Not only do I agree with you 100%, I'll go even further- NO test can show that differences between A and B (whatever the variable is) are inaudible. Now, if people who claim differences fail to show any audibility in test after test after test, one can draw a reasonable working conclusion that the differences are inaudible. BUT... one must be open to the possibility that someone, somewhere might demonstrate that A and B can be distinguished.
Falsifiability - Wikipedia, the free encyclopedia
For the case of electronic effects beyond level, noise, frequency response, gross distortion, overload characteristics, and gross phase errors, nothing but white swans have been seen. The overwhelming probabilities are that there's just white swans. Anyone who claims that, after forty years of fruitless searching, there are black swans, but that they're just really, really hard to find needs to find one in order to be taken seriously, and if they want to spend their time doing so, great, that's their business. Everyone else has moved on to looking for other birds. But if the black swan enthusiast manages to get solid evidence that they exist, great, we now know something new. But no-one is holding his breath.
Anyone who claims that there are black swans all around you, it's just that they turn white when you try to photograph or capture them should be regarded with either suspicion or pity, depending on whether they're trying to sell something or are just gullible.
That is the point here in Australia we do have black swans so the statement that all swans are white was falsified as soon as somebody discovered it.
A correct scientific statement is, all the swans I have seen are white, this is an inductive statement because it is always up for revision and is falsified if you see a black swan.
If you say however that all white birds are swans, then this is an axiomatic statement that is not falsified by seeing a black swan because you have defined swans as all white birds.
In many areas of science if you go through the experimental design people used, you can show that what they are actually saying is the second sort of statement, which results in an outcome that is pre determined because it is axiomatic. In other words what it does is to attempt to demonstrate that a hypothesis is true rather than to test if it is true.
So in the end science can only say that all swans I have seen so far are white, anybody claiming it can do more than this is in error.
rcw
A correct scientific statement is, all the swans I have seen are white, this is an inductive statement because it is always up for revision and is falsified if you see a black swan.
If you say however that all white birds are swans, then this is an axiomatic statement that is not falsified by seeing a black swan because you have defined swans as all white birds.
In many areas of science if you go through the experimental design people used, you can show that what they are actually saying is the second sort of statement, which results in an outcome that is pre determined because it is axiomatic. In other words what it does is to attempt to demonstrate that a hypothesis is true rather than to test if it is true.
So in the end science can only say that all swans I have seen so far are white, anybody claiming it can do more than this is in error.
rcw
In a vague sense, yes. The listener is the detector, but also part of the claim. For the listener, the DUT is the box of gain (or whatever is being claimed to have an audible effect). For the experimenter, the DUT encompasses the listener, the system, and the source material.
Again, it is critical to note that there is no one-size-fits-all experimental protocol. The protocol needs to be designed to accommodate the specifics of the claim. For example, in the case of evaluating the claim "John Smith can hear the effects of 128k MP3 data compression" (something with well-established audibility), the setup will be extremely different than evaluating (for example) the claim that "John Smith can hear a difference between silver and copper wire in his interconnects," for which there is no well-established audibility.
Again, it is critical to note that there is no one-size-fits-all experimental protocol. The protocol needs to be designed to accommodate the specifics of the claim. For example, in the case of evaluating the claim "John Smith can hear the effects of 128k MP3 data compression" (something with well-established audibility), the setup will be extremely different than evaluating (for example) the claim that "John Smith can hear a difference between silver and copper wire in his interconnects," for which there is no well-established audibility.
It does allow you to put his claim in context and know where to put the burden of proof.
In aggregate, when John Smith, Bob Jones, Akimi Poontangawi, Hershel Abramowitz, Lee Chen, Fibbo Basizi, Jane Barrington-Smythe, T'keenah Suggs, Biffy Montague, and Jean-Claude Deschamps all fail to demonstrate that, despite their claims, they are unable to distinguish A from B, that one can absolutely say, "No-one has demonstrated an audible difference between A and B," and provisionally operate under the assumption that A and B are audibly indistinguishable (while keeping an open mind that Dieter Gruenhauser might possibly do so, while not holding one's breath that this will ever happen).
But again, burden of proof shifts- scientists don't waste time chasing every claim of perpetual motion, it's up to the claimant to provide evidence. Of course, as in fashion-niche audio, perpetual motion claims are notable mostly for their creative excuse-making and dark conspiracy theories.
In aggregate, when John Smith, Bob Jones, Akimi Poontangawi, Hershel Abramowitz, Lee Chen, Fibbo Basizi, Jane Barrington-Smythe, T'keenah Suggs, Biffy Montague, and Jean-Claude Deschamps all fail to demonstrate that, despite their claims, they are unable to distinguish A from B, that one can absolutely say, "No-one has demonstrated an audible difference between A and B," and provisionally operate under the assumption that A and B are audibly indistinguishable (while keeping an open mind that Dieter Gruenhauser might possibly do so, while not holding one's breath that this will ever happen).
But again, burden of proof shifts- scientists don't waste time chasing every claim of perpetual motion, it's up to the claimant to provide evidence. Of course, as in fashion-niche audio, perpetual motion claims are notable mostly for their creative excuse-making and dark conspiracy theories.
It seems that ABX tests have never been able to detect "small" differences in sound quality,
only obvoius ones (like the 128k MP3) you mentioned in post #27524.
So this would indicate these "small" differences (e.g. cheap receiver vs. a tube amp etc.)
do not exist at all.
I´m very sceptical. Time to validate (as Jakub2 already said) that ABX listenenig tests
are capable of detecting such differences (let´s take something that has been proven to
be audible by some other scientific methodology) at all.
I can easily imagine a test where John and Jane can´t distinguish a photograph from a
decent camera vs. a 99,- point and shoot. Time to sell my Hasselblad before anyone finds out ?
only obvoius ones (like the 128k MP3) you mentioned in post #27524.
So this would indicate these "small" differences (e.g. cheap receiver vs. a tube amp etc.)
do not exist at all.
I´m very sceptical. Time to validate (as Jakub2 already said) that ABX listenenig tests
are capable of detecting such differences (let´s take something that has been proven to
be audible by some other scientific methodology) at all.
I can easily imagine a test where John and Jane can´t distinguish a photograph from a
decent camera vs. a 99,- point and shoot. Time to sell my Hasselblad before anyone finds out ?
It is an interesting argument (decade after decade of trying) , because it implies that numerous controlled listening tests on these topics were done.
That is quite unlikely due to statistical reasons. Otherwise we had dozens or hundreds of positive test results on amplifier sound, dac sound and so on.
But it seems that the number of listening tests is quite small and the sample size order in these is small as well.
If somebody gets a positive test result it is dismissed as it happened to Michael Fremer with his amplifier listening test.
And so SY slightly adjusts his "experimental question" during the ongoing discussions. It starts with "can John Smith or others detect a difference between two DUTs" but is already "can John Smith or others detect a difference between two DUTs (under these test conditions)" because the experimenter obviously doesn´t care about scientific requirements.
In the discussion, although the first question was never tested, the conclusions is drawn that no difference exists, because "John Smith and other were never able to detect......"
So John Curl might think that this point is only a small subset of ´scientific reasoning´ but imho it covers quite a large portion of it. Design of experiments, Operationalisation, statistical analysis, cognitive psychology, psychoacoustical considerations and so on.
That is quite unlikely due to statistical reasons. Otherwise we had dozens or hundreds of positive test results on amplifier sound, dac sound and so on.
But it seems that the number of listening tests is quite small and the sample size order in these is small as well.
If somebody gets a positive test result it is dismissed as it happened to Michael Fremer with his amplifier listening test.
And so SY slightly adjusts his "experimental question" during the ongoing discussions. It starts with "can John Smith or others detect a difference between two DUTs" but is already "can John Smith or others detect a difference between two DUTs (under these test conditions)" because the experimenter obviously doesn´t care about scientific requirements.
In the discussion, although the first question was never tested, the conclusions is drawn that no difference exists, because "John Smith and other were never able to detect......"
So John Curl might think that this point is only a small subset of ´scientific reasoning´ but imho it covers quite a large portion of it. Design of experiments, Operationalisation, statistical analysis, cognitive psychology, psychoacoustical considerations and so on.
Well, did it again! Wrote a long message and lost it because of timeout of the software ...
Anyway, the gist of it was, thanks for your heartfelt advice, but I do carefully pace myself with sound, and always back off when I pick the signs of overload. An idiot at high school letting off a starting pistol in one ear has made me very sensitive to excessive loudness, so I only listen to high intensity, live quality sound, not the compressed, audibly distorted replay one gets from many systems at higher levels ...
I do appreciate your showing of concern,
Frank
Sy, i try to avoid any intentionally misreading of other posts; you´ve stated what your "experimental question" should have been, and i´ve stated what it really was, if no controls were used.
And you were drawing conclusions due to the socalled fact that nobody (decades after decades failed in tests to show) did show a positive test result.
Please provide a few arguments to back up your assertion.
In which way were the standards suspended?
Any double blind test that fails to take into consideration how our auditory system works, really says nothing all, and shouldn't be taken seriously.
A double blind test that's takes into account that we learn sounds and allows for habituation (this takes time) to take place would be worth something. Why? Because you don't need to see what you're listening to, to hear a difference.
I have ears but I can't echolocate Sy. I've tried. Yet some blind people do Echolocate. They click their tongue and listen for differences in the sound caused by reflections in order to navigate. This takes a whole lot of practice and time, learning these differences, forming conditioned reflexes. It's a difficult test ... but they do it and play games like basket ball or go roller blading down a street.
My ears are incredibly insensitive to these differences and I just can't hear a difference unless it's pretty big. I'm not blind so I'm not going to put in the effort to learn to echolocate.
- Status
- Not open for further replies.