@Salas
Completely irrelevant. People still spend millions on jewelery, ultra expensive clothes and thousands of other items that are not really necessary nor better in any meaningful way.
In any case, please stick to the subject.
What conditions have to be met in order to hear the differences among amplifiers ?
Or - what did all those tests mentioned/cited/published do wrong to invalidate their results ?
Be very careful, as there have been examples offered throughout this thread of people causally picking amplifiers at completely different levels, one-off hearing and likes.
Completely irrelevant. People still spend millions on jewelery, ultra expensive clothes and thousands of other items that are not really necessary nor better in any meaningful way.
In any case, please stick to the subject.
What conditions have to be met in order to hear the differences among amplifiers ?
Or - what did all those tests mentioned/cited/published do wrong to invalidate their results ?
Be very careful, as there have been examples offered throughout this thread of people causally picking amplifiers at completely different levels, one-off hearing and likes.
john curl said:
Not for me, no. Not for any person I have met or seen attempt to differentiate amplifiers without seeing them.
But that's what I've been arguing all along, no ?
<anectodal>
when I replaced my old amplifier with Krell, I invited all my audio friends for a listen. Superlatives galore, all those Pearsonesque adjectives abundant.
What they didn't know is that punny 25Wpc Yamaha was in fact what was playing all along ...
</anecdotal>
Since it's impossible to prove a negative (there is no difference) and no one can be forced to hear a difference who does not want to, this position is unassailable. But it is also unprovable since it is impossible to prove that a person hearing a difference is deluded and not hearing a difference. Therefore this position would be propounded by someone who has no agenda accept to be a PITA.
Sincerely,
Dang.
Sincerely,
Dang.
I am not mixing the luxury spender's fetishism with better amps here. And I am not looking for conditions. I make an amp with my soldering iron, and then I make another phase splitter for it and it sounds different. Same system, same levels, same guy. I cant possibly stay on subject, I just say that the subject is totally irrelevant to life. Nothing more to add.
Why is it impossible to prove that differences cannot be heard ?
I can easily prove that people can't tell difference between shades of color for example, when spectrometer measures tons of them.
Or differences in brightness of stars, when photometer measures again tons of differences. It is easy as ABX 😀
I can easily prove that people can't tell difference between shades of color for example, when spectrometer measures tons of them.
Or differences in brightness of stars, when photometer measures again tons of differences. It is easy as ABX 😀
Its easy, differences can not be heard under ABX, and data backs it up.
I don't argue with that. Wow, what a peculiar phenomenon! Another paradox. So what? But I will not sell my DIY KT88 on e-bay and hook up my 25yrs old Lux solid state in its place. I repaired it one year ago, gave it a go, it was dry like the Sahara, and is ok for my brother's TV output since then.
ABX is not life. Thats my point.
I don't argue with that. Wow, what a peculiar phenomenon! Another paradox. So what? But I will not sell my DIY KT88 on e-bay and hook up my 25yrs old Lux solid state in its place. I repaired it one year ago, gave it a go, it was dry like the Sahara, and is ok for my brother's TV output since then.
ABX is not life. Thats my point.
ABX is not life. Thats my point.
Well, that is exactly what we are trying to establish here.
What is that invalidates blind audio testing ?
Why a person can (with supposedly straight face) say that they can easily tell, "day and night", "laughable" etc. two amps when they can see them (just a casual glance, you know) at the same time can be left to (blindly) switch amplifiers till they turn blue in the face and still fail to tell them apart ? I mean, we are supposed to hear amplifiers, that is use our ears ?
Color me skeptic, but having to know which amp is connected invalidates much more the position of being able to hear the differences. It is like guessing what card is drawn from the deck - yup, I can always guess it correct, but I have to see it. If I don't see it, I can't guess correctly because I'm pressured and conditions aren't right.
Well, that is exactly what we are trying to establish here.
What is that invalidates blind audio testing ?
Why a person can (with supposedly straight face) say that they can easily tell, "day and night", "laughable" etc. two amps when they can see them (just a casual glance, you know) at the same time can be left to (blindly) switch amplifiers till they turn blue in the face and still fail to tell them apart ? I mean, we are supposed to hear amplifiers, that is use our ears ?
Color me skeptic, but having to know which amp is connected invalidates much more the position of being able to hear the differences. It is like guessing what card is drawn from the deck - yup, I can always guess it correct, but I have to see it. If I don't see it, I can't guess correctly because I'm pressured and conditions aren't right.
I was actually referring to scientific types rather than scientists. Of course a real scientist will accept a well documented test, regardless of whether it seems the subject under test has cheated by learning subtle differences in advance or that the test conditions have to be specific to each subject under test. I should say that I use the term "scientific types" somewhat pejoratively (I edited out the word "scientifical" from the first draft - I probably should have left it) referring to those who simply won't accept any premise based on anecdotal evidence and would probably not accept the results of a test constructed specifically to give every possible advantage to obtaining a favorable outcome (without resorting to cheating).
John
salas said:
Oddly enough, this could be exactly one of the cases where blind testing would work just fine.
Among the amps that I could hear the differences were many valve designs. Mostly 300B single ended variety, but some decently powered toobes as well. They would all sport a massive distortion in upper registers, curiously described by some as "air","bloom" and "sparkle".
Even with an A B test, we have to match level, frequency response, damping factor, and polarity. We have done this for MORE than 40 years. Where were many of you, 40 years ago?
We had those standards even then.
We had those standards even then.
But even when amps weren't matched by say damping factor or frequency response (case Futterman vs Pioneer) people failed to tell them apart ?
This should have made it easier to detect, no ?
This should have made it easier to detect, no ?
Hmmm ... problem I see with that is always knowing which amp is A. I can proceed to prefer A (deliberately or subconsciously, we can't tell). Thus it is not bulletproof.
A better test would be where random generator switches in either A or B every time you request the change. If you manage to identify them above statistical noise, good. If not, at the end of the test we check if you have played one of them for significantly longer periods, that would count as positive result as well.
A better test would be where random generator switches in either A or B every time you request the change. If you manage to identify them above statistical noise, good. If not, at the end of the test we check if you have played one of them for significantly longer periods, that would count as positive result as well.
Actually, thinking a bit more, I can't see how can ABX differ logically from A/B testing. I mean, you have the freedom to play A or B against X for as long as you want.
So let's say coin toss chose X=A.
Playing A against X (=A) you're not so sure you can tell the difference (no surprises). Which is fine. Don't record anything.
Then the next test will be B against X (=A). Now you should be able to tell the difference easily, no ? Remember, you can switch B and X as many times as you want, for as long as you feel like, on any material you pick.
How is this different from A/B ? I mean this is now A/B testing !
So let's say coin toss chose X=A.
Playing A against X (=A) you're not so sure you can tell the difference (no surprises). Which is fine. Don't record anything.
Then the next test will be B against X (=A). Now you should be able to tell the difference easily, no ? Remember, you can switch B and X as many times as you want, for as long as you feel like, on any material you pick.
How is this different from A/B ? I mean this is now A/B testing !
Bratislav, we have been thinking about this for 30 years. You are just starting. ABX doesn't work. Hundreds of tests have proven it. However, A or B CAN be arbitrarily designated, if you want. The main thing is that you pick A or B, but you don't try to say whether 'X' IS A or B. It works, I have used it for decades.
I leave it to you to note the difference, and why the brain can't seem to do something so obvious as ABX.
IF ABX worked, we would all embrace it. Why not?
Also, if everything really sounded the same, we would go on to making, cheaper, smaller, more complex designs, and not spend any more effort in making just 'another amplifier'. Do you have any idea how many power amps I have designed in the last 40 years? How many preamps in the last 35 years? How many studio boards, electronic crossovers, analog master recorders? I've been around. Yet I still don't know how to make a 'perfect' amplifier, any more than a winemaster knows how to make a perfect wine, or an automotive designer knows how to make a perfect automoble, etc. Perfection is made by both fortitude and luck.
I leave it to you to note the difference, and why the brain can't seem to do something so obvious as ABX.
IF ABX worked, we would all embrace it. Why not?
Also, if everything really sounded the same, we would go on to making, cheaper, smaller, more complex designs, and not spend any more effort in making just 'another amplifier'. Do you have any idea how many power amps I have designed in the last 40 years? How many preamps in the last 35 years? How many studio boards, electronic crossovers, analog master recorders? I've been around. Yet I still don't know how to make a 'perfect' amplifier, any more than a winemaster knows how to make a perfect wine, or an automotive designer knows how to make a perfect automoble, etc. Perfection is made by both fortitude and luck.
john curl said:
But, John, these are two different issues. One is preference, and that is something I for one won't argue with. Just as I don't want anyone else to argue with my preferences.
But the other issue is whether you can reliably, under controlled conditions, hear differences between pieces of equipment. That has nothing to do with preference. It even has nothing to do with music per se. In such a test, music becomes a test vehicle. In deciding preferences, music plays another role.
Jan Didden
John_Curl, i´d not agree that ABX doesn´t work in general, because at least some people report consistently detecting differences in ABX tests, for example Paul Frindle and Bruno Putzeys.
But, as i´ve noticed several times, the ABX protocol seems to confuse some listeners more than others, what leeds to the prediction that at least extended training should be done.
I´d strongly prefer tests for preference instead of discrimination.
But even if you do train your listeners, you never know what level of sensitivity is reached in the final test situation and that´s why it is so important to include controls.
The test janneman mentioned i.e. the famous Tiefenbrunn/Lipshitz ´desaster´ is a good example for that; no real control was incorporated, but if there was a click sound that remains undetected by Tiefenbrunn, but was finally noticed by Lipshitz, then you should take that as a hint that the sensitivity level reached wasn´t high enough.
A mechanism that was reported in AES meetings as well, see:
http://www.stereophile.com/asweseeit/894awsi/index1.html
As was pointed out before in this thread, this kind of test and the statistics involved can´t produce the thruth, but is a game of probabilities.
Postive test results without negative controls are as much doubtful as negative test results without positive controls.
Jakob
But, as i´ve noticed several times, the ABX protocol seems to confuse some listeners more than others, what leeds to the prediction that at least extended training should be done.
I´d strongly prefer tests for preference instead of discrimination.
But even if you do train your listeners, you never know what level of sensitivity is reached in the final test situation and that´s why it is so important to include controls.
The test janneman mentioned i.e. the famous Tiefenbrunn/Lipshitz ´desaster´ is a good example for that; no real control was incorporated, but if there was a click sound that remains undetected by Tiefenbrunn, but was finally noticed by Lipshitz, then you should take that as a hint that the sensitivity level reached wasn´t high enough.
A mechanism that was reported in AES meetings as well, see:
http://www.stereophile.com/asweseeit/894awsi/index1.html
As was pointed out before in this thread, this kind of test and the statistics involved can´t produce the thruth, but is a game of probabilities.
Postive test results without negative controls are as much doubtful as negative test results without positive controls.
Jakob
- Status
- Not open for further replies.