Objectivist or Subjectivist? Give Me a Break

[Phelonious Ponk]

Excellent post, Alan. Marginally related, at best, to the position that there are some immeasurable or unmeasured qualities that make certain high end electronics superior to all others, in spite of the fact that these qualities either can't or haven't been measured and, therefore, can't have been consistently controlled in the manufacturing of these aforementioned superior components.

But it was an excellent post nonetheless.

Tim

 

[Ronm1]

Not just an excellent post, but IMHO, the most sensible and logical in the entire thread.

 

[rbbert]

Not just an excellent post, but IMHO, the most sensible and logical in the entire thread.

I think it's been previously stated at intervals throughout this thread; common sense for those of us that have been around.

 

[microstrip]

You quoted my question, and you still aren't going to answer it? The question was fully in sight when you said "what question".

The question was once formulated with -120 dB, I answered to it, then you repeated it changing for -140dB.

My answer is surely I do not know - I am not an expert. Van den Hul, a dutch manufactured once said in an interview that in order to see differences between cables we must go down to the -140dB level.
Anyway I would love to learn something about it - but it seems you mostly enjoy asking questions and have little answers to give, except for your love for positive and negative controls in tests and blind tests need.

 

[Alan Sircom]

Excellent post, Alan. Marginally related, at best, to the position that there are some immeasurable or unmeasured qualities that make certain high end electronics superior to all others, in spite of the fact that these qualities either can't or haven't been measured and, therefore, can't have been consistently controlled in the manufacturing of these aforementioned superior components.

But it was an excellent post nonetheless.

Tim

Yes, but that is not necessarily the case. The listening test to product design loop should be empirical. Try A. If A works, use it, if it doesn't, don't use it. Repeat with B. Nothing about that process is inherently impossible to implement, although that depends on the repeatable nature of A and B.

This is almost self selecting, because no unrepeatable element should ever be factored into product deign. And few ever are.

This is why the difference between a conrad-johnson ET3 and an ET3 SE (IIRC) comes down to changing resistors in the signal path to Vishay metal foils and the caps to CJ's own Teflons. The change is entirely capable of being production engineered on the line. Furthermore, because the resistor and capacitor values are the same as the stick ET3, the testing stage of ET3 and ET3 SE preamp on that line are identical.

Notionally at least, the existing resistors and capacitors in the ET3 were of more than sufficient quality to achieve transparency and therefore the two products should perform identically on the bench and in the system.

In this case, it's not that the immeasurable, unmeasured quantity found in listening is incapable of being incorporated, Instead, it's whether the test itself yields real-world results. And that largely depends on how you like to define the word 'test'.

 

[Alan Sircom]

Alan,
I think one aspect that compounds these discussions though is that some measurements may not correlate to what some objectivists (this is about a specific context and not objectivits-measurements in general) on forums argue as being audible or how they affect sound.
Look how contentious jitter is and then reading most forum discussions it is oversimplified with the debate more around noise rather than correlated, and then not discussing the pattern-traits of said correlated jitter and other aspects such as noise from a products mains that is seen for some gear in jitter,etc.

Regarding cables, I could point to digital cables and eye patterns, but many forum objectivitists will argue there is no correlation between average measurements compared to superb ones being audibly different, or how they affect sound.
A point I raised earlier was about energy decay/delay-latency; can any objectivist on here answer the sound difference a listener hears for a speaker driver that has a 2ms quicker decay at say 1khz on a complex musical note?

The behaviour of cognitive perception on this and some other related factors (say +1db difference between left/right speaker drivers, suckout with a narrow frequency range,etc) needs further work; case in point is headphones where you read often that they are much more accurate than speakers and yet the difference between left/right can be 4-5db, response is far from smooth, and critically measurements change substantially for certain frequency ranges (both bass and highs) with the smallest of position changes to the pinna but this is not picked up it seems in listening tests perception of sound quality; see recent test results from Sean Olive where anomalies of such variables can be seen and what Keith Howard has to say also on this subject.

TBH I am not coming out for/against any one position on here as there are valid points all round, but I do feel this is not going to resolve anytime soon.
Cheers
Orb

I entirely agree, which is why the type and nature of a given listening test remains a sticking point, and is perhaps hidebound by years of argument causing a hardening of the orthodoxies.

My take on this is a pragmatic one. There is no one-size-fits-all listening test, because there is no one-size-fits-all listener. The further you go from the home listening room and toward the design lab, the more stringent the listening tests need to be. By the time you get to the dealer, any kind of test more complicated than say a road test for a car is probably asking too much.

To each his or her own, of course.

 

[Phelonious Ponk]

Yes, but that is not necessarily the case. The listening test to product design loop should be empirical. Try A. If A works, use it, if it doesn't, don't use it. Repeat with B. Nothing about that process is inherently impossible to implement, although that depends on the repeatable nature of A and B.

This is almost self selecting, because no unrepeatable element should ever be factored into product deign. And few ever are.

This is why the difference between a conrad-johnson ET3 and an ET3 SE (IIRC) comes down to changing resistors in the signal path to Vishay metal foils and the caps to CJ's own Teflons. The change is entirely capable of being production engineered on the line. Furthermore, because the resistor and capacitor values are the same as the stick ET3, the testing stage of ET3 and ET3 SE preamp on that line are identical.

Notionally at least, the existing resistors and capacitors in the ET3 were of more than sufficient quality to achieve transparency and therefore the two products should perform identically on the bench and in the system.

In this case, it's not that the immeasurable, unmeasured quantity found in listening is incapable of being incorporated, Instead, it's whether the test itself yields real-world results. And that largely depends on how you like to define the word 'test'.

Fair enough. Can the impact of that change in resistors and capacitors be measured?

Tim

 

[microstrip]

Hello Micro

How do you think we got here?? How do you think the systems we all take for granted everyday were designed and tested??

I guess I will have to be happy with entry level and mid-fi. They seem to know what they are doing and actually seem to get measurements. You guys in the High-end can continue to stumble blind wondering how it all happened and how we got here.

Magic beans?

Rob

Rob,

You must separate what is the proprietary knowledge of the manufacturers and their specific measurements from the information that consumers and forum members can get - we only debate this minimal part.

Again , no magic beans at all. Talent and a lot of knowledge from a few manufacturers ...

 

[microstrip]

Fair enough. Can the impact of that change in resistors and capacitors be measured?

Tim

If you call measurement just getting some numbers that show some difference, yes. If you ask for a number showing a systematic correlation with sound quality, IMHO not using the usual measurements.

I have owned and several times upgraded cj equipment. The upgrades were carried replacing passive components with different type, but equal nominal value ones. The first one was a premier Seven preamplifier - Lew Johnson supplied a lot of resistors, capacitors, RCA input plugs and new wire. I did it myself - it took many hours of work.

 

[microstrip]

Micro's position is a fairly nuanced, First-World-Problem one that doesn't quite hold water under close scrutiny.

Although there are some high-end products that seem to have been thrown together with no regard for the objective performance, they are - relatively speaking - rare (interestingly, there's some fairly strong correlation between ears and meters on these products, in that they are usually the ones the 'ears-first' people tend to reject as sounding 'no good', and the 'needles-first' people can point out why. There are also products that are routinely liked by some, despite not performing well on the bench, and these are usually for reasonably predictable reasons). This gives rise to what I think of as the 'audiophile white smoke' effect; because the objective boxes are already ticked in advance, the audiophile can sit around and muse on the nature of sound like some Romantic era poet, wafting around in a haze of opium smoke.

However, the alternate and just as polarised viewpoint is the objective ivory tower of thinking measurement alone fully defines the nature of a product is as flawed in thinking. While it is possible to cite well-engineered products that were designed with minimal listener involvement, a more reliable method of product design involves a great deal of measurement and a great deal of listening along the way. In fairness, thanks in part to extremely sophisticated EE design and modelling programs making electronic circuit design extremely predictable before it even gets to physical prototype form, the closer you get to the transducers, the thermionics and the mechanical engineering aspects of a system, the more important those listening tests become.

An analogy I've seen several times recently is that of the SONAR operator. SONAR calls upon some very well tried and tested technology, with detection algorithms that are capable of seemingly impossible levels of discrimination. And yet, above all that sits an operator, listening for things that should be demonstrably and scientifically impossible to detect according to any current psychoacoustic model. It's not a perfect mechanism, as it ultimately calls upon an inherently fallible mechanism (man); but that fallible human hearing mechanism still proves to be more discriminating than the current limits of technology.

So ultimately, it's not that high-end audio - or audio in general - rejects the objective systems in place in audio. In most cases, far from it. However, there's a realisation that - as Floyd-Toole put it - it's 'science in the service of art' and that even after you have measured absolutely everything, listening (to a thing that is designed to be listened to) can provide its own data points.

Thanks for such a good post - for some reason you are a magazine editor!

I have referred several times in this forum that the capabilities of the powerful electronic simulators has opened new doors to audio design - but a much deeper knowledge of psychoachoustics is needed to deal with this information and with the enormous capabilities of the recent ultra-sophisticated measurement instruments - otherwise you will be flooded with tones of meaningless data.

Unfortunately for must of us, this also means that the manuals of the latest Audio Precision need a lot of more study and comprehension than the manual of the basic THD and IMD meters we got used to.

 

[microstrip]

(...)

This is the game. And why is it that the manufacturers of gear not put in the best technology when they are charging such huge sums for this stuff anyway....well...they got to leave room for "upgrades" and more importantly "market churn".

(...)

My preference is absolute, yours may or may not agree with mine.

I think that the evolution is much more complicated that this ... Most upgrades are due to new types of components, and yes, sometimes just to the evolution of competing equipment or partnering equipment.

And yes, your or mine single preferences are absolutes and have very little value. But the preference of many people, when statistically and properly analyzed becomes valuable objective data.

 

[Atmasphere]

Sorry Atmasphere, I didn't mean to "ambush" you. It just occurred to me that what goes on in LP mastering is many orders of magnitude more significant than the effects of cables, yet people ignore it and can talk for pages and pages about what are minuscule effects - if they exist at all in any form that any human can sense.

That's true. FWIW I don't feel ambushed at all

I seem to be in some trouble whichever word I use ("saturation, "clipping", "overload"). In your example above I would say that the needle jumping out of the groove amounts to pretty hard clipping! If I was preparing my recording for vinyl, I might be tempted to self-administer some compression, so my recording would have a form of saturation anyway, so I don't think it's possible to say simply that vinyl does not suffer from saturation - it's more complex than that..?

Yes it is! Knocking the needle out of the groove usually has nothing to do with saturation, overload or clipping. It has to do with out-of-phase information, usually bass information. The thing about bass in a natural environment is that it tends to not have much in the way of directionality. So in a natural 2-mic recording out of phase bass isn't a problem. It becomes a problem when you start recording in two different rooms where each room has its own bass instrument being recording. With any such multi-track recording you have to beware that there could be phasing problems brought on by the fact that the recording is not true stereo.

The most common means of dealing with this is a circuit that causes the two channels to be combined in mono temporarily for the brief period when you are actually dealing with the out of phase bass. Such a circuit is only effective below a certain frequency, which often can be varied somewhat. Note that we are not changing levels at all, and also that stereo information above this cutoff frequency is unaffected. This is sort of basic 101 stuff for any mastering engineer to learn to deal with. It is a matter of understanding what the playback pickup is capable of and what it is not. IOW, the limitation of the LP's dynamic range exists during playback, not record. In record mode the LP can easily demonstrate dynamic range far in excess of any digital system, but no cartridge could ever be able to play it back.

But we are not talking about saturating a cartridge either. The cartridge is capable of much larger output voltages than what might be encountered with out-of-phase bass. Its really more a matter of understanding the rules of how the recording works. You have similar limitations in digital recordings, for example no part of the recording should exceed the Nyquist frequency, or the bits on the CD surface must be of a certain dimension on order to be properly interpreted by the playback apparatus.

We did a project recently where our playback (a Technics SL1200 with Grado Gold cartridge, and our Triplanar with Transfiguration Orpheus) was able to play a cut we made just fine. But when we sent a copy to the client, there was a certain track that caused the needle to jump out of the groove. Our solution was not to change anything in the signal, but simply to cut the groove deeper. Keep in mind that the dynamic range expressed on the LP is a function of the modulation of the groove, not the depth of the groove. A deeper groove though might mean that we have to have more spacing between grooves, so we might not be able to get as much time on the LP. I am using this example to show that there are a variety of techniques available, and that altering the level does not have to be the only way to deal with it.

I have to admit I didn't know all this when we started. The machine taught us a lot! It also helps to have good advisors

 

[Alan Sircom]

Fair enough. Can the impact of that change in resistors and capacitors be measured?

Tim

Yes. But with regard to the relevance of those measurements to what we should be able to hear... we should be in gilding the lily territory.

And it's difficult to go from "should be" to "is", even if independent anecdotal reports of the results of the implementation of those changes frequently run in close parallel to one another, and without any visible means of contamination. This is because objectively it doesn't matter if it's one anecdote or 10,000 - they remain anecdotal, because they do not provide formal data points for objective evaluation. If you consider the audiophile world as one big and diffuse longitudinal cohort study, things would get a little better for using those anecdotes objectively, but no one is treating the audio world as one big and diffuse longitudinal cohort study, not least because the costs of such a study would be prohibitive.

This isn't meant at some kind of whine against a more objective stance in audio, but merely to explain why it's so difficult to go from what happens on the street and what happens in the lab. Which reminds me - I have some episodes of Breaking Bad to watch soon.

 

[Phelonious Ponk]

Yes. But with regard to the relevance of those measurements to what we should be able to hear... we should be in gilding the lily territory.

And it's difficult to go from "should be" to "is", even if independent anecdotal reports of the results of the implementation of those changes frequently run in close parallel to one another, and without any visible means of contamination. This is because objectively it doesn't matter if it's one anecdote or 10,000 - they remain anecdotal, because they do not provide formal data points for objective evaluation. If you consider the audiophile world as one big and diffuse longitudinal cohort study, things would get a little better for using those anecdotes objectively, but no one is treating the audio world as one big and diffuse longitudinal cohort study, not least because the costs of such a study would be prohibitive.

This isn't meant at some kind of whine against a more objective stance in audio, but merely to explain why it's so difficult to go from what happens on the street and what happens in the lab. Which reminds me - I have some episodes of Breaking Bad to watch soon.

Yes, if you're not up to speed you need to get there fast. The last 8 episodes are running. Game on.

Tim

 

[Alan Sircom]

The conventional tests are identical, but surely the new cap and resistor, measure thier distortions much further down than the units they replaced, and certainly can be measured, as thats how they are know to be "better". Now this replacement gives the manufacturer "facts based on measurments he took" to proclaim that the newer unit "sounds better". So, we keep the audiophiles interested in upgrading.

The thing is, if listening tests were actually done properly, with music (not tones) and no (and I mean zero) changes in listener position, etc, for someone to distinguish between these two equipments, I feel that then we would have some truth to show. I for one fully expect that a real, controlled listenging test, if in the case above just these two parts were changed out, would reveal no differences dbt, but under the sighted test, all kinds of improvements would be heard.

This is the game. And why is it that the manufacturers of gear not put in the best technology when they are charging such huge sums for this stuff anyway....well...they got to leave room for "upgrades" and more importantly "market churn".

In short, if as much effort and all the controls needed were put into listening tests (on like styles of electronics not transducers) even then, as you say, if would be a matter of preferences, as only measurments can reveal (and do) whats differernt between stuff but what sounds best is indeed preference. Also, we really dont know all we know about the ear/brain interface and although our hearing is pretty amazing, it does not exceed the ability of test equipment to resolve differences.

My preference is absolute, yours may or may not agree with mine.

I don't view the situation with such scepticism, but I can see why this can be viewed that way.

The upgrade path (I live in a country that has Naim Audio... we are well versed in upgrade paths as a result) is not designed to sell someone the same product twice. It's more like an electronic lay-away: you buy the preamp and power it from the power amp (because that's all you can afford at the time), then you buy the power supply for the preamp when funds permit. It does psychologically lock someone into buying into the brand's ethos, but in a small and desperately contended market, that doesn't seem like too bad an idea. It in theory also allows people to walk through a faster churn of trade ins and trade ups, thereby creating a pool of second-user products to be recycled through the clientele.

I have to say I believe this is fighting the next war with the tactics of the last. We are moving away from people playing a long game with periodic upgrades and instead buyers tend to buy what all in one hit. Putative future upgrades (either of the components within a device, or additional functionality to the device) doesn't seem to be transitioning well to the 2010s.

And I suspect I'm not the only one to have noticed this. Naim newer products are more geared toward meeting the needs of these one-hit clients, rather than putting them on the first rung of the ladder. It has a lot of different products doing seemingly similar tasks at similar prices, because it's trying to reach a lot of different markets.

 

[Atmasphere]

I have to say I believe this is fighting the next war with the tactics of the last. We are moving away from people playing a long game with periodic upgrades and instead buyers tend to buy what all in one hit. Putative future upgrades (either of the components within a device, or additional functionality to the device) doesn't seem to be transitioning well to the 2010s.

We've taken a different approach, which is that our equipment operates at a revision level, not unlike computer software. Right now we are at revision level 2 in the Mk3 series (IOW Mk3.2). We have a warranty reactivation program that is part of the update process to older equipment. This often means you can buy one of our older products and get it updated, get the new warranty and sonic performance for considerably less than the retail of a new unit. This has resulted in the update program being a considerable portion of our business, something I was not expecting. We instituted it as a means of side stepping 'planned obsolescence' and keeping our gear in circulation, rather than eventually headed towards the landfill. When we can show a measurable and audible improvement we will issue the new revision level, and it must affect our entire lineup. So a typical revision will go for a while (its not 'tweak of the week'). Mk3.1 for example lasted about 2 1/2 years.

 

[Phelonious Ponk]

The conventional tests are identical, but surely the new cap and resistor, measure thier distortions much further down than the units they replaced, and certainly can be measured, as thats how they are know to be "better". Now this replacement gives the manufacturer "facts based on measurments he took" to proclaim that the newer unit "sounds better". So, we keep the audiophiles interested in upgrading.

The thing is, if listening tests were actually done properly, with music (not tones) and no (and I mean zero) changes in listener position, etc, for someone to distinguish between these two equipments, I feel that then we would have some truth to show. I for one fully expect that a real, controlled listenging test, if in the case above just these two parts were changed out, would reveal no differences dbt, but under the sighted test, all kinds of improvements would be heard.

This is the game. And why is it that the manufacturers of gear not put in the best technology when they are charging such huge sums for this stuff anyway....well...they got to leave room for "upgrades" and more importantly "market churn".

Yes, sometimes touting measureable improvements that may not be audible is the game, except when touting "audible" improvements that are not measureable is the game.

Tim

 

[Phelonious Ponk]

I don't view the situation with such scepticism, but I can see why this can be viewed that way.

The upgrade path (I live in a country that has Naim Audio... we are well versed in upgrade paths as a result)

I view Naim with great skepticism.

Tim

 

[Alan Sircom]

We've taken a different approach, which is that our equipment operates at a revision level, not unlike computer software. Right now we are at revision level 2 in the Mk3 series (IOW Mk3.2). We have a warranty reactivation program that is part of the update process to older equipment. This often means you can buy one of our older products and get it updated, get the new warranty and sonic performance for considerably less than the retail of a new unit. This has resulted in the update program being a considerable portion of our business, something I was not expecting. We instituted it as a means of side stepping 'planned obsolescence' and keeping our gear in circulation, rather than eventually headed towards the landfill. When we can show a measurable and audible improvement we will issue the new revision level, and it must affect our entire lineup. So a typical revision will go for a while (its not 'tweak of the week'). Mk3.1 for example lasted about 2 1/2 years.

That seems like a good way of doing things. I think people are still willing - and if anything, actively keen - to see updates to products, in part because they are used to their software going through regular updates. And I also think people are disposed to accept those hardware updates will be chargeable.

I think it's the concept of buying piecemeal and upgrading to better products over the years is starting to run its course. There will still be a lot of people starting with the 'good', trading it in for the 'better' and on to the 'best'. But I think more and more the path is buying upfront, going through the update rather than upgrade process and then eventually moving on. Or maybe up!

 

[andromedaaudio]

all audiophiles are skeptic ,tim

I view Naim with great skepticism.

Tim