DAC audible differences are finally proven with ABX?

[pma]

Trying to convince those that don't even remotely comprehend scientific method is a lost cause

Exactly. Hundreds of posts that would change nothing.

 

[NTTY]

Hi Flo, I’ve been silently following this thread, and I’m sure I missed some of it. I appreciate the effort and the opportunity offered to all of us to get to know more with the purchase of the ABX box.
You’ve reviewed several of my devices in the past, that was eye opening to me, and probably not only for me.
For these reasons, I’d like to contribute to the effort by sending you 200 chf, as I see we’ll get one more opportunity to learn something new (or maybe not).
Talk to you soon!

Thank you William!!
You didn’t have to and I appreciate, of course.
It was fun testing some of your vintage gears, a learning for me! Feel free to send more my way

 

[kemmler3D]

Smacks of either dishonesty, or, at best, lack of confidence in his own test result.

Devil's advocate on that point: a core premise of even interacting with the subjectivist audiophile community is believing your ears and attributing heard differences according to one's intuition about the gear. Doing an ABX at all is way beyond the rigor demanded by subjectivists. I imagine that to them it's like cooking a gourmet meal in a chemistry lab so you can prove it tastes good. A silly exercise for nerds that insist on missing the point.

Subjectivists find questions that don't assume you heard what you think you heard offensive because it cuts at the basic premise of their worldview. (The placebo shaped elephant in the room could not possibly explain what they've experienced.)

From that POV, further questioning of the gear and methods probably smacks of bad faith and sour grapes. So reluctance to go along with additional investigation probably makes total sense from that side. Why entertain a dismal attempt to dismantle a beyond-rigorous result?

The ASR POV is that we'd like to learn something about what was heard, potentially something important, and in the process validate or invalidate the test.

The AS POV is probably that we're a bunch of sore loser cranks who can't accept excellent evidence that was actually superfluous in the first place. The idea that we really look at audio primarily from a technical / scientific lens probably feels less credible to them than we realize, because that mindset is hard to empathize with for them.

The flip side is we find it hard to imagine not wanting to know or not caring about how the audio gear works in a technical sense.

Anyway, I'm not sure that's what's driving it but the reluctance to share further data could be an attitude mismatch of a different type than it seems to us.

 

[NTTY]

Devil's advocate on that point: a core premise of even interacting with the subjectivist audiophile community is believing your ears and attributing heard differences according to one's intuition about the gear. Doing an ABX at all is way beyond the rigor demanded by subjectivists. I imagine that to them it's like cooking a gourmet meal in a chemistry lab so you can prove it tastes good. A silly exercise for nerds that insist on missing the point.

True, but 1) Hell is paved of good intentions, 2) Devil is the details

 

[antcollinet]

could be an attitude mismatch of a different type than it seems to us.

Fair point - I'll amend my post accordingly.

 

[AndyBell]

I'm not sure that's what's driving it but the reluctance to share further data could be

As mind reading is not a human skill (my wife excepted, of course ) the reason could be anything…

It is what it is - and I don’t think we’ll ever find out and further speculation won’t achieve anything.

Once the box is analysed we may have an explanation for the results.

 

[kemmler3D]

As mind reading is not a human skill (my wife excepted, of course ) the reason could be anything…

It is what it is - and I don’t think we’ll ever find out and further speculation won’t achieve anything.

Once the box is analysed we may have an explanation for the results.

Yes, I'm not necessarily saying there isn't something funny going on. I tend to be very cynical on these points myself. But I have no proof of that either. And I can imagine honest (emotionally driven) reasons to not engage.

But at the end of the day we will probably just keep posting what amounts to "why won't he give us the data??" until we get some data. You can quote me on that when we hit post # 1600 and still no data.

 

[RexrothPigeon]

I wonder what percentage of these 1428 posts are people complaining about how many posts there are and claiming there's no point to posting further. Feels like a lot.

 

[gzost]

@kemmler3D - Exactly. The subjectivists know that all DACs sound different, and they experience it with every change to their system. So it has to be bad faith by the objectivists if one of theirs does one of these weird, cumbersome and unnecessary ABX tests, easily proves what they know, and we still do not accept the results.

To them, of course, the opposite result would also not be surprising since Robert Harley proved in a stereophile column long ago that ABX does not work for audio. The doublethink needs to be strong on their side, else heads would explode with all the contradictory beliefs they need to hold depending on the situation, or sometimes all at once.

 

[AndyBell]

I tend to be very cynical on these points myself

Part of my job is as a data analyst and listing possibilities is something I tend to do all the time.

In the absence of data all possibilities are, well, possible. Even ones we’d rather avoid mentioning

 

[Plcamp]

and they experience it with every change to their system.

In fact, scientific exploration has shown that simply believing a change makes an audible difference will cause you to hear a difference (even if there isn’t one).

Which, curiously, admits the possibility that the ultimate experience in HiFi might hinge on you believing nonsense before you listen?

 

[PristineSound]

Devil's advocate on that point: a core premise of even interacting with the subjectivist audiophile community is believing your ears and attributing heard differences according to one's intuition about the gear. Doing an ABX at all is way beyond the rigor demanded by subjectivists. I imagine that to them it's like cooking a gourmet meal in a chemistry lab so you can prove it tastes good. A silly exercise for nerds that insist on missing the point.

Subjectivists find questions that don't assume you heard what you think you heard offensive because it cuts at the basic premise of their worldview. (The placebo shaped elephant in the room could not possibly explain what they've experienced.)

From that POV, further questioning of the gear and methods probably smacks of bad faith and sour grapes. So reluctance to go along with additional investigation probably makes total sense from that side. Why entertain a dismal attempt to dismantle a beyond-rigorous result?

The ASR POV is that we'd like to learn something about what was heard, potentially something important, and in the process validate or invalidate the test.

The AS POV is probably that we're a bunch of sore loser cranks who can't accept excellent evidence that was actually superfluous in the first place. The idea that we really look at audio primarily from a technical / scientific lens probably feels less credible to them than we realize, because that mindset is hard to empathize with for them.

The flip side is we find it hard to imagine not wanting to know or not caring about how the audio gear works in a technical sense.

Anyway, I'm not sure that's what's driving it but the reluctance to share further data could be an attitude mismatch of a different type than it seems to us.

Well said. I don't think it's necessary to think the worse of someone from the audio subjectivist world who actually when to great lengths doing an ABX, issues or no issues, perfect or imperfect as the test may or may not be.

I think he earned some grace with me at the moment, Amir's upcoming detail will provide more insight, at that point further judgement can be decided.

 

[Plcamp]

at that point further judgement can be decided.

I am pessimistic.

As was said previously, unless the original test is repeatable (and it isn’t without the full supply of everything Josh did) we simply can’t come to a definitive judgment. Comparator is just one piece.

Maybe some big revelation comes from the comparator testing, but right now I doubt it.

 

[welwynnick]

Anyway, I'm not sure that's what's driving it but the reluctance to share further data could be an attitude mismatch of a different type than it seems to us.

I think a possible reason is that the ADC recordings were done with the wrong sensitivity setting, and are nowhere near good enough to be useful.

 

[PristineSound]

I am pessimistic.

As was said previously, unless the original test is repeatable (and it isn’t without the full supply of everything Josh did) we simply can’t come to a definitive judgment. Comparator is just one piece.

Maybe some big revelation comes from the comparator testing, but right now I doubt it.

The way how I see the possible outcomes:

1) There is something with the ABX that gives a tell. In such case, can't blame Josh, it wasn't intentional and he never claimed to be an expert (at least I don't think he did). I would give him credit for effort, although he may be a bit embarrassed. But I would give him respect for trying to bring science.

If there is nothing wrong why the ABX. In such case, I would expect Amir to use the device to do ABX himself, and replicate Josh's test using the same DACs (if he had access to it), possibly with many other DACs:

2) if Amir passes such test, I am expecting Amir to investigate the DUTs. And I would be excited to learn more.

3) If Amir fails the ABX, I am expecting few others trying their turn with ABX, I am hoping by that time, the open source ABX device will be available. And if no one else can pass an ABX (which is what I would expect), then we can release our judgement on Josh. I think he should be innocent until proven guilty of forgery.

 

[Plcamp]

The way how I see the possible outcomes:

1) There is something with the ABX that gives a tell. In such case, can't blame Josh, it wasn't intentional and he never claimed to be an expert (at least I don't think he did). I would give him credit for effort, although he may be a bit embarrassed. But I would give him respect for trying to bring science.

If there is nothing wrong why the ABX. In such case, I would expect Amir to use the device to do ABX himself, and replicate Josh's test using the same DACs (if he had access to it), possibly with many other DACs:

2) if Amir passes such test, I am expecting Amir to investigate the DUTs. And I would be excited to learn more.

3) If Amir fails the ABX, I am expecting few others trying their turn with ABX, I am hoping by that time, the open source ABX device will be available. And if no one else can pass an ABX (which is what I would expect), then we can release our judgement on Josh. I think he should be innocent until proven guilty of forgery.

Well I think the major point here is that we will never know all the circumstances surrounding the Josh test (even after Amir tests the comparator), so we will never be able to explain his results definitively.

So we should recognize that and forget about trying to explain it…I am suggesting no data sufficient to explain it is going to become available (unless maybe the comparator test yields something truly surprising - Your scenario 1).

What’s left is a need to conduct well controlled and repeatable ABX on (measured transparent) DACs…maybe best done with level aligned recorded files instead of any comparator box…and with a large # listeners, each doing statistically significant sessions.

That holds some promise to resolve the core question (without ever having the need to ‘pass judgment’ on Josh’s outcome) and probably has the best chance of a wide audience accepting the results.

 

[pma]

What’s left is a need to conduct well controlled and repeatable ABX on (measured transparent) DACs…maybe best done with level aligned recorded files instead of any comparator box…and with a large # listeners, each doing statistically significant sessions.

I agree. It is much better to prepare level matched and time aligned recordings, 64 bit or 32bit precision in SW is much more than good enough and no magic box will meet that. The magic box is only a pain in a game. But doing that makes the mystery go away, and because the show must go on, the mystery is necessary.

 

[Plcamp]

I agree. It is much better to prepare level matched and time aligned recordings, 64 bit or 32bit precision in SW is much more than good enough and no magic box will meet that. The magic box is only a pain in a game. But doing that makes the mystery go away, and because the show must go on, the mystery is necessary.

Amir might hate me for this time consuming suggestion…

…but it would be great if every DAC test had, along with its measurement results, a time aligned, level matched file of standard recordings?

 

[Plcamp]

It is much better to prepare level matched and time aligned recordings

Question: If that was done for any two dacs…would that not allow the difference to be calculated as “X db down” vs signal and essentially prove no audible difference exists? Or are there high frequency phase effects beyond simple amplitude that would defeat this approach?

Looking to defeat the mystery.

 

[pma]

Do you want for me to answer to the best of my knowledge? The are no audible differences between well designed DACs. There are some measurable differences, but not audible differences. Our measuring instruments are in orders more sensitive than our ears. If there will be an audible difference (the verified one), it would be by mistake. Technical error like system ground loops, unmatched level. And this would be unique error, system dependent. Adding recorded music files by Amir would only be tons of work and time spent. And it would not stop audiophile debates at all.