Objectivist or Subjectivist? Give Me a Break

[jkeny]

Some people will claim that tonal balance will be shown in a null test & I'm not disputing that - I would just love to see a control produced that shows it & proves it. This would be the logical conclusion! Controlled null tests that prove a null test is capable of showing these differences would be interesting. The evidence needs to be presented that the test is appropriate, not just a claim that it must be because "who can hear down that low, anyway".

If people want to be an objectivist, be objective with your experiments & present evidence with some rigour! Don't just latch onto a measurement because it is convenient & looks like it could support your argument!

I fully agree. It's the ones that do that raise the ire of the subjectivist crowd (and many of the sincerely objective too). You know, the ones that make it look like they've correlated everything. To me they are no different from the guys that walk into a Casino saying "I have a SYSTEM!". Good luck I say!

BTW I really loved the video posted of Mr.Carver discussing that challenge. What surprised me was when he said distortion had nothing to do with his success and said that in fact he didn't even use the distortion pots. He said it was tonal balance that he tweaked.

 

[jkeny]

What might be of relevance to this discussion is Amir's latest article in Wide Screen Review "Dynamic Range: How Quiet Is Quiet?". I'm sure it might allow us to home in on at least one aspect of what is required in a Null test? Hopefully it will be republished here in due course or some summary of the findings given?

 

[Phelonious Ponk]

A guy loves an audio technology that is consistently characterized by a bunch (the technical term) of harmonic content at the output that is not present at the input. Can I reach the logical conclusion that this harmonic content is what he likes? No. It could be that he likes it in spite of the extra harmonics. It could happen. Is it a reasonable assumption that he likes all that extra harmonic content that is consistently present in the amplifier type he loves? It is to me. But I'm not an objectivist, just a skeptic. I lack the technical expertise to be a true objectivist.

John, it seems you nearly have this argument to yourself now. Good luck.

Tim

 

[opus111]

Is it a reasonable assumption that he likes all that extra harmonic content that is consistently present in the amplifier type he loves? It is to me. But I'm not an objectivist, just a skeptic. I lack the technical expertise to be a true objectivist.

You lack the understanding of philosophy of science to be a true skeptic. Your position is that of a pseudo skeptic, for whom correlation becomes sufficient to conclude causation.

 

[JackD201]

Some people will claim that tonal balance will be shown in a null test & I'm not disputing that - I would just love to see a control produced that shows it & proves it. This would be the logical conclusion! Controlled null tests that prove a null test is capable of showing these differences would be interesting. The evidence needs to be presented that the test is appropriate, not just a claim that it must be because "who can hear down that low, anyway".

If people want to be an objectivist, be objective with your experiments & present evidence with some rigour! Don't just latch onto a measurement because it is convenient & looks like it could support your argument!

I think Carver might have inadvertently sunk the distortion theory of tubes vs solid state with his challenge as he tells it.

 

[jkeny]

It's interesting when the term "consistently characterised" is used for an amplifier & then a THD measurement used as the pretence that it characterises this amplifier! Show all the other missing measurements then you have some chance of characterising the amplifier & then make "reasonable assumptions".
Otherwise, assumptions based on such flimsy evidence are far from reasonable or logical.

 

[Whatmore]

It seems like an explanation of the fundamentals of logic is required here.

If a partial measurement CORRELATES with an audible characteristic then we can say that measurement can be used to determine how two devices will sound with respect to that characteristic. So if A measures better than B in that particular measurement A will sound better than B in that particular characteristic. A might measure worse than B in another audibly CORRELATED characteristic & so it will sound worse than B in that particular characteristic.

Now, it is foolish, illogical & down right wrong to now pick a measurement which DOES NOT correlate to audibility, THD for instance & say that A's THD measures lower than B's THD, but people like the sound of B better than A therefore those people like the sound of higher THD - a complete fallacy & utterly ridiculous. Let's measure the weight of A & find that it is heavier than B but people like B better therefore people like lighter devices

So parial measurements don't have any real value unless they correlate to audible characteristics

If you want to use measurements to draw conclusions then you either have to have measurements that correlate to an audio characteristic (& then you can say something about that characteristic) or at least try to have as full a set of measurements that have some chance of characterising the device.

Latching onto a simple measurement & drawing simple & illogical conclusions is what I object to with so-called objectivists. In fact objectivist is a misnomer because there's nothing objective about doing that!

In order to make your statement about THD, you first need to show that a higher THD is not the cause of people liking B better than A. You can't assume this any more than you can assume that the weight of an audio component has no effect on how people hear it

Wrong! You are making a logical fallacy! Do I have to also show that the colour, weight, size, shape, (any other difference you care to mention) are also not the cause for people liking B over A. You make the claim, you prove the connection/correlation!

It's the basics of logic!

So what you are saying is two devices are different in so many ways therefore I will pick this one random difference & claim that it is the reason that the two sound different? Now it's up to me to prove that you are wrong?

You've lost me John.
Can you perhaps explain to me what you mean by correlation?

EDIT: to clarify my question: you say we should pick something like THD. You then go on to say that if people like the sound of the amp with a higher THD then there is no correlation... that's the bit I don't get. If higher THD goes together with people's preferences then that is what correlation is...

 

[jkeny]

You've lost me John.
Can you perhaps explain to me what you mean by correlation?

Correlation = how strongly one variable is related to another variable.

If a measurement strongly correlates to an audible characteristic, then a change in that measurement means a change in that audible characteristic. i.e the measurement is worse by X% & therefore the sound characteristic is worse by about X%. Similarly, if there is a change in that sound characteristic then there will be a concomitant change in the measurements by roughly the same amount.

This is just one measurement correlating to one audible characteristic. For an amplifier there will be many different measurements that need to be done to characterise that amplifier. To then correlate these measurements to audible characteristics is a big task.

But because it is a big task is no excuse for being lazy & jumping to assumptions with a some measurements which have no correlation to audible characteristics. Either produce all the measurements which characterise the device, or show measurements which you can prove correlate with audibility or the final choice is - don't make unwarranted assumptions & then try to defend them or try to shift the proof to those who point out the logical flaw!

The attempts to pass off pseudo-science as real science by those professing to have science on their side is the thing that bugs me! It would be just entertaining if they stuck to their own little world of pseudo-science discussing it endlessly but it's when it is used as a cudgel to suppress observations, that's it gets outrageous. It's almost hilarious when they then insist these observations have to be proved using their pseudo-science frame of reference.

 

[Phelonious Ponk]

You lack the understanding of philosophy of science to be a true skeptic. Your position is that of a pseudo skeptic, for whom correlation becomes sufficient to conclude causation.

Let me repeat for you what I just said, just above the part you quoted, so you'll understand that you lack the focus required to hold up a true conversation:

A guy loves an audio technology that is consistently characterized by a bunch (the technical term) of harmonic content at the output that is not present at the input. Can I reach the logical conclusion that this harmonic content is what he likes? No.

Tim

 

[Whatmore]

Correlation = how strongly one variable is related to another variable.

If a measurement strongly correlates to an audible characteristic, then a change in that measurement means a change in that audible characteristic. i.e the measurement is worse by X% & therefore the sound characteristic is worse by about X%. Similarly, if there is a change in that sound characteristic then there will be a concomitant change in the measurements by roughly the same amount.

This is just one measurement correlating to one audible characteristic. For an amplifier there will be many different measurements that need to be done to characterise that amplifier. To then correlate these measurements to audible characteristics is a big task.

But because it is a big task is no excuse for being lazy & jumping to assumptions with a some measurements which have no correlation to audible characteristics. Either produce all the measurements which characterise the device, or show measurements which you can prove correlate with audibility or the final choice is - don't make unwarranted assumptions & then try to defend them or try to shift the proof to those who point out the logical flaw!

The attempts to pass off pseudo-science as real science by those professing to have science on their side is the thing that bugs me! It would be just entertaining if they stuck to their own little world of pseudo-science discussing it endlessly but it's when it is used as a cudgel to suppress observations, that's it gets outrageous. It's almost hilarious when they then insist these observations have to be proved using their pseudo-science frame of reference.

So if higher THD is associated with preference for that amp (as per your example), how can you say that the two are not correlated?

 

[opus111]

A guy loves an audio technology that is consistently characterized by a bunch (the technical term) of harmonic content at the output that is not present at the input. Can I reach the logical conclusion that this harmonic content is what he likes? No.

The difference there is in the word 'logical'. I wasn't referring to logic, rather practice. You did indeed reach the practical conclusion even though you admitted it wasn't due to logic:

Is it a reasonable assumption that he likes all that extra harmonic content that is consistently present in the amplifier type he loves? It is to me.

Reason <> logic. For authority, ask Mr Spock.

 

[jkeny]

So if higher THD is associated with preference for that amp (as per your example), how can you say that the two are not correlated?

There is no correlation proven between THD & audibility. Why not say that it is heavier & therefore sounds better or it is red & therefore sounds better? Because correlation to become causation needs a body of evidence backing it up. You have no such body of evidence with regard to THD or weight or colour or any other random difference you wish to cite!

 

[Phelonious Ponk]

The difference there is in the word 'logical'. I wasn't referring to logic, rather practice. You did indeed reach the practical conclusion even though you admitted it wasn't due to logic:

Is it a reasonable assumption that he likes all that extra harmonic content that is consistently present in the amplifier type he loves? It is to me.

Reason <> logic. For authority, ask Mr Spock.

Absolutely inaccurate,Spock. I reached no conclusions, and it is inthat word - conclusions - that the difference liies. I made what i think is a pretty safe assumption, IMO, for my own purposes. YMMV. You may assume that the legendary harmonic content of SET amps has nothing to do with their popularity if you like. But conclusions? I not only came to none, I was very careful to say so.

Tim

 

[Whatmore]

There is no correlation proven between THD & audibility. Why not say that it is heavier & therefore sounds better or it is red & therefore sounds better? Because correlation to become causation needs a body of evidence backing it up. You have no such body of evidence with regard to THD or weight or colour or any other random difference you wish to cite!

Where did I say anything about causation?
YOU said that there was a preference for the amp with the higher distortion, not me. All I am saying is that if that is the case then by definition, there is a correlation between THD and preference.

 

[opus111]

Absolutely inaccurate,Spock.

Fascinating.

I reached no conclusions, and it is inthat word - conclusions - that the difference liies. I made what i think is a pretty safe assumption, IMO, for my own purposes.

But that 'safe assumption' is a conclusion in the scientific sense. All conclusions in science are provisional, pending new data. Assumptions (seldom are they pretty safe) are what we bring to the data, conclusions are what we take away. Indeed you did draw that conclusion for your own purposes, not at all similar to how a true skeptic would infer which would be according to impartial (i.e. disinterested) reasoning. No 'IMO' would be invoked as science doesn't deal in opinions.

YMMV. You may assume that the legendary harmonic content of SET amps has nothing to do with their popularity if you like.

Again you misunderstand the true skeptical position. It would be pseudo skeptical to 'assume' (conclude in my parlance, I don't mind adapting to your idiolect) no connection.

 

[jkeny]

Where did I say anything about causation?
YOU said that there was a preference for the amp with the higher distortion, not me. All I am saying is that if that is the case then by definition, there is a correlation between THD and preference.

Ah, tautology reigns supreme!
The statement made by others is that those who prefer the sound of tube/SET amps do so because they must like distortion!


If you are just looking for an argument - try elsewhere!

 

[Phelonious Ponk]

Fascinating.



But that 'safe assumption' is a conclusion in the scientific sense. All conclusions in science are provisional, pending new data. Assumptions (seldom are they pretty safe) are what we bring to the data, conclusions are what we take away. Indeed you did draw that conclusion for your own purposes, not at all similar to how a true skeptic would infer which would be according to impartial (i.e. disinterested) reasoning. No 'IMO' would be invoked as science doesn't deal in opinions.



Again you misunderstand the true skeptical position. It would be pseudo skeptical to 'assume' (conclude in my parlance, I don't mind adapting to your idiolect) no connection.

Not so fascinating, actually. I'm using logically, and for that matter, skeptical, in the common, conversational sense. I am neither applying, or expecting scientific rigor. I concede. The connection I'm making between the harmonic content of SET amps and their popularity is casual and my position is unsupported. The connection thousands of SET fans make between what they hear in the midrange and the harmonic content of their amps is unsupported as well. Of course that is where all of that extra harmonic content lies and it would be....no, stop me before I push us into the zone where semantics clashes with the demands of the scientific method. I know nothing. Neither do those SET fans. We don't even know that there is any additional midrange content. It's all anecdotal information awaiting verification.

All yours...

Tim

 

[KeithR]

My position is that one can draw conclusions from partial measurements. John's example is a good one: Many SET amps are very high in harmonic distortion, to the point that the distortion is a fundamental characteristic of their harmonic content. I think it is pretty reasonable assumption that the people who love the sound of those amps like that distortion. I suppose they could just be good at ignoring it, but that's a bit of a stretch. Another example is the Harman speaker studies that have been discussed so much here. They only measured FR, in-room, of speakers (though at many points on and off axis), but they successfully predicted preference in later listening tests. Partial measurements, good conclusions -- most people prefer more linear response. Complete characterization of those speakers and that room from a single measurement? Of course not. I don't think anyone would take that position seriously.


Tim

All amps distort though. even small odd -order harmonic distortions are much more audible than even ones, correct?

so therefore you still have a preference, not a set of facts.

 

[KeithR]

btw, let's use Lamm ML2.1 for your measurement talk about SETs. As they measure well according to JA.

http://www.stereophile.com/content/lamm-ml21-monoblock-power-amplifier-measurements

and assume speakers would be 8ohm, 95+db speakers so work with the specified wattage.

 

[Phelonious Ponk]

All amps distort though. even small odd -order harmonic distortions are much more audible than even ones, correct?

so therefore you still have a preference, not a set of facts.

I don't know that odd-order harmonic distortions are more audible, but they are more annoying when they are audible. Fortunately, in competently designed equipment they are not. And yes, that's my preference.

Tim