Because Jan, IF you do a 'well controlled' db test that will meet AES approval, you will NOT find any difference, and everything said here is a waste of time, if you 'believe' the results. That is unfortunate, but many hear what db tests do not show. That is the reason why we still discuss making better audio electronics.
There is no such goal like to persuade someone. Usually the people, who do not evaluate their assumptions (and sometimes results, if there are any) by listening tests do loudly call for DB tests.
There are several flaws: Methodology is poor. Usually some dirty switchboxes are used, that introduce unwanted HF noise. The compactness of original chain must not be corrupted. And listener is pushed to concentrate on presumed difference.
I have much better experience with different kind of test - I invite listeners separately. They write down their impressions, and they are not aware of the impressions of other listeners. I only care about the same sound level during tests.
I have much better experience with different kind of test - I invite listeners separately. They write down their impressions, and they are not aware of the impressions of other listeners. I only care about the same sound level during tests.
I know that Jan does listening tests- I even participated in one with him. He may have come to a different conclusion than you, but that doesn't mean that he didn't do the work it took to come to that informed opinion.
SY, everyone does listening tests, but a dB listening test that will pass AES approval will most likely not give any differences between almost all op amps, such as the 4558 and the 797. AND why pay 20 times more?
I have never seen any useful result from any double blind test, just confusion. That is why I am sceptical. But the people who believe in the DB are lucky - they can have any component and be happy with it, because there are no differences - are they? 😉
SY said:
Are you kidding??
Everyone of us actually tested many opamps. I made tests with e.g. AD797, OPA627, OPA602, OPA134, LT1122, LT1028, ...., AD844. Almost every visitor was able to describe sonic differences.
Blind A-B tests are OK.
I presume that 797's are overkill for audio, BUT there may be other applications out there, maybe instrumentation, like I use them, where the better design is useful and necessary. Is that what you mean?
I presume that 797's are overkill for audio, BUT there may be other applications out there, maybe instrumentation, like I use them, where the better design is useful and necessary. Is that what you mean?
PMA said:
<RANT>
And that is supposed to mean anything? If people are listening for differences then they will hear them, regardless of whether they really exist or not.
Go to any quack online herbal remedy or psychic healing/stick a magnet on your knob/hanging crystal BS store and read the customer testimonials for a similar effect.
Such profound mysteries of da oonnnniverse all boil down to the following mundane fact:
People in general are morons.
So there.
</RANT>
PMA, just so that I understand this, you have done listening tests that would be publishable by the AES which showed an audible difference between those op-amps?
Glen,
you are just expressing your opinion. Thank you, for me it is useless, and for me it indicates what is your real experience. You can listen what you want, it is your choice.
you are just expressing your opinion. Thank you, for me it is useless, and for me it indicates what is your real experience. You can listen what you want, it is your choice.
You've never given a reason to explain why double blind tests are invalid besides saying that they are because the results don't support your beliefs.
A DB test is by definition simply a listening test where the listener and the tester don't know which component they are hearing.
I fail to see how not knowing what brand the amplifier is will impair one's hearing, or how knowing will improve it.
Tim, I have been studying, debating, and experimenting with ABX testing since 1979, when I wrote my first rebuttal to Dr. Lipshitz in 'The Audio Amateur'. For me, it doesn't work, and IF you want to believe that it really works for electronics, such as discrete vs IC or one IC vs another IC, then you might as well give up trying to improve audio electronics. It isn't PROOF that is the problem, but that everything sounds the same, yet it doesn't when evaluated under other, perhaps more realistic, listening conditions. In other words: "If you can't tell the difference between live and Memorex" (cassette tape) stop worrying over high fidelity design. There are many who did this when ABX testing took hold. Now you have Dolby digital! Wow! Made by many engineers who gave up on better electronics. Others found alternative electronic challenges.
It is impossible to prove that a null result is really what it appears to be.
It is possible with this sort of testing to 'prove' that new coke tastes like old coke. Yes, it has been done! If you can't taste the difference, you might want to find the cheapest generic cola in the grocery store. Why pay more?
It is impossible to prove that a null result is really what it appears to be.
It is possible with this sort of testing to 'prove' that new coke tastes like old coke. Yes, it has been done! If you can't taste the difference, you might want to find the cheapest generic cola in the grocery store. Why pay more?
john curl said:
John,
I think this is a very important statement. It bares the gist of the issue. If our existence is in audio electronics, and somehow we would be forced to accept that in DB tests, where ONLY the sound matters, there is no audible difference between opamps, that goes directly against to our raison d'etre. We will NEVER accept that those things we slaved for during the last 30 years can not even be heard in a DB test.
So, being what we are, we will do anything in our power to discredit DB testing. Even while we officially subscribe to 'the scientific way'. Even if we would think someone crazy to take medicine that wasn't thouroughly DB tested, for example.
Look at PMA's post above where he says:
"There are several flaws: Methodology is poor. Usually some dirty switchboxes are used, that introduce unwanted HF noise. The compactness of original chain must not be corrupted. And listener is pushed to concentrate on presumed difference."
A prime example. PMA is not stupid, he knows quite well that there are good methodologies, that there are transparent switches (which probably are already in his equipment anyway), and that ALL listening tests, DB or open, require people to concentrate in some measure. Look at his use of "there ARE", as if it is certain and proven, which it isn't. Look at his use of "usually there are", also completely unfounded. The words 'corrupted', and 'presumed', forgetting in passing that it is HE who presumes, of course.
But he is an audio electronics designer, so the world MUST be that people can hear differences between his and other equipment. If not, life's efforts have been futile.
So, that is why DB testing is doomed for audio. People have incredible facilities to twist arguments, duck issues, outright lie and generally fool themselves bigtime. If your raison d'etre is at stake, there's no holding back.
Jan Didden
Jan,
this is just your interpretation. I am more interested in results than in reasoning. My goal is high quality sound reproduction. In case that me and my visitors can independently describe their impressions and these impressions quite fit, it is much more important for me than to argue here whether there is any difference between 4558 and 797. For those who assume there is not, it is their choice.
I rememeber very well your review, when you returned from your visit to Lumanauw, btw, what is your opinion now, after the time passed?
this is just your interpretation. I am more interested in results than in reasoning. My goal is high quality sound reproduction. In case that me and my visitors can independently describe their impressions and these impressions quite fit, it is much more important for me than to argue here whether there is any difference between 4558 and 797. For those who assume there is not, it is their choice.
I rememeber very well your review, when you returned from your visit to Lumanauw, btw, what is your opinion now, after the time passed?
PMA said:
A fair question. I did try to think back to it a little, and I am still convinced that I perceived a very large difference between the LP and the CD (I assume that is what you refer to). Was it a scientific test? No, it was informal, only one sample. I was also quite impressed with that very fine and expensive equipment. I realize that almost certainly was also a factor generally.
At the same time, if I would be able to repeat it DB and I would not hear a difference, I would feel very, very uncomfortable, I'm sure.
But there is no easy way out. We have ample proof, if we are willing to see it, that our perception of audible differences is determined by many more factors than the sound alone. We also have ample proof, if we are willing to see it, that there are precious few (I really don't know any, actually) DB tests between, say opamps, that convincingly show audible differences.
We need to accept that it is so, and we need to enjoy the fruits of our hobby. I enjoy it when I listen to my latest error correction amplifier and I hear those cristal clear voices, the almost perfect sound stage, that strong yet fast bass. I know that in a DB test it may be different. But so what? Trying to discredit DB testing on the grounds of twisted facts and tendentious arguments is cheating and takes away from our intelligence.
Jan Didden