHz CCIF IM test results?  How’s the common-mode distortion?  Does the output driver stage take a dump on the VA stage when things go a bit tits up?

And so forth...

*Edited for auto-correct mishaps.

 

[JohnRoberts]

I will second what John said for the most part
Sooner or later, the amplifier has to leave the confines of the lab bench and go out into the big wide world and be used in situations and systems beyond the designer’s control.

I’ll assume most of these good amps have about the same dB voltage gain , and for a given power output into 8 ohms nominal, the supply rails for each are within reasonable comparison.

But are all these “good” amplifiers biased the same?  Do they all have the same nominal unity loop gain frequency? Is there loop gain enhancement?  Identical 19+20kHz CCIF IM test results?  How’s the common-mode distortion?  Does the output driver stage take a dump on the VA stage when things go a bit tits up?

And so forth...

*Edited for auto-correct mishaps.

HF two tone IMD is a relatively mature spec now, it wasn't so well known back decades ago. I had to roll my own 19k:20K IMD test by modifying an old SMPTE heath kit analyzer. I found this very revealing for phono preamp design where the RIAA EQ's low pass filter tended to conceal HF nonlinearity by rolling off the higher THD harmonic content (a case of measured good and sounded bad because we were measuring the wrong stuff).

This is pretty well understood now (maybe not) and included in modern bench test systems. An old studio trick to sniff out HF IMD is to jangle a key chain into a microphone... the keys would make a bunch of higher than 20k content that would test the HF linearity of marginal audio paths. HF IMD sounds like thumpy mud riding beneath loud HF content. Again easy to hear if you know what to listen for but much more common several decades ago, than now.

JR 

 

[Gold]

no dog in this hunt, but offering some opinions and observations.
I've built A/B boxes which is a story unto itself as they introduce a measure of degradation.

My console has a passive monitor. It’s a full H pad with a 10K input impedance and a 5K output impedance. Into a 50K load. Source impedances are carefully managed.

Into a consumer loudspeaker system, at moderate levels, McIntosh MC-75's were very hard to pick from a then new Marantz 510m (massively more powerful).
Into a high efficiency system, most certainly less "linear", the comparison was not night and day; rather night and next week.
Not to pick on users of modern monitoring systems fitted with direct radiating drivers, but it appears differences between amplifiers powering such are less obvious.

Although a Hypex is a fine amplifier it wouldn’t be my first choice for powering a pair of A7’s.

 

[Winston OBoogie]

HF two tone IMD is a relatively mature spec now, it wasn't so well known back decades ago. 

Yes for sure.  But an amplifier with a somewhat aggressive two pole compensation, where F1 & F2 are pushed further apart, will obviously have more -tve fb up there.  Test results will, to a greater or lesser degree, differ from those of an amplifier with a simple 6dB slope miller comp.  And there are plenty of expensive amps out there still using a single dominant pole.

I like your jangling keys trick, never done that for sniffing out IMD.   

 

[5v333]

i agree that low level distortion and imd is more objectionable than regular harm dist.

but doesnt speaker drivers show signs of low level dist and imd aswell?

 

[Winston OBoogie]

I have a modest sized room (11ft X 14ft X 9ft ceiling) to put speakers in.  Before I had to recently sell them,  I had a pair of the small Harbeth P3ESR's.  They're an improved version of the tiny BBC designed LS3/5a's.   
No real bass, but it's a small room.  I like the mid range on the Harbeth's (human voice and cello are my tests) and, as something to listen to music through without putting on headphones, they worked for me.

Power amp needs were modest and 20 watts class A is what I cobbled together using some well known "blameless" pcb's as a basis.

Sure, the speakers had distortion.  But it was *MY* distortion.  From *MY* speakers and room.  So I got to learn what it was and my ears adjusted.
However, through both the speakers and my headphones (HD650's), there was a hardness at the top end that prompted me to POOGE the amp some more until I was happier.  It (the amp) tests better, and, despite the speaker/room and headphone distortions, sh*t sounded better too.

 

[Winston OBoogie]

Although a Hypex is a fine amplifier it wouldn’t be my first choice for powering a pair of A7’s.

I think this better sums up what I've struggled to say.  I think there are combinations that work well, and others that don't work quite so well. 
And I'm not talking voodoo here.  I hope 
If one always knew which speakers were going to be driven, a good designer would no doubt zero in on the bits that mattered, make different compromises, choices.  But it's not really a viable business model to dictate which and what bits the customer should have to use with your expensive amp "gizmo".

I've said enough now, I'm still learning anyway.   



 

 

[Gold]

Sure, the speakers had distortion.  But it was *MY* distortion.  From *MY* speakers and room.  So I got to learn what it was and my ears adjusted.
However, through both the speakers and my headphones (HD650's), there was a hardness at the top end that prompted me to POOGE the amp some more until I was happier.  It (the amp) tests better, and, despite the speaker/room and headphone distortions, sh*t sounded better too.

You used your ears to tweak it to sound better? Perish the thought.