How far have ss amps really come in the last twenty years?

@jdl57  If your Sound Labs have the original toroid backpanel, there's a tweak you can do that makes the speakers a lot easier (and better sounding) to drive! Alternatively you could get the new "Toroidal 2" backpanel that pretty much does the same thing.  If you really want to hear what ESLs do, traditionally since the Quads first appeared, ESLs and OTLs have been the way to do it. If your Sound Labs are so old that they have the original EI transformers for the high frequencies then they are already easy to drive, unless the bias controls are damaged which is very common with the older Sound Labs.

Very interesting post. Why is it that you think that the DBX folks won't even consider the type of amp your are suggesting in a DBX trial? 

@daveyf  because the difference is instantly audible and they inherently don't like amps that aren't perfect voltage sources. This could be a made up story on my part, but so far that's been my experience. No matter how hard we humans try to be objective, in the end we fall well short. 

While I agree many amplifiers sound different, they all should sound the same if done right.

This statement is problematic- correctly stated this needs to be added: '...if the intention of the designers were the same and with that in mind also correctly engineered'.  Seehttp://www.atma-sphere.com/Resources/Paradigms_in_Amplifier_Design.phpfor more information.

I too would be interested to know which amps are typically excluded and why. I agree that seems to be a gap in the discussion, and also an interesting point in general about design (would you call it optimizing knowingly inaccurate artifacts for real world listening?). I’d still be curious where the threshold of audibility is for some of these characteristics, and whether listeners can distinguish consistently between characteristics that are harder to measure. These seem like important design questions.

We have the ability to measure what is important IME. What I find problematic is a refusal on the part of the industry in general and 'objectivists' in particular is to do the homework- which is to correlate what we hear with what we measure (for example, a weighting system is needed for the harmonic spectrum of distortion, placing less weight on the 2nd-4th and more weight on the higher orders, since the ear uses the latter to sense sound pressure). But if we did that, then amplifier spec sheets would tell you how they sounded, and the industry doesn't want that!

So, the question that keeps coming up in my demented mind is, if the goal of high end audio is to reproduce the sound of live performances, why isn’t the high end defined by the likes of Crown, JBL, ElectroVoice and other pro audio items of reproduction? After all, unless you are sitting in the first 10 rows of a concert venue, what you are hearing is coming from the pro audio world. I can buy a lot of Crown amplification (substitute any pro audio product here) for the cost of A Gryphon or Dartzeel amp. Just wondering.

The reason for that is simple- sound reinforcement is meant to be loud enough and reliable. The last nth degree of resolution is unimportant. I have done a few stage shows were we did install a high end system though (which was a royal pain); people kept asking us when we were going to turn it on. I'd then turn the volume down and they would sort of stand there with a surprised expression and say 'Oh.'.

In a nutshell pro audio gear for sound reinforcement is usually lower performance than good high end or even mid fi home gear.

Feel free to augment or quibble, but I would call an "objectively audible difference" between amplifiers as follows: when the equipment is gain-matched to be sure the difference is not simply due to different overall volume levels, listeners can still repeatably distinguish between different pieces of equipment *only by listening*.

OK- 'objectively audible difference' sounds different (if you will pardon the expression) from 'objectively audible'; thanks for clearing that up.


The thing that bothers me about this sort of approach is that usually when I see or hear someone talking in this manner, they are usually only allowing for certain circuits (amplifiers) to be in the conversation, while out of hand rejecting others on an arbitrary basis.

The problem is that in high end audio, there is a community of designers that take different approaches to solving amplifier issues. Some of them are quite brilliant and their works stands easily on their own merits, such as David Berning or Nelson Pass and there are many more. Further, a good number of designers including those I just mentioned, are not completely convinced that they should be using amps with ultra low output impedances (which means that their amps may not act as perfect voltage sources either). In my case, I've avoided making our amps work as voltage sources with intention- as I see the approach as highly problematic. Chief among these is the well-known fact that in removing distortion, the operation of negative feedback adds some of its own, which is always audible as increased brightness and harshness. I'm not saying its bad but I am saying that most of the time its inappropriately applied or inexpertly applied, with a mediocre result, which I see as about 95% of all traditional amplifiers made.

Now the approach I take is nothing new- its what was around prior to Mac and EV developing the voltage rules that are used today. Most people don't realize that there was something before that! I'm by no means the only one doing it either- and in all cases the idea is simple: the ear will place a priority on tonality created by distortion over a tonality created by FR error. To this end, a system that has less audible distortion can sound more neutral than one that has flatter FR. You have to be pickier about the amp/speaker interface (which is what the voltage rules were designed to prevent) but it can pay off in spades, but DBX people don't like to even consider such an amplifier in a DBX trial.

I have admitted I sometimes hear differences between amps, but, after reviewing evidence (incomplete?) I am also open to the possibility those were explainable by power/obvious artifacts or simply not objectively audible

What is 'objectively audible'??

@mrdecibel  Its on this thread too:

No, in my limited experience, I often don't hear the difference between amplifiers (with the exception of underpowered or tube-bloated amps)

The comment is ridiculous, but one that is common in this debate, as if, somehow, hearing how tubes do things is an exception to this DBX nonsense.  The thing is, its become so ubiquitous that we don't even see it sometimes. The thing is, tubes get docked for having a higher output impedance and thus causing a frequency response error (which is often true, but also often very slight) while at the same time no-one says anything about the colorations caused by distortion. Brightness and harshness, ubiquitous to solid state, is indeed a coloration. But we just accept that this is so, and somehow its the norm. But if you listen to real music, that brightness and harshness at volume isn't there and a good tube based system doesn't have it either.


What I am saying in all this is if we collectively did our engineering homework, we could have solid state that doesn't sound harsh in any way. But, we don't, so the beat goes on.

I may be missing something (point it out to me), but there’s almost no convincing evidence out there to support repeatably audible differences between high quality cables or amps

The differences have to do with distortion. This is easy to measure, and when measurement correlates with subjective experience then we can call it real. Its well-known that IMD is highly audible. It should then come as no surprise that when adding capacitance to a power supply reduces IMD (due to less modulation of the power supply- this is a simple timing constant in most cases so this is easy to quantify) that you can also hear it.

The higher ordered harmonics have long been known to contribute to brightness and harshness. We’ve known this since the 1930s (see Radiotron Designer’s Handbook volume 3, IIRC page 67). Now days we also know more about how the masking principle of the human ear works. This applies directly to how distortion is audible, since the ear converts all forms of distortion into some sort of tonality. In tube amps, a 2nd or 3rd harmonic dominates (depending on if the amp is single-ended or fully differential). The ear treats both the same. Now tube amps make more distortion than transistor amps, but those lower ordered harmonics are crucial as they allow the ear to mask the presence of the higher ordered harmonics. This is why tubes sound smoother on top; its nothing to do with bandwidth- we can’t hear the higher orders! The funny thing is, its been shown that injection of a 2nd or 3rd order also allows the ear to perceive more detail and greater depth- thus accounting for why tube amps tend to be smoother, more detailed and with greater soundstage. The peculiarity is that this actually makes tubes more accurate! I’m very convinced that we need to do more research to find out why this is so, but if you are pragmatic, you don’t need to know why, just do engineering based on the implications:


Now its accepted that we can’t build amps that make no distortion whatsoever and that is what is needed to get around this issue. So instead, it appears that if a circuit is devised that injects a 2nd or 3rd harmonic **of appropriate level** that we can build a more neutral sounding amp. Of course, this sort of thing is readily audible even in a DBX test; but IME people doing such tests will avoid allowing good examples of such to taint their ’findings’. You can’t do it if the test is rigged! You have to start wtih two amps that are based on different design goals rather than two that are built the same way. What I mean by this is, you can use engineering to design an amplifier with low impedance and low distortion, or you can use engineering to build an amplifier that sound neutral according to human hearing perceptual rules. One is the Emperor’s New Clothes, as its designed to essentially look good on paper, and the other is real as its meant to work with rather than against the way we perceive sound. Naturally the latter will measure ’poorly’ compared to the former, but the latter is far more likely to sound real.


If you compare two such amps as in my example above, even in a DBX you can perceive the difference. I’ve seen it/heard it.


One more thing as a sort of BTW... none of this implies tubes or solid state. Its all engineering. To me it just makes more sense to design a circuit based on human hearing perceptual rules than making a nice piece of paper! This is the very reason why tubes are still around, decades on after being ’declared obsolete’. The audio industry in general really hasn’t tackled this issue at all, preferring to keep their collective heads in the sand rather than deal with physiological knowledge gained in the last half century. Pathetic, when you think about it. We really shouldn’t be having a tubes/transistor debate this far on.

And if you disagree, and can prove an audible difference under his conditions (gain matching), he's offering a pair of his top-of-the-line speakers (see the first few posts).  

That sounds nice. Might run short on speakers after a while...

Like shooting fish in a barrel. Do you have to fly to England to do the shootout?

I am sorry, but neither Pass nor Curl can compete at the top level. Perhaps they are doing something wrong.

:)  Nelson's SIT amplifiers are some of the best sounding out there. But they are only available in kit form IIRC. SITs went out of production a long time ago.

Regarding John Curl, isn't he the one behind the Blowtorch preamp? That is one of the very few solid state preamps that has developed a legendary status.

Distortion is an interesting term. Distortion compared to what? You mention that the human ear treats 2nd and 3rd harmonic distortion favorably. Therefore, one needs to question whether these distortions are distortions at all. Since we, as humans are using our ear/brain to determine what is considered as a true  and pure signal, wouldn’t it make more sense to label a signal that doesn’t include these 2nd and 3rd harmonics as the signal that is actually distorted...?

Even though they are beneficial in many ways they are still distortions. This is why there is that term 'euphonic'. I've often felt though that distortion harmonics should be weighted according to how important and audible they are to the ear.

If you want to hear an amplifier sound "different" get a tube amp with an output impedance of several ohms! Ohm's Law rules here! A frequency response that will follow the varying impedance of the typical speaker.

There is far more to it than that!

Distortion plays a major role in the sound of any amplifier due to how the ear/brain system treats distortion, which is to say it treats it as a tonality. We've known this fact since the 1930s (see Radioton Designer's Handbook, 3rd edition).


Most solid state amps are designed to have minimal distortion and is why so many of them sound the same, which is to say, harsh and bright especially at higher volume levels. Now tube amps actually make more distortion, but because they have lower ordered harmonics in greater percentage quantities than solid state, the ear's masking principle allows the lower orders to mask the presence of the higher orders. This is why they sound smoother. The odd bit is that the presence of the 2nd or 3rd harmonic (both of which are treated the same by the human ear) causes a greater sense of detail and soundstage definition! This is why tubes are generally more detailed and have more depth.

Now if we could eliminate distortion entirely that would be great. But we can't, so it appears that an injection of a 2nd or 3rd harmonic is beneficial to solid state designs. This is recognized by several prominent solid state designers in high end (John Curl and Nelson Pass come to mind) and it should be no surprise that they have turned out some of the more highly lauded solid state amps as a result.


It is this bit of understanding about how the human ear works that is one thing that's changed in the last 20 years and it is perhaps the most significant- parts and topology notwithstanding. Its our ability to take advantage of how the ear works through engineering that yields the most progress. We're only just now starting to crack that nut.