[Slaughter]

So we know the cable debate will go on forever. Sorry if this has been posted already, but has anyone tested an amp's output to see if there is a measurable difference between different tubes or opamps? I hear a difference between an OPA2132 and OPA2107 and a GE 5963 and a Sylvania 5963. I am guessing they wouldn't measure any different at the output, but I could be wrong. But if they don't measure any different, doesn't that mean there are things that test equipment can't account for? Just a thought. If I had the equipment and know how, I would gladly do the tests myself.

 

[jcx]

certainly the Scientific/Rationalist Physicalism/Materialism world view enabling Engineering and production of audio electronic toys strongly suggests that if different amplifiers produce the "same" (within limits of measurement noise) signals at the headphone driver terminals then the headphone produces the "same" (again within noise, environmental, parameter drift of the transducers) sound
 
if under those conditions a listener claims to perceive differences we usually insist on controlled experiments with precise level matching and blinding protocols to reduce the influence of internal psychological "state" from the perceptual evaluation - the overwhelming majority of such controlled tests reveal that most "Audiophile" claims cannot be substantiated in absence of measurable differences:
 
http://www.diyaudio.com/forums/solid-state/12752-blind-listening-tests-amplifiers-28.html#post152392
 
 
for a automated null test tool try a "prosumer" grade soundcard w good ADC, DAC and http://www.libinst.com/Audio%20DiffMaker.htm
 
Engineering principles can be used to predict some dimensions of "difference" from data sheet specs of differing op amps or tubes - in the case of op amps there are many that should give indistinguishable performance by known psychoacoustic thresholds - I heavily discount the op amp rolling subjective difference reports - most are given in the relative frequency response terminology that strongly suggests lack of level matching - "identical" systems heard at slightly different volumes sound different due to Fletcher-Munson loudness curves - way below the differences that are clearly perceived as loudness changes

 

[Shike]

Here's the thing with tube rolling:
 
Tubes can already have wear on them.  Unless they measure identical you're adding a variable which may be exasperated in a circuit.  I'm not mostly a tube person, but it would require measuring the output.  If they're within certain reasonable limits then most psychoacoustic models say they should sound the same.  If they don't a DBT or similar needs to be performed.
 
 
A similar issue is possible with opamps.  For example, some love opamp rolling.  However, few rarely test before and after.  The 2134 can oscillate rather easily in a cmoy circuit for example where the 2132 has less of a chance.  There's people that claim they noticed huge improvements in an amp but wonder why it's dead in a week (due to oscillations).  IMO, if an amp is designed around a specific chip and is measured/tested for that chip - leave it there.  Why tweakers think they know more than the designers is beyond me, but if you trusted the designer making the amp enough to purchase it shouldn't you half-way trust their selection of parts?
 
Benchmark DACs use the notorious NE5532 for example.  Their opinion is that, quite frankly, others have been using it wrong and wonder why it sounds bad.  I agree with this, if it measures bad either the chips bad or the circuits bad.  When Benchmark can make it measure well then the chip isn't the problem - it's the circuit they're sticking it in.  Even better, they found opamp rolling HURT and DEGRADED performance in the DAC1.  With which opamp by chance?  The OPA2134, one of many tweakers go to chips.

 

[Uncle Erik]

Good point, Shike.  It's also important to keep in mind that tubes are constantly changing.  They're not the same value some hours in.  That's why I don't bother with tuberolling and obsessing over NOS tubes.  Whatever "sound" you achieve will probably be slightly different in a few months.  So why make a big fuss over it?  I just buy reasonably priced tubes that are reliable.  Then I enjoy the music and stop worrying about it.

 

[Slaughter]

I just find it interesting that I don't see the skeptics jump all over tubes and opamps the way they do for cables. It seems like testing would be just as easy and could be another good debate. And if successful, would give either side of the debate more ammo. The post jcx referenced didn't really address the issue. It only showed that Bob Carver can make one amp sound like another.

 

[leeperry]

Most likely because it's a pointless debate, ppl come up w/ their real world experience(or lack thereof) and are only interested in proving their point, linking whatever url as a hard proof.
 
The DAC-1 from Benchmark is only using 5532 and 4562's, they claim it to the best opamps for the job...OTOH many ppl claim that this DAC sounds shrill and boring, exactly what I would expect from 4562's as DAC LPF + 5532's as headphones amps.
 
http://forum.rightmark.org/topic.cgi?id=4:504-3

Everybody -in different locations, at different times, without knowing from each other- told the same story, that they found the differences between opamps more important than the differences in dac chips

 
Some ppl seem to measure gear and never listen to it? Opamps are what colors the sound the most, and being higly complicated integrated components, they all sound different. Many audio gear manufacturers like to put nail polish on top of their opamps, because all they are doing is cashing in on the opamps design engineers work IMO.
 
Look at what happens in an AD797(whose B grade sounds lovely):
 

 

[jcx]

the Carver Stereophile Challenge is a strong indication that the same V on the Loudspeaker terminals gives the same sound - to "Golden Ears" - even when one source is a massive "High End" audiophile state of the art tube amp vs a tweaked mid priced SS amp
 
if frequency response matching/nulling per the Carver Challenge is the 1st order audible difference then simple calculation from op amp datasheet open loop gain curves suggests that few op amps recommended for audio applications have so little audio frequency gain that the resulting closed loop gain in a typical headphone or line level audio amp will differ by a fraction of the established DBT ABX level matching threshold:
 
http://home.provide.net/~djcarlst/abx_crit.htm
 
 
there are reasons for the many op amps manufactured - input noise, bias current, output current, voltage swing - and engineering reasons to choose one over another in differing application circuits - its likely that some op amp rolling suggestions can make other properties of the circuit worse (or just maybe sometimes better) - but frequency response in audio line level applications (usually) isn't one
 
 
DAC I/V is a interesting application in which some op amp performance parameters could have theoretical/measurable effects on circuit operation - high GBW product, low output impedance and exceptionally linear front end are desirable for best theoretical performance - but again the expected differences do not appear in audio frequency level changes 
 
in many applications further "theoretical" improvement can be had with composite/multiloop op amp circuits - much more so than just rolling in the single op amp circuit

 

[Shike]

 
Quote:

Most likely because it's a pointless debate, ppl come up w/ their real world experience(or lack thereof) and are only interested in proving their point, linking whatever url as a hard proof.

 
It's better than someone showing up with absolutely no proof besides anecdotes.
 
Quote:

The DAC-1 from Benchmark is only using 5532 and 4562's, they claim it to the best opamps for the job...OTOH many ppl claim that this DAC sounds shrill and boring, exactly what I would expect from 4562's as DAC LPF + 5532's as headphones amps.

And magically anything with silver is seen as "bright", and anything with a tube is seen as "warm".  CD Treatments add resolution, and putting a maple would block under a laptop will not only improve the sound but color.
 
Expectation bias: it's nasty when it pops up.
 
Quote:

http://forum.rightmark.org/topic.cgi?id=4:504-3

Everybody -in different locations, at different times, without knowing from each other- told the same story, that they found the differences between opamps more important than the differences in dac chips

 
Cool, someone with five posts and little to nothing on testing methodology.  At best we see him talk about demoing some circuits, but it doesn't say whether participants were aware of the opamps or not.  Furthermore, it doesn't have any actual measurements - the opamp could have been oscillating all over the place in both their circuits where the other was stable.  It also doesn't cover whether they were level matched either, go figure.
 
Quote:

Some ppl seem to measure gear and never listen to it? Opamps are what colors the sound the most, and being higly complicated integrated components, they all sound different.

 
Based on various DBTs - no, no they don't.  Only golden ears that never do a level matched DBT seem to say that.
 
Quote:

Many audio gear manufacturers like to put nail polish on top of their opamps, because all they are doing is cashing in on the opamps design engineers work IMO.

 
It's a conspiracy, it's not like they're trying to protect their intellectual property or anything.  Oh wait, that's exactly what they're doing.  It doesn't matter that it uses ICs, they may just want to prevent people from copying a circuit and outfitting it with parts that they actually invested R&D in designing around.
 
The best comparison is programming.  There's tons of libraries and stuff you can already import in Java for example, some extremely complex.  The final product though takes lots of work and implementation.  Even if it uses two libraries the work is still protected, so they may want to make it closed source and even slap DRM on it.  In this case, you can think of opamps as paid for libraries - while they play a key aspect in development they alone are not indicative of the final product.

 

[p a t r i c k]

Quote:

So we know the cable debate will go on forever. Sorry if this has been posted already, but has anyone tested an amp's output to see if there is a measurable difference between different tubes or opamps? I hear a difference between an OPA2132 and OPA2107 and a GE 5963 and a Sylvania 5963. I am guessing they wouldn't measure any different at the output, but I could be wrong. But if they don't measure any different, doesn't that mean there are things that test equipment can't account for? Just a thought. If I had the equipment and know how, I would gladly do the tests myself.

I think this is a good point.
 
My own contribution in most of these discussions is to raise the issue of auto-suggestion. It is most likely auto-suggestion which people are reporting when they claim differences in the sound of cables. I'm not saying that there is no difference between cables, but rather that the auto-suggestion eclipses any differences that there might be and can create perceived differences when there is none.
 
However auto-suggestion comes with all hearing activity. Another very powerful drive for auto-suggestion is that something will sound better because it is scientifically demonstrate-able that it is a better design.
 
Simply "knowing" that you are listening to something that is a better design will be enough to drive a great deal of auto-suggestion.
 
Even though something really is a better design doesn't mean that its improvements will be audible.
 
So, the auto-suggestion will often eclipse any sound differences that may or may not be picked up by the ears.
 
People, such as myself, who tend like like scientifically good design are of course given to auto-suggestion just like anyone. We are presented with, perhaps, a higher resolution digital format and so we immediately start "hearing" additional details. However it might be that the higher resolution format simply doesn't make any real world difference. Our perceived improvements are auto-suggestion.
 
The voodoo cable manufacturers know very well that "scientific credibility" matters. Although their fans often eschew scientific methods, at the same time the voodoo cable manufacturers will fill the descriptions of their products with pseudo science. They use pseudo science to add credibility and with that credibility will come auto-suggestion.
 
So, scientific credibility matters a lot. However what is left out is quantification.
 
Earlier I mentioned how a higher resolution digital format may be perceived as better simply as a result of auto-suggestion. Even though it actually is better, measures better, the perception of improvement comes not from its inherent "betterness" rather simply from the auto-suggestion the listener adopts knowing that it is better.
 
The problem here is quantification. Yes, it is better, but no quantification is given as to how much better and no suggestion as to the likelihood of that improvement being audible.
 
The best way to combat auto-suggestion is through well controlled blind ABX tests. This is as true of identifying difference or qualitative benefits between audio equipment that is differentiated by sound scientific design differences as it is between equipment that is different by other types of design difference.
 

 

[InnerSpace]

Quote:

... doesn't that mean there are things that test equipment can't account for?

I love threads like these.  I believe that everything - theoretically - can be "measured", i.e. "described mathematically."  But the point is - it isn't.  Not can't be, but simply isn't.  Suppose you're sick, and you go to the doctor, and you say, "Doc, check me out."  He fusses around for a minute and says, "You're five feet eleven inches tall."  You say, "Well, yeah, but what's wrong with me?"  He says, "You weigh 170 lbs."  You say, "I don't feel good."  He says, "You're slender, so it's possible people might think you're actually six feet."  You say, "Doc, you're not helping."
 
That's where we are with audio measurement right now.  Electrically based audio is about 80 or 90 years old, and hasn't - frankly - attracted the very best scientific brains over the years, and what brains were attracted attacked the low-hanging fruit pretty early, and haven't done much since.
 
By contrast, primate > hominid > human evolution is about seven million years old, and hearing has been developing all the way, pushed by some very high stakes.  An oft-quoted speculative example is being an early hominid and hearing a twig snapping behind you.  Evolutionary biologists explain how we recognize the transient and localize its position, and react - is it a predator? It doesn't take much imagination - we can all still do this - to assume the sound contains other qualities, that told us whether the potential predator was creeping along, or rushing, was light, was heavy ... a whole wealth of implied information.
 
Can you conceive of a modern-day audio test that could define and interpret such a split-second transient sound?
 
So yes, I think there are plenty of things that test equipment can't account for.  Almost everything, in fact.  Not because it's in principle impossible, but simply because no one has tried yet. We haven't even scratched the surface, and I'm not sure if we ever will.  It would be a massive project, and I can't see a real-world motivation for it.

 

[Uncle Erik]

We aren't testing the properties of humans, we're testing the qualities of a physical object.

The entirety of the signal is electrical and can be measured very precisely. And it's a fairly simple signal.

If the signal is the same no matter the cable, then there isn't a difference. Also known is the human threshhold for hearing. Even if there is something that cannot be measured, it would fall below that threshhold.

Interestingly, a client at the office is a cable manufacturer. Not audio stuff, but high power transmission lines and industrial applications. I've had an opportunity to look through some of their work. They very much do use measurements and science to develop cables. All the stuff tha skeptics call for here. Their work is legitimate and they seem to make an excellent product. I can't say more because it's privileged, but it is a good company.

Which leads to an interesting question. If cables can be measured and scientifically developed for certain applications, why do those same methods fail miserably for audio cables?

If test gear is so unreliable, then how does this company stay in business? Moreover, how come their products work as advertised? If there were a bunch of wildcard unknowns (as some would have you believe), then they should get all sorts of wacky results from their tests and probably have a bunch of failures and problems with real-world testing. But they don't. The products do what they're supposed to do and are fully explainable with known science and test gear.

Lacking motivation to explore cables?

Oh, please.

If you haven't noticed, audio cables are likely an eight figure business each year. Do you mean to say that developing a product that will put millions and millions of dollars in your pocket means that there is no motivation?

That's ridiculous. I'd *love* to eat crow on this one. If I could, I'd develop a proven cable. Then I'd sell the rights, buy a nice place on the hill in Avalon over on Catalina, and sip mai tais on a balcony overlooking the Pacific for the next 40 years. I'd give anything to be wrong about cables if I could do that.

In case you're wondering why no one else has apparently tried, it's because people tend not to crow over failures. I'm sure that plenty of people have tried to develop cables over the years. Money is a great motivator, after all.

But all of them failed.

They all failed because there's not a lick of evidence anywhere. If someone had found evidence, you better believe they would have gone after the market with a clawhammer. If I had proof, I'd unload on the competition in a nasty way. I'd demolish all the others and rake in the dollars.

But since no one has done that, you can assume that many people tried and failed. Though some of them went into the cable business anyway. They're making bogus products and they know it.

 

[InnerSpace]

^^ Erik, you misunderstood.  I don't dispute that the cable business is bogus.  I imagine today's cable hucksters would actively fight against better scientific understanding, just like old-time medicine men did.  What I said was - our current palette of audio tests don't penetrate very far into the totality of how and what humans hear.  If you think current audio science tells us all we need to know, you're fooling yourself.

 

[p a t r i c k]

Quote:

By contrast, primate > hominid > human evolution is about seven million years old, and hearing has been developing all the way, pushed by some very high stakes.  An oft-quoted speculative example is being an early hominid and hearing a twig snapping behind you.  Evolutionary biologists explain how we recognize the transient and localize its position, and react - is it a predator? It doesn't take much imagination - we can all still do this - to assume the sound contains other qualities, that told us whether the potential predator was creeping along, or rushing, was light, was heavy ... a whole wealth of implied information.

This ability you mention is extremely important to consider.
 
The mystery and magic of human hearing takes place with what we actually do with the signal after it arrives at the ears.
 
The human being is astonishingly selective when "making sense" of sound. The little twig snapping, the hearing can suddenly hone in on that at great speed to work out if it is possibly a sign of danger.
 
Also, this ability happens unconsciously, we don't have to go through life consciously remembering to "listen out" for the sound of little twigs snapping.
 
However the measurements for audio equipment are not about what humans make of sound, they are simply comparing input signal with output signal and ascertaining if there is any difference. So that is a totally different domain. I don't know how good audio equipment measurement is, but it is not trying to measure human perception, rather it is measuring what is "served up" to the human in the first place.

 

[InnerSpace]

Quote:

We aren't testing the properties of humans ...

 

Furthermore ... we are testing humans - or we should be.  We need to know what it is about real sounds that isn't captured by audio reproduction.  Both ends of the deal - the sound and the brain and their interaction - need to be understood.  There has been some desultory work there, but not much.
 
And of course there's no motivation to really understand cables.  Real understanding would reveal that they don't matter very much.  Real understanding would destroy the lucrative market.  Obviously cable makers aren't going to risk that, and there isn't any real motivation for anyone else to do it.
 
And that's the problem generally ... a thorough understanding of reproduced audio isn't going to put anyone on a hill in Catalina, sipping mai tais.  There just isn't the demand.  Quality audio is too niche.  So we struggle along with second-rate minds doing third-rate work ... plus a bunch of fools and hucksters mixed in ... and it is what it is.

 

[InnerSpace]

Quote:

However the measurements for audio equipment are not about what humans make of sound, they are simply comparing input signal with output signal and ascertaining if there is any difference. 

Exactly.  Thanks.  All we ever hear is yes, the output is the same as the input, or no, it's not.  We never hear about the factors that distinguish "live" from "not live" being absent from the input in the first place.  That's because no one knows what those factors are, and no one cares enough to find out - or, rather, pace Erik and his hypothetical mai tais, there's not enough financial upside to spur anyone to make the effort.  No one is going to make a billion dollars by designing a better stereo.  Sadly!