Slightly OT:
Attended a workshop yesterday by JC (Morrison) at the European Triode Festival. Did not agree with all he said, and I guess it was mutual.
But it was very worthwhile; this guys always starts me thinking about things I tend to take for granted.
He even got away with calling everyone who's into tubes 'losers', which of course included him.
Thank you JC (M)!
jan
Attended a workshop yesterday by JC (Morrison) at the European Triode Festival. Did not agree with all he said, and I guess it was mutual.
But it was very worthwhile; this guys always starts me thinking about things I tend to take for granted.
He even got away with calling everyone who's into tubes 'losers', which of course included him.
Thank you JC (M)!
jan
Almost certainly he does. I can give you the secret, but you have to promise not to tell.
1. Both amps have low distortion (<0.1%) and noise.
2. Neither amp is clipping.
3. Levels are matched.
4. Frequency responses are matched to within 0.1dB between 20Hz-20kHz.
5. Both amps have low source impedance so as to achieve #4 with the speaker load.
If you do those things, the million dollars is safe. Remember, mom's the word!
However, if you now compare (say) two different levels of data compression, any careful listener will walk away with the prize. That DBT stuff is insidious- it seems to allow the detection of all sorts of things except fashion.
Happy audio engineers, many of them do believe in the "secret" you have disclosed.
From my side, I could save 1 million bucks even more easy - I would put into system the speakers from famous canadian manufacturer I mentioned above. Even without exact observing your 1...5 points, nobody will distinguish two different amps.
On the other hand, if one would start with Makarov's system (the photo), and top level of recordings, any person from street will distinguish between any two amps.
Makarov himself, and many who visited him, refer to this system as to "audio measuring device", it brings to one's face any tiny differences.
Attachments
Last edited:
Vlad, you have to think 'outside the box' so to speak when it comes to ABX testing. Something else is involved, than just getting the test right (which often it is skewed toward NOT hearing differences). It has to be something in the way that our brain actually works. I think it is a 'left brain-right brain' thing. I have some 'evidence' to back this up, but it is not 'solid proof' and therefore would be dismissed here. I have to leave it at that, for the moment.
While the ABX-Protocol seems to work if listeners get used to the method, there is no reason to use the ABX-Test .
A/B discrimination tests are easier to implement and need (based on our experiences with quite some tests) less training to give useful results.
But negative and positive controls are still required.
Afair there were some examples where the careful listeners did not catch that something wrong was happening.
In fact, all the way back to the Lipshitz/Vanderkooy/Tiefenbrunn example, it was shown, that listeners missed something that could have been heard.
quoted from:
Soren Bech, Nick Zacharov
Perceptual Audio Evaluation - Theory, Method and Application
John Wiley & Sons Ltd., 2006
A/B discrimination tests are easier to implement and need (based on our experiences with quite some tests) less training to give useful results.
But negative and positive controls are still required.
SY said:
Afair there were some examples where the careful listeners did not catch that something wrong was happening.
In fact, all the way back to the Lipshitz/Vanderkooy/Tiefenbrunn example, it was shown, that listeners missed something that could have been heard.
quoted from:
Soren Bech, Nick Zacharov
Perceptual Audio Evaluation - Theory, Method and Application
John Wiley & Sons Ltd., 2006
Last edited:
Jacob2, you are correct about Ivor and his 'overconfidence' in thinking that he could 'beat' an ABX test run by Lipshitz and Vanderkooy. Of course, he couldn't, and that is why Dr. Rod Rees told me 20 years ago to AVOID being pulled into such test conditions. That is why I politely decline from ABX tests on this forum, I know 'better'.
IF I participated in such a test, I would NEVER hear the end of it, just like it was just brought up, after how many decades, since the test was run? 2 or is it 3?
IF I participated in such a test, I would NEVER hear the end of it, just like it was just brought up, after how many decades, since the test was run? 2 or is it 3?
Please tell me all those wooden thingies are part of the system. The "anyone off the street" type comments only serve to weaken your argument.
PMA is right we have come back again, this thread is both recursive and re-entrant.
No. The ONLY thing needed for any controlled listening test is trust in your ears. That's what most opponents to such tests here lack.
jan
Isn´t that a contradictio in ratio?
If we could trust in someone´s ears, then there would be no need for a controlled listening test.
If it has to be a "controlled" test then the control has to be comprehensive- not an easy task (see the quotes from Bech/Zacharov for a short summary of the reasons) .
I think it is perfectly logical. If you want to know whether some things sound different or not, AND you trust your ears, you should control the test such that only your ears give the inputs.
If you don't/can't trust your ears, you'd have to find all kinds of excuses why a controlled test 'doesn't work', including accusing people of dishonesty and 'rigging' the test.
Isn't that obvious?
jan
Yes, but there are all kinds reasons why blind testing doesn't work.
One only has to take into account how we hear, and many can't even agree on that.
Many butt head ENTs sure don't agree (or are just ignorant) about we hear. Same for neurologists who can believe anything they can't see on a stinking MRI, and dismiss or don't know about clinic evidence.
Too many people just dig into there respective corners and stick there head in the sand.
And we all keep have the same arguments over and over.
John
It's unremarkable that some people catch differences in sound more easily than other; no-one has ever claimed otherwise. For example, Lipshitz heard differences in his test that Tiefenbrun didn't. One notes that Tiefenbrun agreed before the blind portion that the differences were audible and that the test conditions were satisfactory. Once it was ears-only... errr....
It's also unremarkable that people can design a bad test; no-one has ever claimed otherwise. They can also design a good test. With my points attended to, no-one to date has demonstrated audibility of different boxes of gain.
I'm not sure I understand the argument you're trying to make.
Obviously the "control the test such that only...." part is the critical one.
Please remember all the numerous posts needed to get something obvious as "a positive control" (and already well known in the "other world" of scientific testing) in there.
If the experimenter shows with positive controls that the participants reach sufficient sensitivity under test conditions, then there is not much room left for "excuses" .
If otoh the experimenter finds all kinds of excuses why he don´t want to use positive controls then supects are reasonable because from a scientific point of view there exists no excuse for ommitting positive controls (and negative controls in general as well) .
I don´t know if it is a matter of dishonesty or "rigging" but if an experimenter does not care about the (possibly) confounding parameters, then it is (intentionally or not) simply bad science.
The quote from Bech/Zacharov is illuminating because it reflects the discussion back in ?1986? when Les Leventhal pusblished his comments about some statistical drawbacks of the "normal" 16 trial ABX commonly used.
The comments from Clark and Nousaine after that were a bit less focused on "finding the truth" than i´d have expected from objectivists .
Nobody ever claimed controlled tests are perfect and ideal.
But look at the alternative and there's really no competition.
Suppose you buy a couple 100 dollars worth of caps and cable, and you modify one channel of your amp, and you want to do a listening test to find out if it was worthwhile.
Do you realise the enormous mental stress your brain puts on you to find not just a difference but an improvement?
Under such circumstances, attempting a cool, relaxed, objective listening test is a battle that's lost before it begins.
jan
But look at the alternative and there's really no competition.
Suppose you buy a couple 100 dollars worth of caps and cable, and you modify one channel of your amp, and you want to do a listening test to find out if it was worthwhile.
Do you realise the enormous mental stress your brain puts on you to find not just a difference but an improvement?
Under such circumstances, attempting a cool, relaxed, objective listening test is a battle that's lost before it begins.
jan
- Status
- Not open for further replies.