Skip to main content ↑ ↓

Notice

Please note that most of the software linked on this forum is likely to be safe to use. If you are unsure, feel free to ask in the relevant topics, or send a private message to an administrator or moderator. To help curb the problems of false positives, or in the event that you do find actual malware, you can contribute through the article linked here.
Topic: Article: Why We Need Audiophiles (Read 501342 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

Article: Why We Need Audiophiles

Reply #525 –
The German c't magazine conducted a blind listening test (MP3 vs. Redbook) in 2000 with 12 participants. About 300 people had applied to take part and they mainly chose known audiophile advocates and people with audio related engineering professions. Most of them failed pretty badly, but 2 stood out of the crowd. One, Gernot von Schultzendorff, is mastering engineer at Deutsche Grammophon. The other one has a cutoff of 8 khz in his left ear after an explosion accident. He could hear the filter banks' flanging artifacts that otherwise would had been masked by higher frequency components.

Article: Why We Need Audiophiles

Reply #527 –
Quote from: Axon on
Quote from: krabapple on
This sound oh so reasonable, but if one were to actually tally the number and placement of the 'heterodox' writings, I'm rather sure one would find that by far most of them appear in the letters column, not the articles and reviews, and that that the rare peep of doubt about the Emperor's couture in a review is dwarfed by the amount of 'orthodox' content.


That said, IIRC, didn't JA reject Beltism?



'Old Earth' creationists scoff at 'Young Earth' creationists, too.   

Article: Why We Need Audiophiles

Reply #528 –
Quote from: Frumious B on
Quote from: krabapple on
Mr. Atkinson, this is a smokescreen, and one that Stereophile has been puffing out for far too long.  Yes, it would be an overreach to claim that there is never any difference tout court between A and B based on a small dataset (which is why 'objectivist' claims typically are qualified with words like 'likely').  But one can test *you* and *your* particular claim that you already hear a difference, rather more readily, and discover whether you were really hearing what you claimed to hear five minutes ago.


If we really want to pin these guys down then I think it would be better to sort of narrow down our challenges, present a unified front and a unified message and force them to either shit or get off the pot.


No one can 'force' anyone to do anything of the sort.  One is still free to be a raving audiophile in the US of A.  And free to  criticize them.
 
Quote
Otherwise they get to pick and choose which points to which they will respond from which people and generally obfuscate, evade and conflate all they like and then claim victory afterwards.


I'm actually glad people here have gotten a taste of Mr. Atkinson's talent for that, as well as Mr. Fremer's 'bad cop'.  Now you know what 'we're' up against because, this, folks is the face of 'authority' to a large segment of the high end.  Visit the Stereophile forum for a heapin' helpin' of crazy from the infantry, if you really want more.  (Wave to me and Axon and JJ as you drive by.  ;> )

Quote
I'd hate for that to happen.  Here is my proposal:  Atkinson is on record claiming that all lossy compressed music is unsuitable for "serious listening".  He is also now on record in stating that he uses lossless AND also AAC 320kbps on his own iPod.  Presumably this means there are tracks in his collection where he can very access both a lossy and lossless version of the same track without even doing a fresh rip.  That means he can take a few minutes to download Foobar while he is typing his next post and then he can either demonstrate his ability to ABX AAC at the highest quality setting from lossless or he can publicly retract his statement that lossy music is unsuitable for "serious listening".  I think that should be the challenge.  He either takes it or he folds his tents.


Heh.  Don't count on it.

Article: Why We Need Audiophiles

Reply #529 –
Quote from: gerwen on
Quote from: analogcorner on
I took the ABX test as devised and produced by the group and I got five of five identifications correct. My editor, John Atkinson got 4 of 5 correct. The average of all test takers was inconclusive. According to Dr. Stanley Lipschitz, I was a "lucky coin" and my result was tossed.


Maybe you were ambushed by the ABX'ers, but calling your positive results a 'lucky coin' is valid.

Maybe this has already been explained, but i haven't seen it in this thread.

Toss a coin 5 times, chance of it coming up heads five times is 5/32 or 3.1%, chance of 4 heads is 10/32 or 15.6%
Multiply those chances over 20 trials of 5 coin flips, and you will expect to see a few 4 head outcomes, and there's a good possibility of a 5 head outcome (around 60% i think, i'm no statistics whiz)



Actually, a 5/5 score has a p of 0.03...it's the first x/x score to break the .05 threshold.  But we don't settle for an n = 5 , do we?  Neither do scientists or statisticians, unless they're desperate.  Nor is it very consistent of JA  to cite a 5-trial test if he's also going to dun objectivists for small sample sizes.

As if that was our fault....by now, the high-end camp could have done and published hundreds of positive ABX results, if these differences were truly as apparent as they seem.  It has always struck me as remarkable that 'hi rez' was launched with *no* accompanying scientific listening data.

Article: Why We Need Audiophiles

Reply #530 –
Quote from: hybris on
As we all know, ABX only tells wether or not the individual were able to identify the tested amplifiers there and then.


If the individual has already asserted that ability to hear the difference, including 'there and then',  the performance 'there and then' can actually be quite telling about the *individual's* authority.  Stereophile reviewers seem to have little trouble routinely hearing differences there and then  -- at least based on what they write.  The interesting thing would be to test their 'ears'.

Quote
I'm about to loose track of my point here, but I think what I'm trying to say is that many of the issues we're discussing isn't really testable with ABX, so nagging on and on about making such a test doesn't really make any sense (because the differences are so small). Fremer and Atkinson knows this (while they may not be in a position to say it out loud), but they also know that even though it isn't ABXable doesn't really prove that there is no difference at all. And I see nothing much wrong with that.


What audible difference is so small as to be impossible to ABX?

Article: Why We Need Audiophiles

Reply #531 –
Quote from: krabapple on
What audible difference is so small as to be impossible to ABX?


I don't know, loudspeaker cables possibly? 

Is something audible if you can't ABX it? I guess it's a matter of definition

My main point (I found it now) is that you're asking them to do ABX(or similar tests), but we all know (including them probably) that the difference is too small for that to make any sense, as they'll fail. What that fact actually means is probably the main disagreement here.
Thorbjorn

Article: Why We Need Audiophiles

Reply #532 –
Quote from: krabapple on
Quote from: gerwen on
Quote from: analogcorner on
I took the ABX test as devised and produced by the group and I got five of five identifications correct. My editor, John Atkinson got 4 of 5 correct. The average of all test takers was inconclusive. According to Dr. Stanley Lipschitz, I was a "lucky coin" and my result was tossed.


Maybe you were ambushed by the ABX'ers, but calling your positive results a 'lucky coin' is valid.

Maybe this has already been explained, but i haven't seen it in this thread.

Toss a coin 5 times, chance of it coming up heads five times is 5/32 or 3.1%, chance of 4 heads is 10/32 or 15.6%
Multiply those chances over 20 trials of 5 coin flips, and you will expect to see a few 4 head outcomes, and there's a good possibility of a 5 head outcome (around 60% i think, i'm no statistics whiz)



Actually, a 5/5 score has a p of 0.03...it's the first x/x score to break the .05 threshold.  But we don't settle for an n = 5 , do we?  Neither do scientists or statisticians, unless they're desperate.


In the AES Convention tests that Michael and you are discussing, the number of trials (5) was set by the test organizer David Clark. In theory it was possible for conventiongoers to sign up for a second and third session of 5 trials, but the tests were over-subscribed as it was. So Michael and I had to remain "lucky coins" with Michael being a bit "luckier" than me. :-(

Quote
Nor is it very consistent of JA  to cite a 5-trial test if he's also going to dun objectivists for small sample sizes.


I argued at the time that 5 trials was too small, for the reasons mentioned, but that was the number decided upon, presumably to be able to get as many listeners through the listening test as was possible during the convention. The goal was to have a large number of listeners each doing a small number of trials, as I understood from the test organizers at the time.

Quote
It has always struck me as remarkable that 'hi rez' was launched with *no* accompanying scientific listening data.


Bob Katz, Vicky Melchior, JJ, and others have been discussing the design of formal listening tests comparing hi-rez audio with Red Book versions. Eliminating interfering variables turns out not to be a trivial matter.

John Atkinson
Editor, Stereophile

Article: Why We Need Audiophiles

Reply #533 –
Quote from: Stereoeditor on
I argued at the time that 5 trials was too small, for the reasons mentioned, but that was the number decided upon, presumably to be able to get as many listeners through the listening test as was possible during the convention. The goal was to have a large number of listeners each doing a small number of trials, as I understood from the test organizers at the time.
As I think I mentioned before to Michael, I agree wholeheartedly that the results do not support the conclusion that you and Michael were "lucky coins", and it is news to me that you two got 5/5 and 4/5. It's plainly unfortunate that more testing was not done.

At the same time, do you not agree that the overall result of the test - IIRC, that the results overall were fairly indistinguishable from chance (I don't have the paper on me right now) - still has a very important meaning? I am comfortable with the notion that a small fraction of listeners are able to tell a difference in such situations, while most listeners can't. And if you get close to a 50% result, while individual testers such as yourself can still pass a 16- or 32-trial test with flying colors, well..

Moreover, I think that such results can have considerable importance for those who cannot pass such ABX tests, and perhaps can readjust their purchase priorities accordingly. Of course this has to be balanced against the odds of the listener becoming more adept in the future to hearing such details.

So I'm really tempted to just straddle the fence here. That test did not show that you and Michael could hear the difference, because your results were as to be expected as due to chance. But they certainly did not show that you couldn't hear the difference, either - it is quite plausible - and more testing really should have been done on that matter. The test did strongly suggest that a majority of the testers could not tell a difference, and that in itself is important.

Put another way, the test was certainly "flawed" if it was attempting to show that absolutely nobody could hear a difference, but under a more relaxed criteria (of showing that the proportion of discriminators must be below some low percentage), I don't think it's flawed at all.
 

Article: Why We Need Audiophiles

Reply #534 –
Quote from: Stereoeditor on
Regarding your statement that "no true audiophile will listen to music while watching video from the same recording," which you appear to attribute to me, I have on many occasions written the opposite: that rather than SACD and DVD-A, it is the live music DVD that can perhaps be considered the true successor to the CD.

John Atkinson
Editor, Stereophile


Now, this conclusion surprises me. Isn't the audio on DVD lossy coded? AC3? DTS?

Quote from: Arnold B. Krueger link=msg=0 date=
BTW Stanley Lipshitz tells me that with suitable noise shaping, a 16 bit linear PCM system can have a subjectively-weighted dynamic range of > 120 dB, if memory serves.


He might have said that, but I don't agree. I've done a number of listening tests over the last three years (double-blind ones, of course), and I got a perceptual dynamic range gain of about 2 bit, i.e. 12 dB, when using noise shaping. That's roughly in line with Stuart's findings (see his AES papers from the mid 90s... sorry, I forgot the exact titles).

Edit: This holds for 44 and 48 kHz. True, at 88 kHz and beyond, you can gain 3 or 4 bits of additional perceputal dynamic range (also tried that).

Chris
If I don't reply to your reply, it means I agree with you.

Article: Why We Need Audiophiles

Reply #535 –
Quote from: C.R.Helmrich on
Now, this conclusion surprises me. Isn't the audio on DVD lossy coded? AC3? DTS?
Usually, but it doesn't have to be.

Many music titles have 2-channel 48kHz LPCM of various bitdepths (16, 20, 24) along with the video. Some even have LPCM with more than 2 channels - this leaves less bitrate for the video for each channel you add.

Cheers,
David.

Article: Why We Need Audiophiles

Reply #536 –
Quote from: Arnold B. Krueger on
Quote from: 2Bdecided on
If I have mistaken the meaning you intended, it is up to you to clarify.


No, if three 's any doubt about the meaning of what I said, its up to you to clarify before you proceed.

I call a golden ear a person who is a standout in a pre-selected group of people who have demonstrated far better than average sensitivity to audible differences.
If I have to assume that any word you use has been equally tortuously removed from its generally accepted meaning, it will make discussion very difficult.

Which, on the evidence I've seen on HA, is the general intention of your debating technique.

You do the cause of DBT a huge disservice. John Atkins comes across as perfectly reasonable; you come across as a complete twit (I misspelled that). If I had no prior knowledge, and could only judge this subject from the tone of this discussion, I'd be deleting my HA account and subscribing to Stereophile immediately.

I can't put it much more strongly - if this kind of debate and style spreads much further in HA, it'll kill it. Please take it back to Usenet.

Cheers,
David.

Article: Why We Need Audiophiles

Reply #537 –
Quote from: Stereoeditor on
Quote from: krabapple on


It has always struck me as remarkable that 'hi rez' was launched with *no* accompanying scientific listening data.


Bob Katz, Vicky Melchior, JJ, and others have been discussing the design of formal listening tests comparing hi-rez audio with Red Book versions. Eliminating interfering variables turns out not to be a trivial matter.


Indeed, yet isn't it rather after the fact?  SACD was launched how many years ago with how much hype about it's obvious superiority to CD?

(edit: to remove a wasteful amount of quoting by me)

Article: Why We Need Audiophiles

Reply #538 –
Quote
You do the cause of DBT a huge disservice. John Atkins comes across as perfectly reasonable; you come across as a complete twit (I misspelled that). If I had no prior knowledge, and could only judge this subject from the tone of this discussion, I'd be deleting my HA account and subscribing to Stereophile immediately.


As someone who really does have very little knowledge of the subject, I've got to second this. I mean, I totally understand being tired of making the same civil and informative arguments over and over and over (and over) again, but that is what being an advocate for something entails. If you (for whichever value of 'you' is preferred) would rather score points on the Great Electric Arguing Machine that is the internet (people think I'm joking when I call it this) than actually convince people of anything, well done I suppose.

As it is, I think I hate ABX tests on principle now, and will be making my future audio equipment purchasing decisions based on the number of people it will annoy. I wonder if there's a double-blind test for that?



Article: Why We Need Audiophiles

Reply #539 –
Quote from: 2Bdecided on
If I have to assume that any word you use has been equally tortuously removed from its generally accepted meaning, it will make discussion very difficult. Which, on the evidence I've seen on HA, is the general intention of your debating technique. You do the cause of DBT a huge disservice. John Atkins comes across as perfectly reasonable; you come across as a complete twit (I misspelled that). If I had no prior knowledge, and could only judge this subject from the tone of this discussion, I'd be deleting my HA account and subscribing to Stereophile immediately. I can't put it much more strongly - if this kind of debate and style spreads much further in HA, it'll kill it. Please take it back to Usenet.


Seconded. Strongly.

Arny, your cavalier attitude on the simple topic of communication is breathtaking. I have been nearly ostracized - literally - for making far less inflammatory accusations between groups of friends, than what you have accused Michael of. And then you have the nerve to respond not by actually backing your sh*t up, but by attacking Michael's semantics!?! I was willing to just drop the issue like everybody else did, but now that it seems like that's par for the course for you, I think I'll just go ahead and dig that back up.

I think I understand why I am (and Canar is) so flummoxed about this entire topic, going all the way back to when B0RK was the center of attention. You, and also Steven to a certain degree, have turned this thread into what seems like a word-for-word rematch of the exact same argument, made against the exact same people, that has been replayed on Usenet for what seems like a decade - an argument I have absolutely no intention of supporting. I said this to jj, I alluded to this to Michael, and I'll say it to you: HA IS NOT RAO. It isn't even RAHE. Some of us are quite willing to take John and Michael as being rational, intelligent, educated, experienced people - which is an assertion to some degree backed up by their rather good and coherent replies (particularly John's which I have always admired). I cannot say the same about your replies.

Let me put it this way, when your debating style is criticized as being too coarse on Stereo Central, something is deeply wrong.

Article: Why We Need Audiophiles

Reply #540 –
Quote from: Axon on
Quote from: Stereoeditor on
I argued at the time that 5 trials was too small, for the reasons mentioned, but that was the number decided upon, presumably to be able to get as many listeners through the listening test as was possible during the convention. The goal was to have a large number of listeners each doing a small number of trials, as I understood from the test organizers at the time.
As I think I mentioned before to Michael, I agree wholeheartedly that the results do not support the conclusion that you and Michael were "lucky coins", and it is news to me that you two got 5/5. It's plainly unfortunate that more testing was not done.

At the same time, do you not agree that the overall result of the test - IIRC, that the results overall were fairly indistinguishable from chance (I don't have the paper on me right now) - still has a very important meaning?


My take on it is that when very small  but real differences are involved, either the design of the test becomes non-trivial, or any individual identifications disappear into the overall noise from the results of those who do not perceive a difference. You need to repeat thr test to determine what is the problem, and that may not be possible, as at the AES Convention tests we have been discussing.

When I organized an amplifier test at our 1989 Show in San Mateo, I could tell the amplifiers apart under blind conditions -- see http://www.stereophile.com/features/113 for a description of the methodology. (This shouldn't be surprising as the amplifiers had response differences into the test loudspeaker and I had done so many dry runs that I think it fair to claim that I had learned the difference.) But when the results of all the listeners were calculated and adjusted for the inadvertent mismatch between Sames and Differents, we were back to the old null result.

John Atkinson
Editor, Stereophile

Article: Why We Need Audiophiles

Reply #541 –
@Axon :
I am getting confused here.

I just read this post :
Quote from: Axon on
Quote from: krabapple on
IN his 'real life' Fremer was/is a psychiatrist, and in in NYC that can pay pretty well.
Fremer is a psychiatrist? That joke writes itself. Multiple times over actually. Heh.

Quote from: krabapple on
Indeed. Fremer's pricey system didn't restore his ability to hear the LP hiss that the reporter heard.
Nor does it guarentee that his LPs play back with a speed tolerance of any less than 0.6%, as I observed a few days ago with some needledrops he posted.

Next time you hear an audiophile claim that high-mass turntables do not have speed issues, pour that into their cornflakes and shove it up their ass.


Your last post is playing quite a different tune ... correct ?

Article: Why We Need Audiophiles

Reply #542 –
I wasn't intending to contribute but I would add that too much of this comes across as personal grudges and mudslinging. Informative logical argument, most importantly with some civility would be welcomed by the rest of us who are interested in the topics concerned.

HydrogenAudio is wonderfully free of flame wars and their like and rich in information and genuine enquiry and knowledge-seeking. I can understand a degree of personal animosity after many years of conflict, perhaps people feeling they're hounded wherever they go, having their integrity and honesty called into question despite the fact that they're right. However, I'm fairly confident that the rest of us here would like civil discussion unpolluted by personal attacks and "he would say that because..." which are irrelevant to the technical points being discussed and worsen the high signal to noise ratio of this forum.

Having the humility and intellectual honesty to acknowledge the limitation of your own past results and not claim more that they've actually demonstrated, despite your immense contributions to the field, and your obvious intellectual rigour in those contributions, is the kind of behaviour that engenders great respect when sustained for some time, such that statements you do make can be trusted more straightforwardly that having to see if they're too sweeping The effective assertion that no golden-ears exist generally, runs counter to codec-testing experience of people like Guruboolez of which we've seen so much evidence here and in the French hardware forums where he's also reported his detailed listening tests, so this would seem too sweeping and needs qualifying, as you did after some queries.

I'm definitely outside the industry, but scientifically trained as a physicist, and I appreciate the gracious tone of John Atkinson, even if I know little of Stereophile and though I have a general scepticism and a propensity to dismiss as placebo most of the sighted reviews and recommendations unless I see statistically significant evidence or gross measurement differences to support the views expressed.

Anyhow, I'm diluting the information content enough, so I'll stop there.
Dynamic – the artist formerly known as DickD

Article: Why We Need Audiophiles

Reply #543 –
You mean I'm communicating more politely to people to their faces? Shocker.

Read up on the intermediate threads between that post and my last one. I changed my tune only a couple posts after that one, actually. I remain exasperated that a high-end turntable can have such wide speed excursions, but that's a significantly different issue than what the discussion has evolved into (or, at least, what I'd like the discussion to evolve into): a more nuanced discussion on the nature of blind testing, along with a larger debate on the role of high-end audio plays with respect to mainstream audio. And really, given my recent posting history on the philosophy of science and blind testing in other threads, this should not be a surprise.

Article: Why We Need Audiophiles

Reply #544 –
Quote from: Stereoeditor on
My take on it is that when very small  but real differences are involved, either the design of the test becomes non-trivial, or any individual identifications disappear into the overall noise from the results of those who do not perceive a difference. You ned to rpeat teh test to determine what is the problem, and that may not be possible, as at the AES Convention tests we have been discussing.

When I organized an amplifier test at our 1989 Show in San Mateo, I could tell the amplifiers apart under blind conditions -- see http://www.stereophile.com/features/113 for a description of the methodology. (This shouldn't be surprising as the amplifiers had response differences into the test loudspeaker and I had done so many dry runs that I think it fair to claim that I had learned the difference.) But when the results of all the listeners were calculated and adjusted for the inadvertent mismatch between Sames and Differents, we were back to the old null result.

Fair enough.

However, wouldn't it also be fair to say that, if testing an effect that is self-reported by the testers to definitely not be "very small", that they should be taken at their word? That is, for differences like high res vs Red Book, or before/after demag, when the testers universally agree in sighted listening that the difference is obvious, that the proportion of discriminators should be somewhere very close to 1, and so the type II error is actually very low?

That's one of my biggest peeves about this whole debate. We are not talking about effects that are supposed to be "very small". There are reviews on any number of things I can dig up, where the reviewer claims night-and-day differences... and yet DBTs on the issue yield negatives. IMHO, it's only after the DBTs are taken into account that the words "very small" creep into the discussion. It's largely ad hoc.

Frankly, if sighted testing identifies a clear difference that can be consistently identified, but a DBT in the same listening environment yields a null result, it doesn't really matter whether or not the difference was very small to begin with. The sighted testing impressions are critically challenged all the same.

Article: Why We Need Audiophiles

Reply #545 –
Mr. Atkinson,
I have twice put versions of this question to you, but it was apparently lost in the shuffle.  Your published position on the use of lossy compression is as follows:  "MP3s and their lossy-compressed ilk do not offer sufficient audio quality for serious music listening. This is not true of lossless-compressed formats such as FLAC, ALC, and WMA lossless" 

If an individual listener is unable to differentiate a lossy file, either under the conditions of an ABX test or in a more relaxed and conventional listening setting, from its lossless counterpart then in what way, precisely, does the lossy file "not offer sufficient audio quality for serious music listening"?

Article: Why We Need Audiophiles

Reply #546 –
Quote from: 2Bdecided on
Quote from: Arnold B. Krueger on
Quote from: 2Bdecided on
If I have mistaken the meaning you intended, it is up to you to clarify.


No, if three 's any doubt about the meaning of what I said, its up to you to clarify before you proceed.

I call a golden ear a person who is a standout in a pre-selected group of people who have demonstrated far better than average sensitivity to audible differences.


If I have to assume that any word you use has been equally tortuously removed from its generally accepted meaning, it will make discussion very difficult.


No torture intended. Your problem is one of context. You apparently don't understand the mindset of people who have been doing DBTs for 30 years or so. There's no reason why you should, but when you are communicating with people who have been doing DBTs for such a long time, you might consider making some allowances.

BTW, I didn't invent the usage - David Clark did back in early 1990s.

Quote
Which, on the evidence I've seen on HA, is the general intention of your debating technique.


Hmm, my debating technique?

I don't have a debating technique. I just write about my experiences and my thoughts.  BTW, the RAO usage is "Debating trade".

Quote
You do the cause of DBT a huge disservice. John Atkins comes across as perfectly reasonable; you come across as a complete twit (I misspelled that).


John Atkinson strikes you as being perfectly reasonable?  I'm beginning to think that I'm in Hell, complete with sign over the entrance saying "abandon all hope ye who enter in". ;-)

So what's your next move David, run right out and buy a $20,000+ LP demagnetizer, as recommended by your good buddy Mr. "Perfectly reasonable" John  Atkinson? ;-)

Quote
If I had no prior knowledge, and could only judge this subject from the tone of this discussion, I'd be deleting my HA account and subscribing to Stereophile immediately.


I didn't set the tone of this thread all by myself. Far from it.  So you are willing to give these guys free passes for all of their childish name-calling? I logged on one day and found myself described as a"total liar". The worst thing I've said about them is that they are on occasion poorly-behaved and poorly-informed. I've even done their footnotes for them. ;-)

Quote
I can't put it much more strongly - if this kind of debate and style spreads much further in HA, it'll kill it. Please take it back to Usenet.


You know David, this post of yours seems like it is fresh off of RAO, childish name-calling and all. I think you need to start referring to me as "Krooger" so that the RAO effect will be complete. ;-)

Article: Why We Need Audiophiles

Reply #547 –
Quote from: ff123 on
So what I think I just read is that Arnie claims that if there is a pre-selected group (of supposedly sensitive listeners), who have been trained to hear certain types of audio deficiencies, and their hearing has not been damaged (either by trauma or age), then it has been his experience that their standout results regress to the mean in subsequent testing.


Yes, with the caveat that ABX is not a test that all by itself naturally interates towards finding individual thresholds.  IOW if you test amplifiers whose distortion is inaudible to everybody, you aren't obtaining different results for the sensitivity of all of the listeners. None of them can hear the difference, they are all random guessing, and that is that.

If you have a bunch of people who are guessing randomly, then if you do enough tests their results will eventually all converge to the (same) mean. If anybody is a standout, the outstanding results are really a statistical fluke, and statistical flukes have this nice way of converging to the mean as you do more and more tests.


Quote
That certainly is an interesting claim, and not one I would intuitively guess to be true.


It wasn't intuitive to us up front, but when we thought about it and talked about it for a few weeks, we figured it out.

We had a similar experience with the problem of test results that seem to be far *worse* than random guessing. Lots of thinking and talking and we figured it out.  It turns out that results that seem to be worse than guessing are explained when there was some communication among the listeners, and the number of actual trials is less than the apparent number of  trials because of the inter-communication.  Joe and Sam are somehow communicationg and reporting the same thing, so there is actually one less independent listener than your head count.



Quote
I would personally expect, for example, that as typically is the case with human attributes, that there is a bell curve of sensitivity.


I don't know about that. Listening is essentially athletics, so athletic metaphors can work. Is there a bell curve for people running the mile that extends below the current world's record by however many sigma?


Quote
At some point, let's arbitrarily call it 2 sigma, we would have a few people whose sensitivity is so far from the norm, that we could call them "golden ears," for all practical purposes.


In fact the thresholds of hearing various things often seem to be quite abrupt.  Lots of results just above the threshold, but none below.


Article: Why We Need Audiophiles

Reply #548 –
I will delurk to third the sentiment that HA regulars (Mr. Krueger in particular) championing the cause have been sanctimonious and a general embarrassment.
While I do not buy what they are selling Mr Atkinson has been by far the better debater and I wish other more reasonable and level headed members if HA would step in to continue this.

Article: Why We Need Audiophiles

Reply #549 –
Quote from: Dynamic on
The effective assertion that no golden-ears exist generally, runs counter to codec-testing experience of people like Guruboolez of which we've seen so much evidence here and in the French hardware forums where he's also reported his detailed listening tests, so this would seem too sweeping and needs qualifying, as you did after some queries.


Not everything is a codec. Codecs seem to be very different from things like amplifiers and CD players. In the hardware world we have seriously inaudible things like 0.1 dB FR and level shifts, and noise floors that are 30 dB below the noise in the program material. Not the same as ABXing 128 or maybe 320 kbps codecs.

I don't have a lot of experience with testing codecs. I do just about everything including personal listening with pure .wav files and redbook CDs.

Testing real-world hardware, particularly that of the floobydust kind, seems to be an especially different world than testing codecs. 

We've just never seemed to find any people at all who could reliably hear the benefits of demagnetizing vinyl records, for example. ;-)

So, the concept of varying levels of sensitivity were irrelevant for many of our tests.