[bigshot]

Generally, I agree that blind comparative testing is valuable. But I think that such testing also has limitations in the sense of potentially not accurately revealing differences which are more subtle, may be perceived more subconsciously than consciously, and/or may have a more significant effect when listening durations are longer and more like real listening.

Measurements can definitely reveal things that are more subtle, but they can also reveal things that are totally inaudible. Unless you know where the line of threshold of perception lies, a measurement is just an abstract number with no practical context. We listen to music with our ears. It's just as important to measure our hearing as it is to measure the output of our stereo system. Ultimately, our ears are the last processing of the signal in the chain. There is no point pushing sound we can't hear through the chain. If we can't hear it, it isn't going to improve perceived sound quality. Our ears are what they are. We have to produce recorded sound to suit them.

Humans do not hear more discerningly over long periods of time. In fact, the longer you listen, the more your ears adjust to accommodate the sound. Auditory memory is notoriously short. For similar sounds, the length of time is just a couple of seconds. That's why controlled listening tests have direct A/B switching so the samples can be compared right next to each other.

We do not perceive subconsciously. Hearing is a mechanical process. You either hear something or you don't. How you *interpret* the sound might have some sort of subjective spin, but being able to hear a difference in a controlled comparison test is pretty much cut and dried. The purpose of blind testing is to remove subjective or subconscious bias. You can't apply bias if you can't tell which is which. And you can't apply bias if you can't hear it in the first place.

Subjective enjoyment of music is an intellectual process, not a mechanical one. You can enjoy very low fi recordings of great music as much as hi fi. If you want to improve your enjoyment of music subjectively, it isn't hard to do that. Create a comfortable space to use as a listening room. Read up about the musicians and history of the music you listen to. Expose yourself to a broad range of music of all styles and eras to gain a wider frame of reference. Think about what you listen to. All of these things will improve your "subconscious" perception of music more than inaudible sound.

I think the upshot is that standard measurements, combined with listening, are usually 'good enough', but those measurements can still miss things which some listeners can detect.

What things are those? The way you would prove that would be with a controlled listening test. Do you know of any controlled listening tests where people were able to hear things that couldn't be measured? Or are you just assuming that there must be something because you want to believe that?

Human perception has been studied for centuries. Sound reproduction and measurements go back over 100 years. Most of the principles we're dealing with here were totally understood by the time of Bell Labs in 1920. If there is some aspect we're missing, I'm sure scientists would love to know about it so they can study it.

Analogsurvivor:

Response error = changes in tone
Amplitude error = changes in volume
Distortion = changes in the signal

It's possible for an error in one area to affect another I suppose. But timing error is changes in the signal. In its broadest definition any deviation is distortion. There are lots of different kinds of distortion.

 

[KeithEmo]

I don't disagree with anything you said there....

However, I think we are interpreting the semantics somewhat differently.

I absolutely agree that there would be no point in reproducing sounds which we cannot hear or otherwise percieve. However, I don't agree that the "currently accepted standards" are always correct. For example, "everybody knows that the range of human hearing is from 20 Hz to 20 kHz"... except for at least one recent test which found out that "under certain lab conditions humans can hear frequencies as low as 10 Hz". Likewise, virtually every test I've read about that was performed to test "the range of human hearing" used continuous sine waves. Well, that is NOT conclusive.... because our hearing MAY respond differently to non-continuous tones, or to various combinations of tones. You're actually talking about the results of one very simple, and very limited, test. That test was also designed to determine "the typical range of human hearing" - which is not the same as "the absolute limit". If you did a test for perfect pitch you would find that very few humans have it.... but a few really do.

As for "subconscious differences"... here's an extreme example...
Let's say I play a monaural music clip, centered exactly between the left and right channels. Now, I'm going to add to that a 21 kHz test tone, switching on and off at one second intervals, only from the left speaker. Will the apparent sound source of the music change due to the ultrasonic test tone (perhaps because it causes the gain in my left ear to shift slightly when it's on)? I don't know... and neither do you... because, at least as far as I recall, I've never read of anyone performing that test. And, if it turns out the result is perceptible, it will ONLY work on systems capable of reproducing that ultrasonic 21 kHz tone... and we'll have to be very careful to choose an amplifier and speakers which don't introduce other differences - perhaps intermodulation products from that ultrasonic tone.

And, as for your comment that "we don't perceive subconsciously".......

Here's my claim.....

I'm going to play the exact same audio test signal or piece of music for you from two different amplifiers.
The amplifiers will be identical in every measurement - except that one will add a 21 kHz sine wave tone at 110 dB SPL.
I predict that both will sound exactly the same - except that, with the one with the ultrasonic noise added, you'll soon get a headache.
I would call that headache... as it relates to the music... to be "a subconscious effect" (you certainly cannot hear it).
It's a physical effect, but, since you can't hear the 21 kHz tone, all you'll know is that, with THAT amplifier, you get a headache.

And, yes, I carefully chose that example to prove the point......
- a 21 kHz tone would be easy enough to verify by measurement
- but, being above 20 kHz, "it should not be audible and so shouldn't make any difference"
- it's not entirely implausible that an unstable amplifier might oscillate at 21 kHz

I also find that, with modern subwoofers, I can often FEEL vibrations that are clearly subsonic.
So, is a music system that fails to reproduce those "inaudible but clearly FELT vibrations" accurate or not?

I simply suggest that most tests are designed to determine "typical responses and averages".
I agree that MOST people can't hear sounds above 20 kHz...
But I would not be willing to say with absolute certainty that nowhere on the planet is there a human who can hear 25 kHz.
Likewise, recordings and equipment are generally designed for "a typical user" - whatever that means to the designer.

Back when vinyl was still very popular, all turntables had some tiny speed variations.
However, the better models were designed so their speed variations were "totally inaudible to most listeners when listening to music".
But, from what I'm told, some of those few people out there with "perfect pitch" found those "inaudible" variations quite annoying.
And if, as your test signal, you play a continuous 3 kHz sine wave instead of music, most of us notice even tiny variations in pitch.

As "a scientist" I tend to think and define things in absolutes.
Peanuts are NOT "100% non-toxic" - because I know at least one person who (according to his doctor) would die if he ate one.
Therefore, peanuts are "non-toxic to the vast majority of humans" - but we cannot say "all".

Measurements can definitely reveal things that are more subtle, but they can also reveal things that are totally inaudible. Unless you know where the line of threshold of perception lies, a measurement is just an abstract number with no practical context. We listen to music with our ears. It's just as important to measure our hearing as it is to measure the output of our stereo system. Ultimately, our ears are the last processing of the signal in the chain. There is no point pushing sound we can't hear through the chain. If we can't hear it, it isn't going to improve perceived sound quality. Our ears are what they are. We have to produce recorded sound to suit them.

Humans do not hear more discerningly over long periods of time. In fact, the longer you listen, the more your ears adjust to accommodate the sound. Auditory memory is notoriously short. For similar sounds, the length of time is just a couple of seconds. That's why controlled listening tests have direct A/B switching so the samples can be compared right next to each other.

We do not perceive subconsciously. Hearing is a mechanical process. You either hear something or you don't. How you *interpret* the sound might have some sort of subjective spin, but being able to hear a difference in a controlled comparison test is pretty much cut and dried. The purpose of blind testing is to remove subjective or subconscious bias. You can't apply bias if you can't tell which is which. And you can't apply bias if you can't hear it in the first place.

Subjective enjoyment of music is an intellectual process, not a mechanical one. You can enjoy very low fi recordings of great music as much as hi fi. If you want to improve your enjoyment of music subjectively, it isn't hard to do that. Create a comfortable space to use as a listening room. Read up about the musicians and history of the music you listen to. Expose yourself to a broad range of music of all styles and eras to gain a wider frame of reference. Think about what you listen to. All of these things will improve your "subconscious" perception of music more than inaudible sound.

Analogsurvivor:

Response error = changes in tone
Amplitude error = changes in volume
Distortion = changes in the signal

It's possible for an error in one area to affect another I suppose. But timing error is changes in the signal. In its broadest definition any deviation is distortion. There are lots of different kinds of distortion.

 

[KeithEmo]

And the best way to separate facts from "unsupported perceptions" is to do some sort of blind or double blind test.
Note that this is really not at all precluded by any of the normal arguments from "subjectivists".

A double blind test does NOT have to last five minutes.
If you believe that "a certain product is more fatiguing" then, by all means, test each sample for a full week instead of five minutes.
Take your two products, set up ten one-week trial periods, randomly listen to one each week, and correlate the results.
And, if you believe that "the stress of the testing protocol" is affecting the results, then let people hit the button themselves, whenever they want to.

Pharmaceutical companies routinely run tests that run for weeks, or even months or years, to get long term results.
People in the audio industry tend to shy away from this sort of test because of the expense and time involved.
Nobody in the audio industry really wants to hand out 500 amplifiers.... and then quiz the recipients a month later to see who has had more headaches.

To be honest, it is my personal opinion that most of the differences people claim to hear with cables are psychosomatic.
Furthermore, in the situations where tiny differences are measurable and perhaps audible, they are most often random.
For example, different cables, with slightly different electrical characteristics, may actually sound (and measure) slightly differently when used with certain equipment.
HOWEVER, it then becomes a matter of perception and marketing for the manufacturer to convince you that one or the other is BETTER.
(From my experience, audiophiles have an above average tendency to be willing to accept that "different = better", based on vague or unscientific claims.)

I would also point out that it's not unreasonable to simply accept both types of "experience" for what they are.
I've certainly paid more for equipment that looks nice, or that's more fun to use... but, if I'm paying for aesthetics rather than performance, I'd like to know.
(And, in that context, does it really matter whether 'the music sounds better" or "I enjoy it more"?)

None of us dispute that people respond to ''improvements'. The dispute is whether the perceived improvement is true or misperception. I certainly believe that people hearing improvements with expensive cables (and in many situations, amplifiers) are not hearing anything that can be reliably tested (as is backed up by the vast majority of blind tests conducted).

None of that changes the fact that certain gear 'sounds better' to some people. The phenomenology is not in doubt.

 

[bigshot]

Feel free to consider 10Hz significant, but it isn't needed for reproducing music and it can't really be heard- it can only be felt. It's the same with super audible frequencies. You can turn up a 28kHz squeal loud enough that you will be able to perceive it as sound pressure. But feeling and hearing are two different things. There's a point where frequencies no longer convey pitch. That point is well above 20Hz and well below 20kHz. Not to say that feeling an undefined sub bass thump isn't nice. It's just that the established thresholds are already overkill for listening to music. By 14 Hz or so, you're getting to about the limit that normal speakers can produce anyway.

Audiophiles focus on the extremes of the frequency range much more than they should. The difference between mediocre sounding music and great sounding music depends a lot more on the quality of the core frequencies than it does the extreme ends. If your core frequencies are clear and balanced, you could probably roll off the top and bottom octaves and music would still sound good.

 

[KeithEmo]

You really do have to consider the context... and that includes dates.

Human perception may have been studied for centuries....
But the equipment available to do so has changed considerably.
The best equipment in existence in 1920 could not generate what we would now call a "low noise" or "low distortion" test signal.
And, comparing what people considered to be "audible differences", while listening to music recorded on wax cylinders really was somewhat imprecise by today's standards.

Measurements can definitely reveal things that are more subtle, but they can also reveal things that are totally inaudible. Unless you know where the line of threshold of perception lies, a measurement is just an abstract number with no practical context. We listen to music with our ears. It's just as important to measure our hearing as it is to measure the output of our stereo system. Ultimately, our ears are the last processing of the signal in the chain. There is no point pushing sound we can't hear through the chain. If we can't hear it, it isn't going to improve perceived sound quality. Our ears are what they are. We have to produce recorded sound to suit them.

Humans do not hear more discerningly over long periods of time. In fact, the longer you listen, the more your ears adjust to accommodate the sound. Auditory memory is notoriously short. For similar sounds, the length of time is just a couple of seconds. That's why controlled listening tests have direct A/B switching so the samples can be compared right next to each other.

We do not perceive subconsciously. Hearing is a mechanical process. You either hear something or you don't. How you *interpret* the sound might have some sort of subjective spin, but being able to hear a difference in a controlled comparison test is pretty much cut and dried. The purpose of blind testing is to remove subjective or subconscious bias. You can't apply bias if you can't tell which is which. And you can't apply bias if you can't hear it in the first place.

Subjective enjoyment of music is an intellectual process, not a mechanical one. You can enjoy very low fi recordings of great music as much as hi fi. If you want to improve your enjoyment of music subjectively, it isn't hard to do that. Create a comfortable space to use as a listening room. Read up about the musicians and history of the music you listen to. Expose yourself to a broad range of music of all styles and eras to gain a wider frame of reference. Think about what you listen to. All of these things will improve your "subconscious" perception of music more than inaudible sound.



What things are those? The way you would prove that would be with a controlled listening test. Do you know of any controlled listening tests where people were able to hear things that couldn't be measured? Or are you just assuming that there must be something because you want to believe that?

Human perception has been studied for centuries. Sound reproduction and measurements go back over 100 years. Most of the principles we're dealing with here were totally understood by the time of Bell Labs in 1920. If there is some aspect we're missing, I'm sure scientists would love to know about it so they can study it.

Analogsurvivor:

Response error = changes in tone
Amplitude error = changes in volume
Distortion = changes in the signal

It's possible for an error in one area to affect another I suppose. But timing error is changes in the signal. In its broadest definition any deviation is distortion. There are lots of different kinds of distortion.

 

[gregorio]

My understanding is that FR curves can be generated in various ways, and can't reflect timing errors. Again, a measurement which can't tell the whole story.

Define the "whole story" or "whole truth". If you're talking about a measurement down to infinity, no we can't do that and therefore we don't know the whole story/truth but a measurement beyond or way beyond even the vaguest chance of audibility, yes, we can do that. If the "whole story" is the whole story of what our ears can respond to, then yes, we can get the whole story.

The point is that real signals can't be fully characterized by any measurement or set of measurements.

So you're saying that Nyquist's theory was wrong, that Shannon's mathematical proof of that theory was wrong and consequently digital technology does not exist? This is a problem is that your "point" obviously seems entirely reasonable and logical to you but, because you don't know why it's wrong, you don't understand the consequences of what you're asserting, and those consequences are not in the slightest bit reasonable or logical. The consequences are, that for your assertion to be true, Shannon's proof must be wrong and Shannon's proof is the foundation of modern information theory. If modern information theory is in fact wrong, then all the digital devices which rely on it would not work; all mobiles, computers, laptops, tablets, digital TVs. So, without apparently realising it, you have made an assertion which is clearly ridiculous, presumably even to you.

But, there is an even more obvious reason why you're "point" is incorrect. If we can't measure it, then we can't record it and you can't reproduce it! So, even if there is something we cannot "characterise" (which there isn't) then it's irrelevant anyway because it does not exist in any recording you're trying to reproduce. You can't just keep making assertions, completely ignore the simple logic of why your assertion must be false and then keep repeating that assertion!

Two real signals could differ slightly in some way, but a given measurement or set of measurements may miss the difference and indicate that the signals are the same.

I've already given you the details of a very old, well known and widely used test which is GUARANTEED to precisely indicate if two signals are the same. This is another example of you making an assertion, having that assertion proved incorrect, simply ignoring that proof and then repeating your original assertion again, although this time in a different thread! There's only one logical conclusion to draw if you continue to do this.

G

 

[bigshot]

Human perception may have been studied for centuries....
But the equipment available to do so has changed considerably.

If you haven't ever done any reading on Bell Labs, you should. It's amazing how many of the things we discuss every day in Sound Science come from research done at Bell Labs. And you'd be surprised at how advanced they were technologically. Just because something didn't exist in consumer audio at the time, it doesn't mean that it wasn't being worked with at Bell Labs. EMI also was involved in pioneering early research.

 

[Phronesis]

I agree with what KeithEmo said in his last two posts.

Measuring is good, but has limitations. Listening tests are good, but they also have limitations and pitfalls, which aren't eliminated by 'rigorous' blind testing. People may hear real things that particular measurements missed. People may hear things that aren't really there physically. Measurements may reveal real things which listeners didn't hear. In the end, physical and psychological phenomena are complex beyond our full understanding (so we have to use models), and uncertainties can't be entirely eliminated (only reduced).

Are any of these points really controversial?

 

[bigshot]

What sort of real sounds can be heard that can't be measured? I'm interested to know. And can the things that can't be measured be recorded?

I know that we can measure things we can't hear, but that doesn't mean that the inverse is true. The way I see it is that measurements tell the whole story and listening tests tell you what part of that really matters.

 

[KeithEmo]

I agree...
And I like your camera analogy...

However, especially when it comes to digital audio, the catch is often in what we choose to measure... and what we choose to claim "doesn't matter".
(And the difference between how machines "perceive things" and how humans do.)

I'll give you a really simple example.

Let's say that I generate two entirely separate monaural channels of "white noise" - just random noise.
I put one in the right channel of a stereo file and the other in the left channel.
This would technically be referred to as "uncorrelated white noise" - which simply means that the two channels aren't specifically related to each other.
If we compare those two channels, they will have the same frequency response, and the same frequency spectrum, but the actual bits will be randomly different.
And, if we listen to the stereo pair, we'll hear a nice gentle "hiss" spread across the sound field.

Now, instead of doing that, I'm going to generate a single channel of monaural white noise and copy it to BOTH stereo channels.
This would be considered to be "monaural white noise" or "channel correlated white noise".
Now, if we compare those two channels, they will have the same frequency response, and the same frequency spectrum, but the actual bits will be the same.
And, if we listen to this pair of channels in stereo, we will hear a single white noise source centered between the speakers.

More to the point, if we compare our two samples, using most typical measurements, they will both appear simply as "stereo white noise".
Standard measurements will not indicate whether the noise in the two channels is correlated or not....
However, the two files will sound very different.

This sounds trivial....

However, as part of its process to minimize storage requirements, one of the early multi-channel CODECs would automatically eliminate duplicate high-frequency information.
So, with EITHER of our files, this CODEC would recognize that "the high frequency portion of the spectrum was the same for both channels"....
And, based on that fact, in order to minimize storage space, it would simply store the high-frequency portion of one channel, and discard the other.
Then, when reconstructing the file, it would use the high-frequency portion of the spectrum of single channel it had stored to fill BOTH reconstructed channels.

As a result, if you encoded and then decoded the sample with the monaural white noise, it would come out more or less the same as it went in.
However, if you encoded and then decoded the sample with the uncorrelated white noise, when you played it back it would play back as correlated white noise.
My point here is not that the CODEC "got it wrong".....
My point is that you would be UNABLE to detect the error using almost all standard types of measurements.
The difference would be audible extremely obvious, but the normal measurements many people use when checking files for problems would overlook it.
(A "vectorscope" or "stereoscope" or "separation meter" would pick it up immediately... and mastering engineers use those... but not usually people "just checking files".)

My point is that, while this error would be easy to detect if you were looking for it...
And immediately audible to a human being...
It would probably be overlooked if you were "analyzing the file by meter alone".

I don't know if you're arguing against this point or not, but just for the usual sake of pedantry, if a human being can detect something, so can measurement equipment - if you have the time & budget to do the appropriate measurements. There is often no good reason to do the measurements, but that doesn't mean they can't be done in principle.

Let me put it another way. How could something be reproducible but not measurable? Meaning - if it exists in a recording, it was "measured". The signal passed through equipment that "measured" the signal, and came out the other end. To say otherwise is like asserting people's eyes can see things in prints of photos that cameras can't detect.

But the photo CAME FROM a camera. So that would be a fairly silly thing to assert.

Is there maybe some aspect of live sound that is not reproducible in principle? I don't think so, but I don't know enough about acoustics to rule it out, out of hand. But we're talking about recorded audio, which in a real sense is sound that has been measured already.

Now if you want to argue that there is distortion that is so subtle that it's beyond reproducibility of any kind - with any equipment that exists on earth - you're basically at the point of saying you can hear things like quantum tunneling in your headphones, so at that point I get off the argument train and walk back to town by myself.

 

[Zapp_Fan]

Good example, and I'm certainly not arguing that all audible differences are EASY to measure. It's well known that measuring something as quotidian as high frequency response in headphones is at least very tricky to get right. My point is simply that there's nothing out there in recorded sound that is audible and is also *literally impossible* to measure, or unknown to science altogether. As you've explained quite well, there are many cases where the measurement to use is not obvious, obscure, difficult, or expensive to carry out.

The point as applied to audiophile stuff is simply that there is no hidden / mysterious thing about consumer audio gear (or even pro gear) that *can't* be measured, but can be heard. I will concede that there is probably, once in a while, stuff that can be heard but nobody has bothered to measure on a given piece of gear. And there are probably, (as in edge cases like in your example) many isolated cases of things that *really shouldn't* be audible, but are due to unusual circumstances.

But yeah, to say there is no way to measure something audible in a recording is like saying the recording contains something that is impossible to record, and you also heard it with your ears. And then to make it even worse this is sometimes why people are spending money on things.

 

[bigshot]

Everyone agrees that there are elements of sound in the real world (i.e. the effect of space on sound) that can't be captured or measured, but here we're talking about recorded sound. Let's keep focused on the point of this thread. We're tracing the degradation of the signal from the source (the original recording) through conversion from digital to analogue, through amplification. As far as I'm concerned, I demand that this part of the chain be audibly perfect, and in my experience, it isn't hard to achieve that. The transducers and space at the back end are the wild card.

You can say that you don't know of any people who can hear 26kHz, but you're keeping your mind open that someone like that might exist. That's fine. I can keep my mind open to the possibility of flying pigs and children that can distort space and time with the power of their will. That doesn't mean that it exists. We deal with the evidence we have. Rigorous scientific tests tell us people hear from 20 to 20. Measurements and listening tests tell us that modern amps and DACs are audibly transparent. You can believe that there may be an exception out there, and you may think that there may be some audible thing out there that a DAC produces that we can't measure. Great. Until there is evidence to show that, it isn't relevant.

Controlled listening tests and measurements are powerful tools. The only thing they can't help with is curing the wiggly logic of preconceived notions and bias. If we want to know the truth, we have to strive to eliminate that kind of thing, not embrace it.

 

[Phronesis]

Once again, I agree with KeithEmo's latest post.

I think there's an issue in this discussion with ideas being taken to extremes. I don't think anyone is arguing that standard measurements are necessary and useful, and good enough for most audio purposes.

The issue is that there could be audible differences which aren't detected by standard measurements, because those particular measurements aren't able to detect those differences. Those audible differences could be subtle - maybe too subtle for many or most listeners to consistently notice or care about - but they may significant to some listeners who are looking for the last increments in accuracy or fidelity, and can hear the difference.

I don't doubt that most of the differences that people claim to hear with DACs, amps, and cables aren't real physical differences, but I'm open to the possibility that some of those differences are real physical differences which are audible to some people. Blind testing is a good method to help judge if the reported differences are real, but it's not 100% conclusive, because there are issues with generalizing from the test conditions, and the statistics will only provide indications of probabilities (e.g., "there's only a 4% chance that he could really distinguish between A and B" isn't the same as a 0% chance).

And again, I would note that we don't have everything figured out about how the physical world works - this is why science evolves, rather than being static. We have good models, but they're models, not mirrors of reality. Can two decent cables sound different? It's hard to believe based on our current models, but we can't rule out the possibility on that basis alone. Similarly, we can't rule it out based on measurements either, since the measurements are also made in the context of those measurements ("all observation is theory-laden"), and they may not be sufficient sets of measurements anyway.

 

[sonitus mirus]

Once again, I agree with KeithEmo's latest post.

I think there's an issue in this discussion with ideas being taken to extremes. I don't think anyone is arguing that standard measurements are necessary and useful, and good enough for most audio purposes.

The issue is that there could be audible differences which aren't detected by standard measurements, because those particular measurements aren't able to detect those differences. Those audible differences could be subtle - maybe too subtle for many or most listeners to consistently notice or care about - but they may significant to some listeners who are looking for the last increments in accuracy or fidelity, and can hear the difference.

I don't doubt that most of the differences that people claim to hear with DACs, amps, and cables aren't real physical differences, but I'm open to the possibility that some of those differences are real physical differences which are audible to some people. Blind testing is a good method to help judge if the reported differences are real, but it's not 100% conclusive, because there are issues with generalizing from the test conditions, and the statistics will only provide indications of probabilities (e.g., "there's only a 4% chance that he could really distinguish between A and B" isn't the same as a 0% chance).

And again, I would note that we don't have everything figured out about how the physical world works - this is why science evolves, rather than being static. We have good models, but they're models, not mirrors of reality. Can two decent cables sound different? It's hard to believe based on our current models, but we can't rule out the possibility on that basis alone. Similarly, we can't rule it out based on measurements either, since the measurements are also made in the context of those measurements ("all observation is theory-laden"), and they may not be sufficient sets of measurements anyway.

In your first sentence you point out an issue with the discussion topics being taken to extremes, and then your last paragraph about cables seems to pull in some potentially pathological extremes to justify your position. I mean, maybe in some extreme situation would 2 cables change the sound enough to be audible while still measuring similarly.

 

[Phronesis]

In your first sentence you point out an issue with the discussion topics being taken to extremes, and then your last paragraph about cables seems to pull in some potentially pathological extremes to justify your position. I mean, maybe in some extreme situation would 2 cables change the sound enough to be audible while still measuring similarly.

I'm a skeptic about cables making any difference at all. It's hard for me to believe, based on the current theory/models. I've never bought an expensive cable, and the cables that come with my Hugo 2 and LCD-3 seem pretty ordinary. I used the example mainly because, even with something apparently as simple as a cable, there may be physics involved which we don't sufficiently understand, and therefore could affect accuracy in some way.

Again, I doubt it, but I can't rule it out -- that's what I mean by not going to extremes. It's ok to say 'I doubt it' or 'I strongly doubt it' or 'I'm pretty sure', but when people say things like 'I'm sure' or 'it's impossible', that often indicates dogmatic belief which isn't at all in the spirit of science. It took open minds to develop quantum mechanics - which has some apparently 'crazy' ideas - yet it has a lot of applications in technology.