[robertbudding]

Niouke you're making a bold claim, so I will examine the words used very closely. I understand that normal communication doesn't allow endless caveats; however some distinctions need to be made in my opinion.



As written, absolutely not. Simply an absurd statement. I realise this is not what you (probably) meant to propose, so let's unpack the statement.

A home listening audio amplifier made in the 1960's is most certainly going to sound different than one made in the 1970's or 2010's. Not to mention from the 1930's, say the Loftin-White or similar.

Let's take an open loop tube amp with a basic topology from the retro days, but with a lot of modern stuff added in. Let's say LED cathode bias and stuff like that. Nothing drasticly altering the topology. It is made in the same decade as the 700 eur decent quality receiver bought at a big box store. These amps are in effect contemporaries, however they will most certainly sound different.

Your statement, if interpreted in good faith, is talking about gNFB amps, more specifically amps utilizing a lot of gNFB. By any knowledge of the basics of electronics, gNFB is going to 'flatten' almost every difference. (Also it is going to flatten the sound, but that is my subjective opinion.)

So; amps that use a lot of gNFB are going to sound identical.

This is a truism, a tautology.

However what about the amps that do no utilize a lot of gNFB, or that are open loop? Are they not in the picture at all, or are they considered as being of a quality not worth considering?

I'm an open loop guy who dislikes distortion, as I explained in my previous message here. I don't enjoy "tube sound" in my HIFI yet I detest gNFB sound.

Is my amp included in that blanket statement?

(For the purposes of this discussion "my amp" can be considered an abstract thought experiment since this thread is not about schems but about concepts, but for those more interested in practicalities look at Sonic's thread: https://www.head-fi.org/threads/modern-balanced-tube-amp-build.852879/ )




Yes, SS amps which by necessity utilize huge amounts of gNFB. gNFB will sound like gNFB.




I do somewhat agree, but not to the level proposed in the quote. There is a lot to be said about focusing on the bottleneck. However from working with guitar tube amps a lot, I understand very well what "distortion on distortion" is and sounds like. There is a difference in final output depending on how gradually you introduce distortion, and what kind of distortion to introduce at what point. The audio chain does in my opinion operate somewhat as a whole. That said, if you have crappy speakers, it's not going to matter a lot if you shave off some distortion from the amp.



'Scientifically' is one of those potentially weaselwords that instantly bring up my skepticism senses.

Humans bred animals and crops millenia before anybody found out about DNA and genes even existing. And nowadays that we know about DNA, somebody can say "show me the exact gene that produces X in this species", and after you inevitably cannot (at least yet), following the same logic of SCIENTIFIC PROOF the sophist can claim "trait X is not heritable in that species" even though it clearly is.

The fact that tests have not proved something proves that tests have not proven something. This is not a claim of fact ("it is not so"), but rather an indicator among others. This is not to say science is all bunk, but rather to understand it's limitations.

To bring it closer to topic at hand: which listening tests are scientific and which are not? I have a lot of logged listening test hours, by my ears. Are they scientific?

Wavebourn used his cats (and house guests) as scientific test subjects. If the cat or the houseguests arriving at the house in another room thought the sound was a real instrument being played, his amp was realistic. Is this scientific? If not, why not?




This is the "distortion on distortion" thing again. In my technical and listening opinion it is not at all insignificant to add distortion to an already distorted sound.

Think about an old recording that has let's say 5% THD, mostly second and third. Made in the late 1930's for example.

So you might think this recording is already so distorted in comparison to the acoustic performance it was captured from, that it doesn't matter whether the amp replaying it is producing 1% THD or 0.01% THD.

Completely wrong in my opinion.

Basic level understanding of wave physics (sound is just a waveform, whether electric or air pressure) tells us that when two waveforms are added, the resulting intermodulation products are of higher order than either of the original waves contained. The bigger the 'new' wave (added by the amp), the bigger the 7H 8H 9H components that are produced by the intermodulation. This is the very worst kind of distortion there is.

This means that if you play your bad recording thru a bad amp, the results is significantly worse than if you play it thru a good amp. This same effect in my opinion makes all the 'distortion shaping' and such nonsense a very bad idea.

So no matter what happens in the studio, the way you play it back matters a lot.

I personally found that the most retro stuff I listen to, jazz and blues for 1930 to 1950, became the most alive after listened thru an open loop low THD amp. SS amps don't have this effect and after thinking about the intermodulation effect and knowing the THD spectra of gNFB amps this is not a mystery.

You seem to have ignored the long list of qualifier, such as low distortion, no clipping, and flat frequency response. Don't believe it? Then set up ABX tests with volume levels matched to 0.1 dB.

 

[MrCurwen]

I don't completely disagree that open loop can sound different from gNFB, but which one represents the input signal with the best accuracy when working into a real world complex load?

Categorically? I don't know. But I asked are they not worth considering? Or, are only "the best" wrt THD and identical to "the best" qualified to be considered?

On a practical level, yes, there are some open loop schems that do in practice reach the perceived cleanness of super low THD SS amps (meaning they have none of the "tube sound"). This means they are at least in the same class with the gNFB SS amps in the THD category. Which one would come on top in detailed measurements, I do not know.

Two problems here:
1. An instrument amp definitely becomes part of the total instrument sound, and is often operated well into an area of clearly audible distortion. There is no parallel in hifi audio for this kind of operation or system.

True when considering magnitude of the effect (audible distortion, however I protest because clipping and audible are two very different things; very small distortions in HIFI amps are in fact often audible), absolutely false when considering the method.

Let's say you have a two stage tube amp. Both of these stages produce some 2H, and less 3H. No further measurable harmonics. Let's say they both have "a green zone" on their load lines where no harmonics are created (this is a fantasy example to clarify a larger principle), then a yellow zone where harmonics are created.

Having the same output level, would you rather have;

1) Both tubes operate just a little bit on the yellow zone, both producing only a little bit harmonics.

2) The first one on the green zone, and the next one deep into yellow zone.

I would have the second one, every time. Even if the resulting THD is similar or very small in both cases, the second one will sound significantly better.

This effect doesn't have to result in significant increase in THD to be noticeable (at least by my ears). Therefore this "distortion on distortion" is (in my opinion) very much a factor in HIFI audio.

2. Yes, the entire chain operates as a whole, but all speakers are crappy relative to an amp of any hifi design in every single aspect, FR, distortion, intermod, phase response, everything but noise.

Yes. The effect described above still stands in my opinion. Speakers are the crappiest part of the chain (or headphones), but there are significant sound quality gains to be had by improving the chain as a whole. In my opinion.

How did the cat input his observation data?

If I recall correctly by reacting to some nature noises, while not reacting when the sound quality was lacking. I could be mistaken this was a long time ago.

It depends entirely on what makes up the 1%, but your 5% figure for the 1930s 78 record is low.

Yes I know.

And the 1% THD amp would be questionable in today's world, would definitely require much more information. Remember, a single figure THD spec is useless, it's a 3D data block with magnitude, frequency and resulting products as axis.

Yes I know. These numbers were randomly picked for example use. I should've left them plain numbers with no %.

Yes, but for intermodulation to be produced between two waveforms there must be a nonlinear element. Amps with better linearity produce less IMD. Generally, IMD and THD track each other, though. While IMD is much more obnoxious, there will also be measurably high THD in that system too.

Amps without nonlinear elements do not exist.

Let's say we would even have a SS gNFB amp that had it's H6 and up clearly under the level of sensory observation. Non-observable by ear. Now introduce IM, and you can bring these just above the level of observation (sound becomes "just a little bit harsh" or "uptight").

There is no situation where this effect doesn't happen to at least a tiny amount. Whether or not this is observable by measurement or by ear is up to each specific instance. I submit that it is probably noticeable to my ears, since I enjoy the output more if I reduce THD anywhere along the reproduction chain, provided the harmonic structure remains slanted. I.e. stay open loop for the most part.

Generally agreed, but that's a very binary way to look at it.

Let me amend my statement: This same effect in my opinion makes all the 'distortion shaping' and such nonsense a very bad idea unless distortion shaping is achieved in a way that does not increase THD. As long as THD goes down or stays the same, shape away. I would prefer it to lower.

 

[MrCurwen]

You seem to have ignored the long list of qualifier, such as low distortion, no clipping, and flat frequency response. Don't believe it? Then set up ABX tests with volume levels matched to 0.1 dB.

Which part in my post specifically are you referring to?

 

[bigshot]

to me all well conceived amps sound the same and are audibly transparent

That's a bingo!

There are amps that are designed to color the sound, but that is intentional. The goal of an amp designer should be to create a clean and balanced sound. I've found that in my Oppo HA-1 and in my $70 Radio Shack amp. I've never bought an amp since the 70s that wasn't clean and balanced. If they were colored, I would have returned them as defective.

Yes, in my experience these amps sound boring and unrealistic.

Balanced and clean shouldn't sound unrealistic. If it does, you probably have to correct your response. Do you use EQ? I don't want an amp that is anything other than clean and balanced. If amps and players all sounded different, I would have to recalibrate my whole system every time I used a different source or amp. The place to make corrections and apply DSPs to spice up the sound is at the last stage. Amps used to have jumpers in the back that you could patch an equalizer into. They should do that again. Then we could get a "sweetening box" to plug in that would apply adjustable levels of euphonic distortion, equalization, timing correction and reverb. Then everyone would have *exactly* the sort of sweetening they wanted and coloration wouldn't have to be hard wired into an amp or player.

DSPs offer the opportunity to introduce whatever sort of sound shaping you want and it is totally adjustable. Right now AV amps are leading the way with this (as they are with multichannel and most other big innovations in sound). The tendency in amps intended for music is to double down on more and more purity. But amp manufacturers have already solved that problem. Amps are audibly transparent... as they should be. The next step is plug ins to modify and sculpt the sound. That is DSPs. I use a half dozen DSPs in my system and there's a thread here on DSPs that duplicate the harmonic effect that tubes have on sound. You don't need a tube amp to have tube amp sound.

 

[castleofargh]

@coinmaster . you can't go making many statements about many things and talk like you're an expert, then say that you have better things to do when asked to elaborate on your own claims. aside from how hypocritical it makes you look after about a dozen posts in this thread in just 2 days, it goes against the very nature of that sub section of the forum. ideally people come in here because they are interested in the evidence and demonstrations leading to a conclusion. they have questions/ideas and wish to see some experimentation or some of the available data on a subject to disprove or suggest a conclusion. else the "science" in "sound science" is lost.

evidences aren't systematically required to come with any statement we make, if only because we usually don't need to demonstrate a consensus, or because we can simply have an opinion on the matter. but at the very least if you're willing to engage in a technical conversation, don't be a snob when someone asks you to elaborate.



now back on topic, almost, amps sounding the same or not. IMO amps should sound the same within the given range of gains and loads they were designed for. that still leaves aside:
-all the amps failing to reach audible transparency even under nominal use.
-all the amps working fine under other circumstances but used in a way where they aren't audibly transparent.
-some more or less significant changes going on with the damping ratio, so the result very much depends on the amp and headphone used.
-and of course all the amps where the designer was looking for his own sound. not only they exist, but in the fancy elite audiophile world, euphony is likely to be assimilated as technical superiority by the consumers. so it can be a better bet than trying to achieve high fidelity like everybody else to end up with the same sound.
and I'm most likely missing a few other cases.
so I disagree with the idea that all amps sound the same, but not because I believe only special amps can sound transparent. I actually think it's fairly easy to get a reasonably transparent amp. it might not have 200W of output power, it might not have a noise floor at -130dB, but it might sound pretty clean anyway for some more or less specific use.


about the price, if an amp does more or objectively better than another, I'm ok with it being more expensive. if a non transparent euphony consistently felt better to me than transparency, I might let myself pay more for subjective preference, but so far I never met such amp. I've heard many that were really nice sounding, but I wouldn't wish to have that all the time. as a DSP I could turn OFF, that would be great. not as my main amp. but that's personal taste. I also wouldn't personally seek the objectively very best amp because I don't really care about anything beyond audible, but would find reasonable for it to have a cost in relation to being the king of the hill in measurements.
what is absolutely ludicrous to me is when we end up seeing some gear sold for ten times the price of typical amps, while still failing to actually measure better or be more versatile than typical amps. with great prices should come the duty to show superiority somewhere. it's at least my view on elite devices. the best is king and does whatever he wants because there is nowhere else to go if we want the same performance. but bad gears trying to pass as kings dressed in a king's price, IMO those deserve to be publicly humiliated for the scam they are.

 

[robertbudding]

Which part in my post specifically are you referring to?

The post you replied to listed the conditions under which Amps sound identical. But you seem intent on insisting that they don’r sound identical. That is true only if they are clipping, distorting, or have a non-flat frequency response.

 

[MrCurwen]

robertbudding, are you referring to this article he linked in his post? https://numeralnine.wordpress.com/2013/10/09/a-brief-guide-to-audio-for-the-skeptical-consumer/

What conditions are mentioned there, I quickly read it through and nothing much sticked to my eye. Please elaborate.

But you seem intent on insisting that they don’r sound identical. That is true only if they are clipping, distorting, or have a non-flat frequency response.

Based on my understanding at this moment of what you are saying (I don't know the exact conditions you are referring to, except the ones you list here), let me clarify.

Amps that utilize a lot of gNFB that are not clipping, have comparably low THD, and have a flat frequency response will sound different than open loop amps that are not clipping, have comparably low THD and a flat frequency response.

 

[upstateguy]

Indeed. But from a consumer perspective. You could have two amps with "0.001%" THD that sound completely different.

Personally I've never found a solid state amp that excited me, and I've heard a lot of very expensive ones.
Tube amps are the way to go for me. There's a magic in the sound that typical solid state designs just don't have.
Of course design matters most and unfortunately, making an amp produce practically no distortion is comparatively easy to making an amp produce good distortion, since there are infinite ways to do it. Tubes make this job much easier though.

The more things change the more they stay the same. Bob Carver is your Huckleberry.

 

[bigshot]

I have a Carver subwoofer. It is unbelievable. I had a bigwig from one of the biggest audio companies come by to visit and hear my system and he was amazed by my sub in particular.

 

[upstateguy]

That's a bingo!

There are amps that are designed to color the sound, but that is intentional. The goal of an amp designer should be to create a clean and balanced sound. I've found that in my Oppo HA-1 and in my $70 Radio Shack amp. I've never bought an amp since the 70s that wasn't clean and balanced. If they were colored, I would have returned them as defective.



Balanced and clean shouldn't sound unrealistic. If it does, you probably have to correct your response. Do you use EQ? I don't want an amp that is anything other than clean and balanced. If amps and players all sounded different, I would have to recalibrate my whole system every time I used a different source or amp. The place to make corrections and apply DSPs to spice up the sound is at the last stage. Amps used to have jumpers in the back that you could patch an equalizer into. They should do that again. Then we could get a "sweetening box" to plug in that would apply adjustable levels of euphonic distortion, equalization, timing correction and reverb. Then everyone would have *exactly* the sort of sweetening they wanted and coloration wouldn't have to be hard wired into an amp or player.

DSPs offer the opportunity to introduce whatever sort of sound shaping you want and it is totally adjustable. Right now AV amps are leading the way with this (as they are with multichannel and most other big innovations in sound). The tendency in amps intended for music is to double down on more and more purity. But amp manufacturers have already solved that problem. Amps are audibly transparent... as they should be. The next step is plug ins to modify and sculpt the sound. That is DSPs. I use a half dozen DSPs in my system and there's a thread here on DSPs that duplicate the harmonic effect that tubes have on sound.

You don't need a tube amp to have tube amp sound.

Right-o

Edit to say: Expand

 

[upstateguy]

...amps where the designer was looking for his own sound. not only they exist, but in the fancy elite audiophile world, euphony is likely to be assimilated as technical superiority by the consumers. so it can be a better bet than trying to achieve high fidelity like everybody else to end up with the same sound.

Yep.

 

[pinnahertz]

Categorically? I don't know. But I asked are they not worth considering? Or, are only "the best" wrt THD and identical to "the best" qualified to be considered?


On a practical level, yes, there are some open loop schems that do in practice reach the perceived cleanness of super low THD SS amps (meaning they have none of the "tube sound"). This means they are at least in the same class with the gNFB SS amps in the THD category. Which one would come on top in detailed measurements, I do not know.

What I was getting at is the typical open loop amp, especially using tubes, cannot drive a complex load with an accurate replica of its input signal. The primary reason is output impedance (but not the only reason). It's much higher than the gNFB amps, and there easily can be a resulting audible difference. It may be a preference or not, but that's subjective opinion, and may depend at least partially on what source impedance the speaker designer intended for his product to be driven from. In other words, audibly different may or may not mean better. An open loop tube topology does lack the ability to replicate its input signal into a complex load, though.

True when considering magnitude of the effect (audible distortion, however I protest because clipping and audible are two very different things; very small distortions in HIFI amps are in fact often audible), absolutely false when considering the method.

Let's say you have a two stage tube amp. Both of these stages produce some 2H, and less 3H. No further measurable harmonics. Let's say they both have "a green zone" on their load lines where no harmonics are created (this is a fantasy example to clarify a larger principle), then a yellow zone where harmonics are created.

Having the same output level, would you rather have;

1) Both tubes operate just a little bit on the yellow zone, both producing only a little bit harmonics.

2) The first one on the green zone, and the next one deep into yellow zone.

I would have the second one, every time. Even if the resulting THD is similar or very small in both cases, the second one will sound significantly better.

This effect doesn't have to result in significant increase in THD to be noticeable (at least by my ears). Therefore this "distortion on distortion" is (in my opinion) very much a factor in HIFI audio.

If the above is about instrument amps, we can stop right here. Those are a unique application that bears no relationship to a high fidelity amplifier.

The relationship of THD levels and audibility is complex and quite situational. Research has shown many cases where the inverse of your statement is true, but without a complete definition of the test conditions, discussion is pointless. However, what you statement, specifically relating to an unstated amount of THD ("very small" is not an amount) must be incorrect as it stands, because it lacks the necessary detail of the conditions and full amplifier characteristics. It's an iceberg, and you've only to the tiny tip.

Yes. The effect described above still stands in my opinion. Speakers are the crappiest part of the chain (or headphones), but there are significant sound quality gains to be had by improving the chain as a whole. In my opinion.

Not completely unfounded, but you say "significant quality gains", that's an unscaled, unquantified statement. What is significant? In terms of audibility, I would generally disagree, but since this seems to be a binary discussion, and you'll just pick out the exception, perhaps we should just table this. There are as many variables that result in audible difference as those that don't.

Yes I know. These numbers were randomly picked for example use. I should've left them plain numbers with no %.

That would be even less meaningful, but as I've said, distortion audibility cannot be defined by any single numbers in % or dB. Again, you're responding in a binary manner to the world of analog signals that are completely time and frequency variant. You just can't make that statement correct without defining some very specific conditions.

Amps without nonlinear elements do not exist.

This is what I mean by a "binary discussion". Your statement is correct, but irrelevant. Yes, everything is a bit nonlinear, but the degree of nonlinearity makes all the difference in the world. Can we work in the real world please?

Let's say we would even have a SS gNFB amp that had it's H6 and up clearly under the level of sensory observation. Non-observable by ear. Now introduce IM, and you can bring these just above the level of observation (sound becomes "just a little bit harsh" or "uptight").

You cannot increase IMD without increasing THD, they are generated by the same nonlinearity, though not necessarily directly coupled to each other. Your example is not in the real world at all, and again, without specifics of both the amount of distortion, the rate at which the amp enters nonlinearity, the resulting spectrum, and the specific test signal, the discussion is meaningless. You are both right and wrong.

There is no situation where this effect doesn't happen to at least a tiny amount. Whether or not this is observable by measurement or by ear is up to each specific instance. I submit that it is probably noticeable to my ears, since I enjoy the output more if I reduce THD anywhere along the reproduction chain, provided the harmonic structure remains slanted. I.e. stay open loop for the most part.

I believe the second sentence above defines your position quite well: That of an "enlightened audiophile". Just a guess, but I'd bet you have zero test data to support your super-human hearing abilities.

Let me amend my statement: This same effect in my opinion makes all the 'distortion shaping' and such nonsense a very bad idea unless distortion shaping is achieved in a way that does not increase THD. As long as THD goes down or stays the same, shape away. I would prefer it to lower.

The audibility of distortion is multi-dimensional data set around which exists a set of conditions of observation. Simply looking at THD is invalid and meaningless. Simply looking at THD levels is equally invalid and meaningless. Thinking you can isolate IMD from THD is naive, and not possible in the real world.

Just as an example or case in point, there is an entire professional product line that gained popularity using a process that increased distortion in a way that had some sonic desirability when used sparingly, and on specific types of signals. Pioneering ABX testing showed that nonlinear distortion generators could not always be detected reliably at levels we would consider unacceptable (see "High Resolution Subjective Testing Using a Double-Blind Comparator", D. Clark, JAES 1982 Vol. 30 #5).

However, assuming that the audible differences between amps of different topology are somehow related to THD, IMD, or some form of unquantifiable magic will be routinely shot down without full test details. Any fully sighted observations can and should be just simply ignored.

 

[MrCurwen]

What I was getting at is the typical open loop amp, especially using tubes, cannot drive a complex load with an accurate replica of its input signal.

Binary thinking. In reality there is a whole spectrum of amp schemes ranging from retro to quite modern that utilize tubes. Mostly in the DIY side of things, but if we are talking reality, DIY is a big portion of the "high end" game. People aren't building copies or similar designs of the big box 700 eur amps, they are copying high end stuff at home. Especially because the price of parts is low compared to buying the amp.

If we are to step away from the binary thinking, there is no reason to make these kinds of statements. They are not accurate in reality; they apply only to the retro type amps.

The primary reason is output impedance (but not the only reason).

We can play word games around what is and what isn't a tube amplifier. However there is no reason that an amplifier that uses tubes to have a bad output impedance.

An open loop tube topology does lack the ability to replicate its input signal into a complex load, though.

Categorically false. Only true in certain cases (retro amps).

If the above is about instrument amps, we can stop right here.

No. Look at the exchange and my wording again.

However, what you statement, specifically relating to an unstated amount of THD ("very small" is not an amount) must be incorrect as it stands, because it lacks the necessary detail of the conditions and full amplifier characteristics. It's an iceberg, and you've only to the tiny tip.

Principles and methods of action can be described without using specific real world events. It is 100% valid to explain some case of mechanics and state "this example disregards resting friction". This is done for the sake of making a point. Not every explanation is a detailed real world case study. You do actually understand this. I guess you are either not reading my words in good faith or you are trying to send me a message that I am way off topic and should stop writing. I personally do think this stuff is relevant to stuff in this thread but it might not be.

Not completely unfounded, but you say "significant quality gains", that's an unscaled, unquantified statement. What is significant?

This was defined by the part you left out in your quote; by my opinion. If this is insufficient, you should disregard it.

That would be even less meaningful, but as I've said, distortion audibility cannot be defined by any single numbers in % or dB. Again, you're responding in a binary manner to the world of analog signals that are completely time and frequency variant. You just can't make that statement correct without defining some very specific conditions.

You are fully aware that was an example to demonstrate a principle of action. It is 100% ok to explain a case of simple mechanics by using Newtonian physics, even though they are not in fact taking into account Einstein's stuff, and later stuff. It is "good enough" to demonstrate something the demonstrator wishes to demonstrate.

Again, you know this, so either you are writing in bad faith or I am way off topic.

This is what I mean by a "binary discussion". Your statement is correct, but irrelevant. Yes, everything is a bit nonlinear, but the degree of nonlinearity makes all the difference in the world. Can we work in the real world please?

What is the level of accurateness applicable to this discussion?

On one hand my clinging to strict accuracy is "binary" and on the other hand you make broad generalizations without taking a lot of stuff into account. Look above your sweeping inaccurate generalizations regarding open loop amps.

Is it ok to make some assumptions for the sake of communication, or is it not? One man's accuracy is other man's irrelevant nit picking and vice versa.

I believe the second sentence above defines your position quite well: That of an "enlightened audiophile". Just a guess, but I'd bet you have zero test data to support your super-human hearing abilities.

This line would give some indication to me that you are in fact not writing in good faith. I hope I am interpreting wrong, I'm not a native english speaker.


I don't mean to define the parameters of this discussion. If my writing style is not ok here then I should stop writing. I personally think it doesn't take anything away from scientific discussion (whatever that exactly would mean in a forum like this) to provide personal views and abstract examples.

 

[pinnahertz]

Binary thinking. In reality there is a whole spectrum of amp schemes ranging from retro to quite modern that utilize tubes. Mostly in the DIY side of things, but if we are talking reality, DIY is a big portion of the "high end" game. People aren't building copies or similar designs of the big box 700 eur amps, they are copying high end stuff at home. Especially because the price of parts is low compared to buying the amp.

You have a distorted reality of the actual DIY world. It's very, very tiny. Vocal, and with strong Internet forum presence, but tiny with respect to commercial audio manufacturing companies.

If we are to step away from the binary thinking, there is no reason to make these kinds of statements. They are not accurate in reality; they apply only to the retro type amps.

We can play word games around what is and what isn't a tube amplifier. However there is no reason that an amplifier that uses tubes to have a bad output impedance.

To be clear, I never said tube amps have "bad" output impedance, now did I? But, show me one single tube-based open-loop amp with an output impedance down around 0.1 ohm that is capable of driving a complex load without that load modifying it's otherwise flat frequency and phase response. Just one! Output impedance isn't a word game, it's an electrical property. And please don't add to all the other inaccuracies by misquoting me.

Categorically false. Only true in certain cases (retro amps).

Give me one example of a modern tube-based open loop amp where this isn't true then.

Principles and methods of action can be described without using specific real world events. It is 100% valid to explain some case of mechanics and state "this example disregards resting friction". This is done for the sake of making a point. Not every explanation is a detailed real world case study. You do actually understand this. I guess you are either not reading my words in good faith or you are trying to send me a message that I am way off topic and should stop writing. I personally do think this stuff is relevant to stuff in this thread but it might not be.

If an example is not relevant to conditions in the real world then why use it? It proves nothing.

This was defined by the part you left out in your quote; by my opinion. If this is insufficient, you should disregard it.

If you think you've provided a clear definition of what you think is "significant", please quote it directly. What you have provided is completely unscaled and vague.

You are fully aware that was an example to demonstrate a principle of action. It is 100% ok to explain a case of simple mechanics by using Newtonian physics, even though they are not in fact taking into account Einstein's stuff, and later stuff. It is "good enough" to demonstrate something the demonstrator wishes to demonstrate.

Again, you know this, so either you are writing in bad faith or I am way off topic.

But we are not demonstrating a principle of action in the real world by using an impossible to produce set of conditions or devices. No, your example was not realistic at all. What if I began discussing an amplifier that acted like a straight wire with gain, and a zero output impedance, frequency response to 100mHz, phase linear? Would that mean anything to anyone trying to relate to practical realities? Not in the least.

I could say this: if I had a perfect speaker with zero distortion, perfectly flat response, and perfect impulse response, then differences in amps would be more audible. But that's silly, there's no such speaker, so why use the example?

What is the level of accurateness applicable to this discussion?

Well, in "Sound Science", I'd have to say as much as possible. You've avoided any and all specifics, and thus there's no accuracy in your statements. There's no Sound Science being done.

On one hand my clinging to strict accuracy is "binary" and on the other hand you make broad generalizations without taking a lot of stuff into account. Look above your sweeping inaccurate generalizations regarding open loop amps.

They aren't broad generalizations, though. You don't seem to realize some of the important details behind what I'm saying. Eliminating global feedback creates more than a few issues, and when dealing with tubes, it's slightly worse in some ways, better in others. Negative feedback results in higher bandwidth, lower distortion, lower output Z, lower gain, to name a few. Take that away and the inverse is true. However, you should also be aware that there's pretty much negative feedback in every amp anyway, like it or not. A triode has inherent neg feedback, for example. And even amps with zero gNFB have local NFB, so it's not ever gone. But the issue with tubes is high characteristic plate impedance and high output voltage, which to drive a speaker load, has to be converted to low(er) impedance, lower output swing with a transformer. All of that imposes distortion and bandwidth limitations.

So, again, show me a low output Z, zero gNFB amp that can drive a complex load with a precise replica of its input.

edit: Even when discussing gNFB, it's not binary! It's not gNFB or not, you can design for low gNFB, high gNFB or anything. There is wisdom in designing the most linear stage possible before polishing it off with a bit of gNFB. Where does all of that fit in your binary argument?

Is it ok to make some assumptions for the sake of communication, or is it not? One man's accuracy is other man's irrelevant nit picking and vice versa.

Not when the irrational nature of them key to supporting your point.

This line would give some indication to me that you are in fact not writing in good faith. I hope I am interpreting wrong, I'm not a native english speaker.

I'm responding to your more than classic position in the argument, "it's clearly audible to me, you can't test it, gNFB is bad...blah blah blah...with the more than apparent lack of factual support.

I'm not sure what you mean by writing in good faith, I'm writing about what I believe is true, but I'm always willing to learn. If you have something scientific to convey, though, you'll need to come up with actual details and real-world examples, even test data. And yes, listening tests do produce data, if properly done.

I don't mean to define the parameters of this discussion. If my writing style is not ok here then I should stop writing. I personally think it doesn't take anything away from scientific discussion (whatever that exactly would mean in a forum like this) to provide personal views and abstract examples.

When you make a claim that isn't supported by science, or even specific real world examples, much less actual test data, it's going to be rough getting people to accept them in a Sound Science forum. On the other hand, if you had real world data to convey, I'm more than happy to look it over.

 

[Niouke]

thanks everyone, I will now make the distinction between open loop and gNFB.

I've read the discussion from the consumer's standpoint, trying to find hints on how to shop for amplifiers. From what I've understood I'd need to compare amps in tandem with the load I intend to drive, and with a full laboratory to measure the distortion in relation to the frequency, and that's not very helping. No wonder people spend their days arguing about the amp/speakers tandem with no proof whatsoever of anyone's claim, as measures don't exist.

Now if I'm wrong, please guide me in comparing two amplifier of similar power, the Yamaha A-S701 (which I own) and a McIntosh MA5300. I there any reason I should be drooling for the McIntosh's amp stage? The published specs look very close, while the McIntosh costs 12 times more...