"Blind is the reality" makes a nice hypothesis but it is quite hard to find corrobating for it, isn´t it?
Assuming it is meant in a way, that under "blind" test conditions the result is _really_ reflecting what people "in reality" hear (purely responding to an actually different activation of the physiological apparatus), you seem to neglect the actual evidence for the contrary, which is imo a bit disturbing for a scientist.
It is fact that listeners in controlled "blind" listening experiments incorrectly report hearing a difference, when listening to an identical stimulus twice in a row. The false answer rate is as high as 80% in such trials. (When evaluating a multidimensional task)
So if you state that "blind is the reality" then you have to conclude that in reality there exist a difference between the identical stimulus, so you´ve find the "ghost in the machine".
The hypothesis " blind creates another internal representation of reality" seems to be a lot more compatible to the factual evidence than yours.
In accordance to this hypothesis (and the scientific evidence ) is the extension, that "ABX" creates another internal representation then other "blind test protocols" .
Sorry to be pedantic at this point, but no. Reliability in the "test context" means precision and repeatability of an experimental result.
What you mean is the validity; if a listener in a sighted test always prefers "Brand A" because he believes that "Brand A" is superior to "Brand B" (and has always been superior) despite any real difference (if any given at all) then it´s reliable (because it´s always the same result) but not valid, because it does not test the real independent variable but the bias instead.
The main problem with sighted listening is, that nobody can show the internal validity of the procedure (without additional measures).
Same situation as under "blind test conditions", people still hear differences that not exist in the stimulus (known from trials where the identical stimulus is presented twice in a row).
It´s a problem of multidimensional evaluations -which listening to music in a test is - much less common if testing onedimensional differences.
Can be, but you have to ensure that listeners, participating in "ABX tests", do not underperform under this special conditions.
It is known, since roughly 70 years, that "ABX conditions" evoke a higher proportion of wrong answers in trials when compared to other "blind test" protocols (when testing the same sensory difference).
Still, the results of the tests speak for themselves. Those with most to say are also those least likely to participate, nothing new there. Like polytricksians 
They probably think the test is somehow beneath them and consequently hold those who do it in contempt. It could be that they are afraid to, I think this is very likely, imagine how it would be for them if they found themselves coming to a different conclusion than the one they expected to.
This test is good because it's inclusive, anyone can do it, even us poor souls who can't do an ABX test, also it's anonymous, so you don't have to display your inability if you don't want to. There is consequently no pressure whatsoever, people can be entirely honest, and why wouldn't they be?
They probably think the test is somehow beneath them and consequently hold those who do it in contempt. It could be that they are afraid to, I think this is very likely, imagine how it would be for them if they found themselves coming to a different conclusion than the one they expected to.
This test is good because it's inclusive, anyone can do it, even us poor souls who can't do an ABX test, also it's anonymous, so you don't have to display your inability if you don't want to. There is consequently no pressure whatsoever, people can be entirely honest, and why wouldn't they be?
Last edited:
Those who know something about perceptual testing have the most to say - those who don't know, do the test & accept the results unquestioningly - same as it ever was (as David Byrne sings).
It's like any test - if you don't know what you are doing, it's easy to be led astray. Do you want to find out about what you are doing or remain in ignorance?
What do the results say?
Last edited:
That would be a false conclusion. The fact that something was heard in both sighted ad blind tests is not evidence that the sighted test was unbiased; it is merely evidence that the 'something' is actually present. It may have been 'heard' in the sighted test for some reason other than actually being present; as things which are not present can be 'heard' in sighted tests we cannot deduce that something being 'heard' in a sighted test is evidence of anything.mmerrill99 said:
To use a police analogy, sighted tests may be useful for detection but only unsighted tests produce evidence for conviction.
I try not to blindly accept anything. I leave validation of blind tests to those whose expertise is appropriate for this task.
Obvious differences (sighted) have a habit of disappearing when blind; in some cases it is known that the obvious difference is not actually present (e.g. seeing A and hearing B, and finding an obvious difference with seeing B and hearing B - some have tried to claim that under these circumstances B and B really are different). Naturally, people have two different explanations for this.
Audible means we can detect a difference by hearing. How we do that will vary from person to person, from test to test etc.; it does not require that we can identify and describe the difference. Whether we are detecting a difference is found by looking at the statistics.
What sort of controls would you use for a sighted test?
No. People interpreted this correctly to mean that no audible difference was heard. That is all ABX delivers. The sine wave test was an attempt to explore a possible reason why no audible difference was heard. Whether an audible difference could be heard in another test is a separate issue.
That would depend on what you think the result was. If the result was 'no difference was heard' then that is valid. If the result was 'no difference could be heard by anyone in any test' then that is invalid, but I am not aware of anyone making the latter claim.
As I understand it, the question being asked was 'can you detect a difference between 'this' and 'that', using your audio setup?'. I assume that most people doing the test would have expected that 'this' and 'that' are sufficiently different and their own setup sufficiently good that they might expect to detect a difference. "Can you tell original file from tube amp record" seems to me to be an adequate description of the intended test.
seems like fear to discover that one of the lies that's been sold to us and has introduced a bias in our thinking is the immortal "everyone hears differently!"
is there no commonality to what we can and can't hear?
i know what a pure sine tone sounds like i use them for servicing neither of the sine tone files sounded anything like that but i got positive results with the fruit files....?......where's my scotch...
is there no commonality to what we can and can't hear?
i know what a pure sine tone sounds like i use them for servicing neither of the sine tone files sounded anything like that but i got positive results with the fruit files....?......where's my scotch...
Yes, it's an anecdotal listening impression, I agree. It's the attempts to promote it above this level that I object to.
Obviously so. But there will be feedback from the system to tell you whether you/him (listener) are imagining things or not. The more the experience, the more honest he can be with himself.
The listeners will always 'underperform'. No doubt.
I don't know what 'ABX conditios' was. Is it 'more controlled condition'?
I believe that is what Jakob means & he also previously linked to research that showed ABX is significantly less sensitive a test for differentiating between samples that are audibly different (gets more wrong results - more false negatives) - other blind test protocols show this differentiation is audible more often.
Of course you could use DF96's twisted logic & claim that, in these other blind tests, they could be hearing something that doesn't actually exist - I'll leave you to ponder that logical flaw
Perception is a fickle thing, isn't it
- you see the expected patterns unless you pay close attention
I often keep wondering reading your posts. What are you talking about when saying "they turn into intro for the fruit files"? Check settings of your foobar, maybe you have crossfader or something like that activated. The sine files are only sines, amplitude looks like this (rising-constant-decaying). Takes a lot of time answering your strange questions and re-posting the files.
Attachments
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.