 
|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
|
|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
83.248.95.107
In Reply to: RE: To be entirely clear the Retraction I mention should come from you. posted by John Atkinson on December 17, 2007 at 13:56:49
...the research in which Lipshitz has "confirmed that a reversal of absolute signal polarity will be subtly audible on music to a 99% confidence limit!"?
The one whose results I've linked -- yes or no?
In other words, your claim above is in error -- yes or no?
I note that, quite as expected, you prefer to be evasive and say nothing of either of these in your reply.
- Or do you know of another, secret study by Lipshitz hitherto unknown to the rest of us? (Open in New Window)
Follow Ups:
> Would you now also care to point out ...the research in which Lipshitz
> has "confirmed that a reversal of absolute signal polarity will be
> subtly audible on music to a 99% confidence limit!"?
I suspect that I misread the original article's conclusions as referring
to the results on music _and_ test tones combined as referring to music
only. As I said, the BAS Speaker article is in storage and I haven't yet
checked their website to see if it has bene reprinted yet. So while my
original statement read "on music to a 99% confidence limit," you are
welcome to read it as saying "on music and special signals to a 99%
confidence limit."
Ansd you still appear to be confusing the actual score in a blind test
with the statistical confidence in that score. The two are not the same,
as a little thought on your part will show.
Let's say I take part in a blind test and score 6 identifications out of
10 correct. The 60% score is not statistically significant, as the chance
of my achieving that score by chance is high. Now, say, I repeat the test
with 1000 individual trials and score the same 60% identification, ie, I
get 600 correct. Statistically, this is now highly significant
identification, even though the score in percentage terms is the same. I
don't have time to calculate it right now, but the chance of my achieving
a score of 600 out of 1000 are way less than 1%, I would say.
So, same score but different statistical significance. Do you see? (And
why am I, the subjectivist writer "in hock to the snake-oil salesmen,"
according to a deleted post, having to point this basic math out to you
the believer in science?)
John Atkinson
Editor, Stereophile
I take this as a Yes, indeed your description of Lipshitz's results was factually incorrect and erroneously characterized. And that therefore, as you have stated in the past, you will publish a correction in your magazine's online edition, as you have claimed here on AA that willingness to do this is what distinguishes Stereophile from many of its competitors.Now, just one curious point: Are you seriously in claiming that, in the case of Lipshitz, results indicating audibility on music and results indicating audibility on sound material consisting of a mix of both music and special signals are basically interchangeable and the same thing?
I tell you what (why am I, a casual reader of your texts, having to point the significance of this difference to you, an industry pro who frequently does and uses this kind of research): On music, Lipshitz' score was 60/113; on test tones, it was 24/24.
That's a big difference: 53% correct vs. 100% correct.
May I point out that it seems to mean two things, each with a very different meaning for us as music listeners: Polarity reversal is audible if you listen to test tones. Polarity reversal is not audible if you listen to music.
You will therefore immediately understand that this makes the two cases anything but equivalent. Or can you not understand why this is so? Let me put it differently, just to be sure: 60/113 or 53% correct responses (on music alone) is different from 84/137 or 61% correct responses (on a music and test tones combined) and both are very far from "confirming the audibility" of polarity reversal in any case.
You see how suspicions of intentional misleading of readership can creep to mind?
I am well aware of what "confidence level" is about and so you don't need to keep diverting more attention to my unfortunate conflation of the two different figures above as already explained. Thank you. I simply remain wondering why would you include a mention of a confidence level but not of the actual test score, as this can be very confusing (given it's not always about probability)? You see, again...
> I am well aware of what "confidence level" is about...
My apologies. Given that you had made the same error of confusing actual
test scores with statistical probability in several postings, it was a
reasonable conclusion for me to reach, that you didn't understand the difference.
> I simply remain wondering why would you include a mention of a confidence
> level but not of the actual test score, as this can be very confusing
> (given it's not always about probability)?
In my own work, I always give both actual scores and the confidence level.
Given that the citation I made of Stanley's work was in a footnote, I
thought it sufficennt to give the reliability, as that is by far the more
important result of any blind test. And it's _always_ about probability.
> On music, Lipshitz' score was 60/113; on test tones, it was 24/24.
> That's a big difference: 53% correct vs. 100% correct.
Correct. What it means is a) that the effect of inverting absolute
polarity with test tones is detectable under blind conditions to a
highly significant degree (24/24), and b) that if it is audible with
typical music, the effect is subtle, as both Dan Shanefield and I have
stated (me in the footnote to which you object) and that with a score of
60/113, this test does not provide evidence that the effect is audible
with a statistical degree of confidence greater than [x]%. (Again, I
don't have time to calculate the signficance right now, but maybe someone
else will.)
> Polarity reversal is audible if you listen to test tones. Polarity
> reversal is not audible if you listen to music.
No, that is incorrect. You cannot derive an absolute negative result
from the results of a test such as this. You really do need to read a
textbook on the statistical design of experiments. I use a well-thumbed
copy of Neville and Kennedy's "Basic Statistical Methods for Engineers
and Scentists" (InterText) that I bought when I was doing my bachelor's
degree nearly 40 years ago. Which statistics textbook do you use?
John Atkinson
Editor, Stereophile
Re: confidence levels: It's OK, not a biggie.Re: the actual score of the Lipshitz-Vanderkooy test:
(1) Why on earth would you omit the actual test score and only mention the confidence level, if your purpose was not to mislead the reader? Especially as you were making the big claim that what was established was audibility quite specifically " on music " (and not on test tones)? Besides, this claim of yours that polarity is "audible...on music" was not in a footnote but in the body text of your "As We See It" column in which you developed your whole claim about the importance of the polarity issue.
(2) Yes, thanks for admitting Lipshitz was correct in his test results that polarity reversal, if at all, is at most subtly audible with music. But you didn't state this in your Stereophile column we are talking about; quite the contrary, you state that it is "audible...on music."
(3) Yes, my statement was truly simplistic to make the above point clear, which I am happy to see you now in agreement with. Everyone understands that here we are only talking about the Lipshitz-Vanderkooy experiments and not of a definite proof positive of anything other than what was specifically shown for this particular set of questions & with this particular test setup. No need to belittle, hence; I think anyone who bothers to read exchanges like ours knows well enough what experiments can and what they can't do with hypotheses and what the status of the latter is. So no problem, we can easily restate if you wish: "In this experiment polarity reversal was shown to be audible when the specific test tone designed for the purpose was used but not shown to be audible when musical excerpts were used." Better now?
Lastly, I don't think I have objected to any footnotes. What I have objected to, again, is:
- Your claim in the Sphile "As We See It" column that polarity reversal is audible on music (it is not, you now admit); and
- you claim in the same text that Lipshitz and Vanderkooy have demonstrated this audibility on music (they have not, you now admit); and
- your claim still in the same text that such audibility on music is indeed basically the one thing that has been reliably detected in DBTs (this has not been detected in DBTs, you now admit).
Since you have now, de facto if not de jure, retracted all three of these false claims, I will feel satisfied to see your correction also printed in Stereophile as you have promised to do by now several times already.
I for my part will qualify my earlier claim that you were intentionally misleading the reader from the start, with an addendum that it may all have been due in part to your being confused about terminology and interpretation of the test results, as it now seems to me. But it would much help make a more generous assessment of the facts possible if there was a readier acknowledgment of errors (and willingness to set the record straight) in evidence in lieu of what we now see above in the tortured twists and turns of this thread that can only make it all seem like so politically motivated instead.
TL
1) Do you agree that the following is taken from the referenced material that Atkinson cited in the 1988 article?
"The authors have demonstrated the two-tone experiment described above to numerous people on different systems. No one has ever failed to hear the timbral change with phase, and discern the polarity reversal on this signal with unvarying accuracy. Indeed, in a double-blind demonstration to eleven members of the SMWTMS audio group [13], the accuracy score was 100% on the summed 200-Hz and 400-Hz tones over loudspeakers, and overall, including musical excerpts, the results on the audibility of the polarity inversion of both loudspeaker channels were 84 correct responses out of 137, this representing confidence of more than 99% in the thesis that acoustic polarity reversal is audible."
...........YES or NO?
2) Do you see any mention in that text of a score of 60/113 on music?
...........YES or NO?
3) Do you see any mention in that text of a score of 24/24 on test tones?
...........YES or NO?
The answers to those question should indicate whether you're playing with a full deck. Frankly I don't hold out much hope at this point that you are... But hey!, surpirse me!
Everything matters, don't forget to tweak your placebos!
1) YES.2) NO. (Guess why this was omitted by JA?)
3) NO. (Guess why this was omitted by JA?)
I suppose in your eagerness you are not realizing that this was my whole point: misleading by omitting the key information.
To begin with, you may note that "overall, including musical excerpts" (i.e., the 24/24 test-tone score and 60/113 music score summed up) is not at all equivalent to (same as) Atkinson's claim that these results were obtained "on music." If you didn't get it yet, this is the whole crux of the matter (the very thing that makes it a distortion or misrepresentation of facts on the part of JA/Stereophile).
So thanks for your help, but I'm afraid JA may not be patting you on your back any more at this point.
By the way, also this very text was discussed heatedly for about two months without a break already last spring (over at Critics), with also your participation IIRC, so all those following the discussions on this subject will find little use for your constant reiterations of the same.
- In other words, he decided to omit all the critical details (see the same scores in this breakdown). (Open in New Window)
"I suppose in your eagerness you are not realizing that this was my whole point: misleading by omitting the key information."
I can only conclude that you're implying that back in 1988 John Atkinson chose source materials that he knowingly knew to be "misleading". Surely you have something to substantiate such a seemingly fantastic notion?... I mean WOW!
.....
"To begin with, you may note that "overall, including musical excerpts" (i.e., the 24/24 test-tone score and 60/113 music score summed up) is not at all equivalent to (same as) Atkinson's claim that these results were obtained "on music." If you didn't get it yet, this is the whole crux of the matter (the very thing that makes it a distortion or misrepresentation of facts on the part of JA/Stereophile)."
Oh I "get it", no question about that, but your case depends upon the proposition that back in 1988 Atkinson was in possession of the knowledge (results on music vs. on test tones) and moreover was "misleading by omitting the key information".
I again invite you to provide something that would substiantiate such a claim, please do!
Everything matters, don't forget to tweak your placebos!
...one shouldn't try to use for one's own ends research that one is not familiar with.
This guy was really into statistical process control.
In his cubicle, he had a poster with a picture of a drunk leaning against a street light. The caption of the poster was something like this:
"Statistics are often used as a drunk uses a lamppost - for support, rather than illumination".
I always liked that one :-).
Though in my profession they are mostly used for the formulation of new hypotheses and development of fresh platforms for further action, not so much for validation of existing positions.
TL
"one shouldn't try to use for one's own ends research that one is not familiar with" appears a step down from "misleading by omitting the key information".
That being the case allow me to ask are you formally retracting the "misleading by omitting the key information" statement, or do you stand by it?
If you stand by it then I repeat I would be interested in seeing your evidence to justify the claim.
Thank you.
Everything matters, don't forget to tweak your placebos!
(1.) Key information : Audibility of polarity reversal on music has never been demonstrated in any DBTs in general and in the Lipshitz-Vanderkooy experiments in particular.(2.) John Atkinson in Stereophile : "Work by Stanley Lipshitz...confirmed that a reversal of absolute signal polarity will be subtly audible on music to a 99% confidence limit!" (Boldface mine.)
(3.) John Atkinson in Stereophile : "Indeed, it [audibility of polarity reversal on music] is one of the few things that can be reliably detected with double-blind testing ." (Boldface mine.)
(4.) Discussion : The key information given in (1.) is omitted from statements by JA in both (2.) and (3.) (and elsewhere in the Stereophile article in question); in fact, what JA states in (2.) and (3.) is something diametrically opposed to known facts ("key information").
(5.) Conclusions : Draw your own.
For evidence, try for a change to ACTUALLY TAKE A LOOK at any of the articles and reports discussing the Lipshitz-Vanderkooy experiments and review any and all reports on DBTs related to polarity audibility on music. If interested in further detail on the Lipshitz-Vanderkooy test, see link (first two rows of the table give the score first for test tones and then for music). You should also re-read the passage from the Lipshitz JAES piece abstract that you've quoted yourself above in this thread.
...I haven't yet checked their [BAS] website to see if it has bene reprinted yet.
I checked, and unfortunately it's not there.
True, but a letter to the editor from Dr Shanefield discussing the results in The BAS Speaker, Volume 17 number 3 contains the following interesting note. Bolding mine.
"The first has to do with a detailed study of the audi-bility of various phase effects as reported by Stan Lip-shitz and co-workers in theJournal of the Audio Engi-neering Society(vol. 30 no. 9). Johnsen cites a statementon page 583 of this report as evidence for the audibilityof absolute polarity. He certainly is correct that absolutepolarity has sometimes been (barely) audible in double-blind tests.
But Johnson failed to mention another critical point. The Summary section of the Lipshitz et al report (onpage 593) says that the audibility of polarity changes inmusic played over loudspeakers is only "extremely subtle." This evaluation is confirmed in an exchange ofcomments between me and Lipshitz et al published intheJ.A.E.S. vol.31 no. 6, in which we all agreed that formusic on loudspeakers absolute polarity is (1) notstrongly audible to most people, (2) not usually audibleabove 500Hz (more so below), (3) possibly not as audibleover loudspeakers which are comparatively free ofasymmetry effects (such as second harmonic distortion).In these letters I said that all kinds of phase effects were"of negligible importance," and Lipshitz et al said at thevery end, "We are basically in agreement with Dr.Shanefield."
The mystery deepens.
That's an interesting article. Thanks for posting it. I've been hoping to find an article that goes into detail about the actual data and procedures of the tests that Lipshitz referred to in his phase distortion article. So far, I've only found the cryptic info at the ABX site. The connection between the ABX site data and the BAS test unfortunately seems to be circumstantial - though the chances of the numbers matching up exactly with Lipshitz's by chance seem pretty remote. Also, Krueger has written Audio Amateur articles with Muller, so it's not unreasonable to assume he had access to the data.
As far as Lipshitz's subjective comments, I have tremendous respect for the guy, but his subjective opinions on audibility don't really mean any more or less to me than those of any other sane audiophile :-).
I have no beef with the absolute polarity crowd either. As I see it, audibility of absolute polarity has been proven beyond a shadow of a doubt on test signals. Audibility is audibility. I suspect that somewhere, someone can find a recording of an instrument that produces a very asymmetrical waveform, and for which care has been taken in the recording process, and that such a recording could be used to establish audibility with music. Hey, after all a synthesizer is nothing but a signal generator anyway. I see this situation as being similar to the tests used to determine people's ability to detect changes in level. People don't do well on this test with music at all, but do much better with pink noise. It's the most sensitive tests that determine the facts of the matter IMO. If those tests don't use music, then so be it. That said, I find some of the claims of certain absolute polarity fanatics overstated.
I agree with most of that, save that raw voice has plenty of asymmetry. By the time producers or broadcasters are done with it little remains though. We should remember the ABX site is mostly irrelevant to the question of honesty in Atkinson's article, written years before the WWW was public. The BAS article is what counts.
...that polarity reversal is audible ON MUSIC (to a 99% confidence limit or whatever)?IT DOESN'T.
Lipshitz et al. couldn't show that at all.
Yet that is what JA claims, and that is where he is misleading and in error. And to try and claim like him that "on music" and "on material combining test tones and musical excerpts" are one and the same thing is obviously nothing but an attempt at plain old BS as I am sure you will agree. I'll post a separate note somewhere here describing exactly why this is so, maybe in reply to JA or bjh who seem most ignorant of the (statistical) meaning of the difference.
"...the ABX site is mostly irrelevant to the question of honesty in Atkinson's article, written years before the WWW was public..."
I'm not sure if you are saying that JA probably didn't have the data on the experiments that he was trying to publicly draw upon to bolster his own claims about the audibility of polarity? That would be either extremely sloppy or extremely cynical of him. The ABX site has no monopoly on this test data; it's just one of the many sources out there that have bothered to record the results for posterity. The details were obviously available from multiple sources, including the authors of whom JA writes most warmly, i.e., he could have always checked with them, too, if he felt unsure for some reason even after perusing all the written reports.
TL
Sit down. Breath through your nose. The question was simple, does the BAS article - or for that matter the Lipshitz paper - Atkinson references clearly break down the tone and music results and make the difference explicit, or do they report the confusing combined confidence level? If the latter, by what right do you accuse Atkinson of the things you have? By your response though you appear to have the BAS article and can settle this quickly with quotes.
On the larger question of audibility of polarity, you contradict what the BAS letter to the editor says of Lipshitz's thinking, substituting 'not aubible' for 'subtle'.
I wasn't shouting, I just didn't have time for html codes.Your simple question has a simple answer: Lipshitz [OK, I'll go for the codes :-)] clearly states that overall , including the scores for both the test tone and the musical excerpts (this is all directly lifted from the Lipshitz JAES piece that JA himself cited -- see Andy_C or bjh above for the exact quote), the results were 84 out of 137 correct responses (representing 99% confidence level, but see meaning of this term; and which 84/137 result could already be questioned as to its significance).
But the question here is: How did JA turn all this into a simple, unqualified statement proclaiming that audibility to be there "on music" (with even an exclamation mark to heigthen the significance of his assertion!)??? There was certainly enough information there in his own references to know better.
Even if we assume your mildest-case scenario as the most accurate one (JA simply was unable to understand statistics and/or terminology here, and so erred without aiming to mislead, which we of course do not know as he refuses to clarify the issue), what he could and should do is to correct the misinformation/error after it has been brought to his attention.
This has not been done, even though in public (also here on AA) he has presented Stereophile proudly as one of the few accurate & reliable publications out there that actually issues corrections of factual erros. Only this isn't happening, at least not with this particular issue. What are we going to make out of that, then?
I may have the BASS issue but will have to look it up in storage and I can't do that now. If I find it, I'll let you know, but what do you expect to find in it? If you cite a test in support of your position, you better ensure that this test indeed supports your position. It's no one else's responsibility but your own. And it's too bad if you aren't capable of deciphering research reports or did a sloppy job reading them, but then you shouldn't have tried and used them in the first place. It's that simple really.
I don't care about what Stereophile or John Atkinson does, but I don't like hypocrisy and arrogance especially when they combine, and I don't like it when people keep denying undisputed facts for ideological reasons, and I don't like pointless objections for objections' sake when I point to the existence of such facts. Hence my posts here, if you want to know. Nothing personal. I hope with you, too.
TL
I'm not sure what counts at this point - LOL!
At the bottom of the ABX page with the absolute polarity data, it says:
The tests of absolute polarity were done on the original Quad electrostatic speakers. They were used in hope of minimizing the known potential problem of speaker distortion revealing polarity. The audibility of the Stan Lipshitz / John Van der Kooy polarity training signal was confirmed on those speakers. However the audibility of polarity on music was not confirmed whether the sound source was vinyl or a 1/4" 2-track master tape I recorded of a soprano recital with piano accompaniment. Stan and John came to Detriot for the first pair of tests. The system included a Hartley 24" subwoofer, Leach amplfiers, an ARC SP3-1, and a Kilmanas modified Rabco turntable.
The information above, and the total trial and success numbers matching Lipshitz's numbers exactly seem like too much of a coincidence. Also, this text appears on the ABX page:
May 7, 1977 SMWTMS did the first ever audio double blind subjective listening tests. An argument over the audibility of differences between amplifiers at a club meeting in November 1976 resulted in an agreement that a double blind test could settle the question. Just six months later, Arny Krueger gave a lecture on his design of a double blind comparator and the first three double blind tests were done. The results include the first three listed in the Power Amplifier Comparison Table in the data. Thus we credit Arny Krueger and his opponent in the argument, Bern Muller, as the inventors of the ABX Comparator. The agreement to create a company to manufacture comparators was informally made the following summer.
Same SMWTMS that Lipshitz mentions. Also, Bern Muller was the author of the Audio Amateur article that Lipshitz references. I don't think this had anything to do with the BAS at all.
When Lipshitz in his AES article is writing that:"The authors have demonstrated the two-tone experiment described above to numerous people on different systems. No one has ever failed to hear the timbral change with phase [TL: the 24/24 score with test tones], and discern the polarity reversal on this signal with unvarying accuracy. Indeed, in a double-blind demonstration to eleven members of the SMWTMS audio group, the accuracy score was 100% on the summed 200-Hz and 400-Hz tones over loudspeakers [TL: again, that 24/24 test-tone score], and overall, including musical excerpts, the results on the audibility of the polarity inversion of both loudspeaker channels were 84 correct responses out of 137, this representing confidence of more than 99% in the thesis that acoustic polarity reversal is audible."
...he is indeed describing the very same test whose results are given in greater detail on the ABX comparator site. Compare the numbers; on the other hand, the 11 test participants are described more in the Audio Amateur article (though are of course also noted on the ABX comparator site). Lipshitz isn't famous for clarity and he is clearly struggling to avoid direct confrontation with the fact that with musical material the score was no better than 60/113.
The prob with the "cryptic" way of presenting the info on the ABX comparator site is that their system for doing so is actually a bit too basic. The line "Van der Kooy / Lipshitz Training Signal" gives the score obtained with the test tone (it was specifically designed for this purpose by the test authors); the line "Sheffield & Other Phonograph Records" gives the score obtained when music, not the test tone, was used ("phonograph records" were utilized as this was after all still the pre-digital era...). The last two lines report on further tests for which neither Lipshitz nor Vanderkooy were in attendance (not addressed by any of the articles discussing their "own" test, then). The columns giving the p value and the number of participants in each test round are self-evident.
In a word, all of these sources that have been mentioned talk about the the same test results (the ones whose closer numerical breakdown is given on the ABX comparator site but only more generally described in the papers referred to).
TL
I just wish the ABX site would have put some footnote on the data, saying, "This test was performed by XYZ and is further documented in the article ABC". That would take care of my "cryptic" and "circumstantial" concerns.
I just realized that Krueger posts to usenet regularly, so I'll bet it would be possible to get some clarification from him, to verify that these are indeed the data in question. I have very little doubt, but it might convince (ha ha :-)) some of the doubters.
Instad, they just put the numbers up, possibly including some circumstantial-looking sentence or two about who came from where, a line header obvious to those only who were present during the experiments, and that's it. They could really function as a highly useful data bank if they did something along the lines of what you suggest.
I'd be curious to hear if you get any comments out of Krueger, so I'll keep my ears wide open, just in case!
Cheers,
TL
I just sent an email to the webmaster, David Carlstrom. I asked about the origin of the data, and whether I could have his permission to post his reply. We'll see what happens.
No. of trials, no. of test runs with test tones, no. of test runs with musical excerpts, no. of participants, the description and history of the specifically devised test tone, the test dates, the p values, and the names of the test authors are all identical. The chances that at some point the two authors replicated this experiment in a totally identical format with regard to every single one of its above parameters (incl. date...backdate!!) is: zero.TL
I'm completely convinced at this point, but I get the impression that there's still some diehards holding out.
My email to David Carlstrom bounced, so I'll post a message to usenet. I'm sure Arny Krueger won't be able to resist responding :-).
See link: "An experiment involving polarity inversion of both loudspeaker channels resulted in an audibility confidence rating in excess of 99% with the two-component tone, although the effect was very subtle on music and speech."
I'm not sure where the mystery resides. It's not that this research by Lipshitz & Vanderkooy isn't known; it's simply that the papers discussing it only very seldom give the detailed breakdown of the results (score with music vs. score with the specifical test tone designed by Lipshitz to maximally reveal polarity shift). Usually the results are only verbally characterized -- except in the ABX comparator website which collects all DBT results done using an ABX comparator. So the mystery results are there in their full glory and in plain olde English.
TL
You do understand the difference between "not audible" (your description) and "very subtle," don't you?
> You do understand the difference between "not audible" (your description)
> and "very subtle," don't you?
I don't think tlyyra does grasp that "subtle" distinction, any more
than he does the concept of statistical confidence limits. :-)
John Atkinson
Editor, Stereophile
But would you care sharing your view of exactly what that 99 % confidence limit has to do with the audibility of polarity reversal on music and where you obtained this figure?
.
.
.
I think not.
.
.
.
But it would be actually interesting to hear you answer some of these questions, instead of trying to sidetrack the issues with efforts to undermine the credibility of the questioner's character.;-)
- I actually agree with your own conclusions. Or do you mean to say you are now backtracking from your admission of error? (Open in New Window)
> would you care sharing your view of exactly what that 99% confidence
> limit has to do with the audibility of polarity reversal on music and
> where you obtained this figure?
From the reference cited, and confirmed in a link you yourself gave,
tlyyra: "An experiment involving polarity inversion of both loudspeaker
channels resulted in an audibility confidence rating in excess of 99%
with the two-component tone, although the effect was very subtle on music
and speech."
Seems clear enough. And I will ask again about your own familiarity with
the statistical analysis of experimental results, as you really don't
seem familiar with the concept of applying a confidence limit. The 99%
limit you asked me to explain in another posting means that there was 1%
or less probablility that the identification in the test results was the
result of chance.
By contrast, you seem to want to deal with a black and white world,
where an effect is either audible or not. That is not how the results of
tests like these can be discussed; what is important is the _probability_
of the test results being due to a real effect rather than to chance.
With something like absolute polarity, the effect is possibly audible
only some of the time and not to all listeners with all kinds of music
program (particularly as with multimiked programs, you can have random
mixings of inverted and correct absolute polarity). As both Stanley said,
and I said back in 1988, the effect is subtle. You need, therefore, to
design both the experiment and the statistical analysis of the results
with those facts in mind.
Look, rather than you keep dredging around the facts of what happened
and what was written two decades ago, test the matter for yourself.
Stereophile's first Test CD, which I produced, includes an absolute
polarity test track (#8) with which you can test yourself blind.
The track, which is purist miked, is decribed at the link below. (The key
can be found at www.stereophile.com/musicrecordings/176/index14.html.)
Email me your street address and I'll send a copy of the CD.
> it would be actually interesting to hear you answer some of these
> questions, instead of trying to sidetrack the issues with efforts to
> undermine the credibility of the questioner's character.
What efforts? I haven't made any mention of my opinions of other
posters' characters.
John Atkinson
Editor, Stereophile
OK. Let's go into this a little deeper, then, as you still seem to prefer to avoid addressing the points raised.(1.) Where did you get the result that polarity reversal is audibile on music to a 99% confidence limit?
The latter applies to the overall results only: test tones and musical excerpts combined -- not music alone as you claim.
The quotation you refer to above is by a grad student at U Miami. Why don't you cite Lipshitz himself? Or any of the two sources you cited in your original Sphile article?
But let's stay with this "99% confidence limit" a little that you misattributed to the results on music (and that gives a misleading impression that polarity reversal is somehow audible on music):
The results obtained by Lipshitz & Vanderkooy were: 24/24 correct responses on test tones, and 60/113 correct responses on music. So we get the following as regards:
(i) Audibility on Test Tones :
24/24 has a 2*2^(-24) = 2^(-23) chance of occurring randomly (we should include 0/24, since that would be considered equally significant, which is why it's multiplied by 2) -- which is less than 1/1,000,000. So this result is highly significant, which no one denies: "Audible on test tones."
(ii) Audibility on Music :
On the other hand, the 60/113 score (or more) on music has about a 29% chance of occurring randomly, and if we use the two-tailed significance we must again multiply it by 2 -- so it has a greater than 50% chance of happening by guessing. In other words, no significance: "Not audible on music."
(iii) Audibility Overall :
Combining both data sets gives the score of 84/137, with a two-tailed significance of almost exactly 99%. Obviously, all the significance here comes from the test-tones results, as anyone can immediately see from the above.
(iv) Therefore it is indeed possible to say that overall , polarity reversal is audible with 99% confidence; and that it's audible on test tones with better than .000001 confidence; BUT it is NOT possible to say (like you do) that it is audible on music with any confidence at all .
So we can only conclude that your claim is totally false (either a misrepresentation or simply erroneous, I'm not sure as you have preferred not to clarify this).
In other words, the effect, as has been summarized by many already, if audible at all, is certainly "very subtle," given that no one thus far has been able to detect it in double blind tests on music (contrary to what you claim; but please feel free to cite evidence showing otherwise). Including Stan Lipshitz.
To repeat : You claim that polarity reversal is audible on music to a 99% confidence limit ! (Exlamation mark and italics yours.) And that Stan Lipshitz has demonstrated this in his experiments discussed above. And that the audibility of polarity reversal on music is indeed one of the few things that can be reliably detected with double-blind testing (Italics yours).
None of these three claims are true, as we can all clearly see: There are no DBTs that show this audibility on music. The confidence level you cite was obtained by Lipshitz not for music but for overall score combining also test tones. Lipshitz himself couldn't establish any audibility on music, either. And if no one has shown it with DBTs, it obviously cannot be detected with them with any kind of reliability, no matter how flexibly you yourself might prefer to use the concept.
Seems clear enough? Or still uncomfortably black-and-white for your taste?
(2.) If it seems people like I keep "dredging around the facts" for too long (we actually like facts), why not put the matter to rest? You know what to do, since you've proposed it yourself: Edit that Sphile article to reflect the actual facts of that Lipshitz experiment, and issue a correction of this erroneous claim in Sphile as you've already said at least two or three times that you will. End of story.
(3.) As for the Sphile test CD, I have it and have run a few series of little experiments precisely with that polarity track. I am afraid my result may disappoint you (and I can assure you it's not because of my audio system is inadequate).
(4.) And as for those "efforts to undermine the questioner's character" that you deny, I obviously referred to your post in response to Robert Young in which, instead of addressing any matters of substance, you instead decided to try and raise doubts about my general competence and credibility in precisely those areas in which you now seem to be somewhat lacking yourself, if I may say so, judging from the above.
TL
You don't need any help from me, or from JA, as you seem to be doing a smashing job of this all by yourself. It is your insistance that JA has willfully distorted the facts for some nefarious purpose that constitutes the significant character-questioning in this thread.
I have asked if he would like to explain or correct a factual error, which very much seems like a misrepresentation of facts and in any case amounts to a false claim, whether accidental or intentional.
You did indeed say some of that!
You specifically charged Atkinson with being "misleading by omitting the key information" and those two "Guess why this was omitted by JA?" s (same post) leave very little doubt.
I'll remind you I asked asked you to provide evidence to back such a claim, and that you have not produced any such! So where's the beef?
Everything matters, don't forget to tweak your placebos!
"Misleading by omitting information" is something completely different than what Robert Young is asserting.Then again maybe not in your book. It seems a bit different as a stylistic guide anyway.
The evidence? That he has mislead by omitting information? And even made claims that contradict the facts in his own sources as well? Is reading comprehension the area where you feel challenged, or the numbers part? I just yesterday put it in a grammar-school type of a nutshell for you. That's as pared-down as I can get.
I suggest a more thorough investigation into the meaning of "misleading by omitting information." You seem to be alone in not understanding it.
Here's a hint: "misleading" is not accidental.
I think by now we are all similarly prepared to draw our own.
However, on the point of grammar, if I may: The verb "to mislead" in itself does not necessarily imply the presence of conscious intention. Think, say, of a "misleading sentence": sentences do not possess either consciousness or intentionality (the latter in particular is a fundamentally human* quality, as we have known since the early phenomenologists at the very latest). But it is true that often this expression connotes deliberateness on the part of the one that misleads.
In the absence of any clarification from JA on this particular point, we simply do not know with certainty to which extent his actions may or may not have been intentional and deliberate, so usage of a term that leaves the question at least little bit open seems appropriate.
*Today we might want to say "humanoid" or something even less restrictive.
Your example - the "misleading sentence" - is not a strong one (as there is still an active subject that creates the condition of being "misleading" - someone had to write your "misleading sentence"), but your point that the word "misleading" is not specific regarding intent is true. However, put in the context of your posts, it is clear that "misleading" is used as a synonym for "delusive."
Regarding the lack of clarification from JA, given the absence of any proof of deception, we must presume innocence before guilt.
That all you got left?
My how the mighty have fallen.
Language is at the base of understanding and communication. That you continually abuse it means that any substantive additional effort spent with you is wasted time.
This is why I prefer to post plain facts such as numbers:
60 / 113 = 53%
I'm still waiting Atkinson to do the same.
'nuff said.
Can you describe for me the difference between "confidence level" and "percentage correct?"
Thank you.
It's beside the point.But feel free to re-read my reply to JA on this. You seem to have posted a reply to it so I would have assumed you read it already. Maybe JA could help with the unpacking of the stats content? ;-)
- Discussed already, so what is your aim if not just arguing for the sake of arguing? (Open in New Window)
You have been obsessively focussed on claiming a consciously deceptive math error, but you can't acknowledge that you have substituted terms. Your position is so ludicrous that you can't even maintian a consistency of attack.
You really should have given up ages ago. Now you'll have to go much further, just to get the moderators to ban you. That way you will have constructed your own vindication.
Robert Young said to tlyyra:
> You have been obsessively focussed on claiming a consciously deceptive
> math error, but you can't acknowledge that you have substituted terms.
> Your position is so ludicrous that you can't even maintian a consistency
> of attack.
I am not sure what else I can do, Robert? I have acknowledged that I may
have made an error -- this was 20 years ago, for goodness sake -- and I
have amended the text in the offending article in Stereophile's Web
archive so that it is now literally correct. Yet the man is still
demanding I answer his questions! :-(
John Atkinson
Editor, Stereophile
wasn't that it?
LOL
Everything matters, don't forget to tweak your placebos!
But I don't mind you keeping the thread alive.
That's for you, bjh."I have amended the text...so that it is now literally correct" :
Are you kidding? Your text reads:
"Work by Stanley Lipshitz in the late '70s (footnote 9), using carefully organized double-blind testing, confirmed that a reversal of absolute signal polarity will be subtly audible on music [and definitely audible on test signals] to a 99% confidence limit! (Indeed, it is one of the few things that can be reliably detected with double-blind testing.)"
I am not sure if you can understand this, but you have done nothing to correct the three patently false claims you make. It's literally incorrect on the same three points as before:
- Stan Lipshitz never confirmed audibility of polarity inversion on music with his DBTs.
- Stan Lipshitz never confirmed audibility of polarity inversion on music to a 99% confidence limit.
- None of this has ever been confirmed in any other DBTs, either.It's not even very good cosmetics. I'm sure your personal audience will buy it, but do you think anyone else will?
"I am not sure what else I can do, Robert?...the man is still
demanding I answer his questions!" :Poor you, people pestering you with requests that you kindly not misrepresent someone else's research results. Maybe you should avoid appearances in public, then, if that's too inconvenient.
Are you trying to make it look like you are a victim of a witch hunt by these terrible AA fact dredgers or something?
I'm reminded of the old proverb, "Don't wrestle with pigs: you're both going to get drty, but the pig likes it."
> Regarding the lack of clarification from JA, given the absence of any
> proof of deception, we must presume innocence before guilt.
I think that you forget, Robert, that as a member of the Evil High-End
Audio Establishment, my guilt is to be assumed until I am proved
innocent. :-)
To be honest, I don't have a sufficently good memory to be able to
perfectly recall what I was thinking on the day I wrote that essay 20
years ago. But it is possible that I conflated the 1978 Lipshitz test
results on music and test tones. When I can retrieve the specific issue
of The BAS speaker from storage, I will check out the Lipshitz article.
In the meantime, perhaps I should be gratified that people are studying
things I wrote decades ago with such intensity. :-)
John Atkinson
Editor, Stereophile
We are and have been talking about publishing misleading information on a topic of vital interest (in the audio world)......a false statement
....an error of fact
...an extremely serious "error" in that, it seems, as it got another poster already banned and deleted from AA for bringing it up.
To be honest, you and I both know that this error is easily detected at the first reading of your essay, if one is familiar with Lipshitz' experiments (as you claim to be yourself, citing it a number of times in your essay).
No need to go back to scratching memory or digging archives (a well-known wait-it-out manoeuvre). Just read what you yourself wrote. So plain even I can see it... ;-)
But what you probably should do is consult someone reliable on statistical analysis. See for instance KlausR's remarks above, if our mutual exchanges so far have failed to convince you.
TL
TSP didn't get banned for "bringing it up." He got banned for making slanderous remarks without backing them up. Actually, if not for your cute little backtracking on the meaning of the word "misrepresent," you wouldn't be far off from him either. But your agenda, and your meaning of what you think JA did is extremely obvious to everyone who reads this Forum. "An extremely serious "error..."" is how you put it, and that's coming from someone who stil doesn't understand (or just refuses to) understand the difference between "confidence level" and percentage correct.
Now think just a bit harder: if TSP had been banned for "bringing it up," then you would have been banned as well.
Did Atkinson's Stereophile column misrepresent Lipshitz' research or not?That's a rhetorical question since I know you won't be answering it. Doing so would confirm what we all know by now: that TSP got banned & deleted for pointing out that which seems obvious.
What did you say about "abuse of language" just a minute ago?
As for why I'm still allowed to post, since you ask, my best guesses for now are:
1) Deleting my account would make things a bit too obvious, since TSP just got silenced for the very same reason of pointing false claims advanced by Atkinson. The whole forum would appear as nothing more than a tool of audiophile industry, in open contradiction with what it claims itself to be (an open forum for debate). So it would be better to at least let things to quiet down first.
2) But I think it's perhaps more likely that the moderators themselves understand how they screwed things up and, even when reluctant to admit it and reverse their actions, are unwilling to repeat their mistakes, at least at this point and in public.
3) We're probably not done yet.
(see link below)
Tlyyra: I don't wish to alarm you, but I fear that as we speak, agents from the high end audio establishment are plotting against you to do unspeakable things. Perhaps you ought not to say anymore about "The Lipshitz Fiasco". I mean only if you value your life and freedom, of course. I know people who have disappeared in seconds, as if they never existed. And for a lot less. In fact, I've already said too much....
Pande
I only ventured an answer to a specific question put to me by that poster.
Besides, I spoke only of this site, not the entire industry... ;-)
Tlyyra, I wasn't being "dramatic", I was just being humorous. I'm sorry that that also wasn't obvious to you, but I can't say I'm surprised. Anyway, the point I'd like to make is, there really aren't agents of the high end audio establishment that have been provoked by your prime truthseeking mission and which are on high alert and intensely monitoring your activities right now. So rest assured, you are free to speak your mind against the powerful forces of HE audio, to expose the gritty secrets of "The Lipshitz Affair" as you have been making so many efforts to do, and are under no danger at any present time.
p.s. If you ever had any goals in life that you hoped to achieve before you shuffle off the old mortal coil, I would strongly suggest you do so within the next 48 -- no, make that 24 hrs. Give or take. Good luck! (thumbs up) :-)
Pande
in case you missed the smiley face.
That's what I said in my earlier reply to you, too, which curiuosly enough seems to have been deleted in the meantime.
TL
Hi,
I think it's great that you're still alive and posting (against all odds). Especially since you just made me $50 bucks on that fact. If you can keep this up, I figure that by the time you get to the end of the downward spiral in your "vindicating TSP campaign", I'll have won enough to buy 10 Brilliant Pebbles, 9 Clever Clocks, 8 Tru-Tone Covers, 7 Shakti On-Lines, 6 Hallographs, 5 tuning dots, 4 Silver Foils, 3 Myrtle Wood Blocks 2 Bybee Filters and a Partridge Family greatest hits album.
"That's what I said in my earlier reply to you, too, which curiuosly enough seems to have been deleted in the meantime."
Well, be thankful it was the only thing that got rubbed out. Thus far.
Pande
"Well, be thankful it was the only thing that got rubbed out. Thus far. "It wasn't, at all. Hope I didn't make you lose that $50 by telling this.
Talk about trolls.
completely delusional.
TSP didn't get banned for the reasons you claim, and you know it. At this point, you continue to make a sham out of intelligent debate.
You are totally wrong but I am not going to say why and on what point! You just are! And you know it! I know it for sure! Totally delusional! But of course no evidence is necessary! As with everything else! That's what I say! Now go away and leave us intelligently debating!
There you go with your "royal 'we'" again. I see no one coming to your assistance on this one.
Your agenda has been laid bare.
Of the approach practiced by you.And your kind.
It's alarming that you don't see the irrationality in it.
But not surprising.
What earned the response, Tlyyra DEAR is the breadth of your delusion.
one "dumb----" on that other (now departed) dumb----! You only get one you know, at least that was my understanding the last time I examined the Privileges Section of the Member Rights document distributed to contributors to the Influence Peddling Slush Fund for AA Moderators .
Shucks!
ps.
Please don't remind me it should be "dumb ----" or "dumb-----", I've put up with about as much annoyance as I can bear over the past few days. Thanks for your cooperation.
Everything matters, don't forget to tweak your placebos!
> [Your] Assumed Guilt"? Are you trying to become an Audio Asylum martyr?
No, I was trying to inject some humor into the thread. My apologies if
you were confused.
John Atkinson
Editor, Stereophile
Gotcha, sorry.
Perhaps that's the best way to proceed for you at this point. :-)
TL
It would also be interesting to see what's in the Audio Amateur article that was the primary reference for these tests that Lipshitz quoted in his paper - if you have it. Bern Muller was apparently a guy from SMWTMS that was in on the tests. Here's the Lipshitz quote again for reference:
"The authors have demonstrated the two-tone experiment described above to numerous people on different systems. No one has ever failed to hear the timbral change with phase, and discern the polarity reversal on this signal with unvarying accuracy. Indeed, in a double-blind demonstration to eleven members of the SMWTMS audio group [13], the accuracy score was 100% on the summed 200-Hz and 400-Hz tones over loudspeakers, and overall, including musical excerpts, the results on the audibility of the polarity inversion of both loudspeaker channels were 84 correct responses out of 137, this representing confidence of more than 99% in the thesis that acoustic polarity reversal is audible."
And reference 13 is this:
[13] B. F. Muller, "Third World: The Scientific Subjectivists," Audio Amateur, vol. 11, p. 64 (1980 Jan.).
> reference 13 is this: [13] B. F. Muller, "Third World: The Scientific
> Subjectivists," Audio Amateur, vol. 11, p. 64 (1980 Jan.).
Like the BAS Speakers, my AAs are all in storage. But Old Colony has
published a complete AA archive on CD-ROM. I'll get hold of a copy.
John Atkinson
Editor, Stereophile
...over this formulation of a grad student the way John Atkinson now agrees to put it, closely following the conclusions by the test author Lipshitz himself:" If it [reversal of absolute polarity] is audible with typical music, the effect is subtle...and this test does not provide evidence that the effect is audible with a statistical degree of confidence greater than [x]%." (My emphasis.)
If I may fill in that x for JA in his last sentence, the chances that a 60/113 result is obtained by just guessing are greater than 50% (about 58%), meaning that the score has no significance at all. His own point proven: not shown to be audible.
In other words, if audible at all, the effect is so subtle it cannot even be heard (in the tests conducted so far, in case JA disagrees with this formulation...).
TL
"In other words, if audible at all, the effect is so subtle it cannot even be heard (in the tests conducted so far, in case JA disagrees with this formulation...).
"
Well according to this thesis it IS audible (if subtly) on music.
Not per Lipshitz, no longer per Atkinson either, not per any of the tests/research I've seen.
Perhaps per clarkjohnsen but that's another story...
TL
"See link: "An experiment involving polarity inversion of both loudspeaker channels resulted in an audibility confidence rating in excess of 99% with the two-component tone, although the effect was very subtle on music and speech.""
That's in your post, T. And you link to the text from which it is drawn at the bottom of your post.
You can't have your cake and eat it too.
Seems to me he can, and does, again, and again, and....
Perhaps he was spoiled as a child?
Everything matters, don't forget to tweak your placebos!
That's a grad student summing up the Lipshitz experiments. Did you not manage to read it?
I myself, like I said, prefer the way both Atkinson now and Lipshitz always have put it: if it's there at all, the effect is so subtle it cannot even be claimed to be audible. That's a bit more precise way of saying it than "very subtle."
TL
"That's a grad student summing up the Lipshitz experiments. Did you not manage to read it?"
That is not what that thesis is about...otherwise its not really a thesis now is it? What is it with you and Lipshitz? Read this quote from the introduction:
"This thesis research is not meant to be exhaustive by any means, but to ascertain some permissible levels based on careful experimental design and analysis. The Kwalwasser-Dykema Music Tests format will be used in the implementation of the thesis research with other considerations. These permissible levels may be beneficial in the design and application of audio equipment, especially in the area of transducer and loudspeaker system engineering.
"
Where did you get that this thesis is all about Lipshitz?? It is research in the same area but it is not regurgitation of his work.
From the conclusion:
"A psychoacoustic experiment and analysis, proposed to ascertain permissible levels of phase distortion in various audio signals, addressed the significance of the audibility of phase distortion."
further;
"These permissible levels may be beneficial in the design and application of audio equipment. It is concluded that phase distortion is an extremely subtle effect, complicated by reverberant listening or original recording conditions"
"The audibility of phase distortion in audio signals seems to be highly dependent upon individual ability"
Gee, where have we heard that before? You guys are simply not willing to accept that people have significantly different hearing abilities, some hear these anamolies better than others.
"The selection of the all-pass center frequency of 3.5 kHz where the Robinson-Dadson curve displays the lowest threshold of audibility may suggest that for broad-band signals (such as the impulse), mid-range phase distortion is most audible."
"Improved irradiation methods, such as the use of phase-equalized loudspeakers in an anechoic environment, may also aid in ascertaining more accurate permissible levels."
So it was audible in many cases even with phase incoherent loudspeakers? Well then possibly the effect could go from "very subtle" to "rather obvious"...or not but it is a rather important variable I think.
"the effect is so subtle it cannot even be claimed to be audible. That's a bit more precise way of saying it than "very subtle."
"
Its only a more precise way of saying it that is what is really implied by it, which in this case I am quite sure it is not.
Your reading comprehension simply sucks! "Very subtle" still connotates audible. Lipshitz is being rather cute with his twist of it but "very subtle" still implies audible, perhaps "unimportant" depending on the listener (as we have established that it affects people differently) but audible.
Sorry the misunderstanding. No one said this thesis is about Lipshitz & audibility of polarity or anything like that. I brought it up since it contains a reference to the very same experiments we have been discussing, in the hopes that it helps people to understand that there is only one (as far as I have been able to establish) such research reported by Lipshitz, and that there are no other, secret and hitherto unknown research results hidden somewhere that only Atkinson has access to (and that would help explain his very different characterization of these test, compared to Lipshitz's own).
That was all. And if you think this grad students personal characterization of Lipshitz's results ("subtly audible") should somehow take precendence over the actual test results Lipshitz got and reported, go ahead. This is a make-believe world anyway as it seems.
TL
You said:
"That's a grad student summing up the Lipshitz experiments. Did you not manage to read it?
"
and now you are backtracking, saying:
"Sorry the misunderstanding. No one said this thesis is about Lipshitz & audibility of polarity or anything like that"
Excuse me but the quote above clearly points out that you WERE saying that the thesis is about Lipshitz and the Thesis IS about audibility of phase inversion. Why lie now? Just admit you were wrong. I seriously think you didn't bother to read what was in fact a very interesting thesis.
Is this a simple misquote, like JA made 20 years ago, or are you deliberately trying to deceive me??
"that it helps people to understand that there is only one (as far as I have been able to establish) such research reported by Lipshitz"
I think everyone got that...a long time ago. Who gives a Lipshitz?
Look, it is all very transparent what you are attempting to do. You are not the noble policeman trying to protect us from misinformation. You have a clear agenda to embarrass a man for a misquote in an article from nearly 20 years ago. Your agenda is to try to "expose" his supposed agenda and it is malicious and grasping at little fine details in an attempt to say "gotcha". Pathetic.
Everyone can do the simplistic math and yes you are correct that with music only 53% were correct but that is not the way the data was presented and give the TOTAL of test signal and music the audibility fit a 99% confidence limit. All this proves is the fallability of reporting statistics without taking everything into account. So you are the big genius who figured that little clever bit out? Good for you. I don't have access to this report so I had to assume that 84/137 was for music and tones but not how it broke down.
"This is a make-believe world anyway as it seems"
No, this is where you are wrong and it is fundamental to the problem here. You sneer at people who claim to hear differences when they change things in their systems. The reality is that often time there IS a difference in the sound and yes it is usually "very subtle". But you know the human brain has a funny way of "locking onto" signals that while subtle become MUCH more obvious once consciously noticed. To some people these things become very pleasant or actively annoying. To others less sensitized they are irrelevant. I will assume you fall into the latter category but you should accpet the fact that there are those whose hearing is attuned better to noticing and finding importance in subtle effects.
The thesis you posted states very clearly that they found the results HIGHLY listener dependent.
"The audibility of phase distortion in audio signals was also highly dependent upon individual ability, although for statistical analysis individual data was not considered. For example, while most test subjects were very good at recognizing what was in general perceivable as phase distortion such as the impulse and the 70 Hz sawtooth wave, a few others had greater difficulty. Specifically, a few subjects seemed to hear clearly the presence of phase distortion in the jazz-vocal test signal for the headphone listening test, while a few test subjects seemed to perceive phase distortion better than others during the loudspeaker listening test."
Now maybe you are one of the ones that would have trouble with the sawtooth pattern but that doesn't mean its not audible, just to you.
It brings up many questions though about the statistical results reported in this thesis (if we really want to break it down like you have the Lipshitz results). If it is highly listener dependent then it suggests that SOME of the listeners were able to reliably detect phase distortion with TONES AND MUSIC. This means for these listeners the effect was not "Very subtle" but rather "readily audible". For those stone ears who couldn't even tell the sawtooth, well I am sure they drag the whole data down into statistical irrelevancy resulting in the authors to confirm "very subtle" when in fact it is very subtle for some, impossible for others and readily audible for a few. Did Lipshitz break it down for individual listeners, looking for individual outliers?
You know how you treat statistical data from listening is not how you treat data for weighing things or measuring voltages etc. The individual response is relevant. Maybe you will learn this and maybe not.
I'll try to be more brief than you:1. "The quote above clearly points out" ONLY what it states, i.e., that in the cited paper the grad student sums up Lipshitz's experiments (even if in passing). As I said that's precisely why I included a reference to it.
What is it that you would rather have me say instead? You claim that I have an agenda but the only agenda I can see signs of seems to be there in your posts such as this.
2. Good if everyone understood what Lipshitz has done and what he hasn't done. Only it doesn't appear that way to me, judging from the replies I keep getting that aim at continuously contradicting the facts or trying to brush them under the carpet.
3. "Playing a policeman": One poster got already banned and had his account deleted a few days ago for pointing out this very same "error" of Atkinson's. Who are the policement? Those who post the facts for everyone to see, or those who go about deleting them where they become inconvenient and eliminate those who post them? Keep watching what happens to people like him. My own posts below and above are selectively deleted as well, even though they are nothing but polite and factual as anyone can see, removing facts given in answer to questions addressed to points I've made, and then archiving (closing) the thread so I can't repost my clarification or answer. Anyone venture a guess why this happens to the extent it does?
Talking about "policing"...
4. Atkinson himself has suggested that any errors of fact in Stereophile be forwarded to his attention so he can issue a public correction of them. I have done this as suggested. Look at the reactions, including your own. The public may draw its own conclusions.
As I've said the issue is easily settled: rewrite the column to correspond to facts, issue a correction of the error, and it's all yesterday's news by then already. This is normally not a big deal for anyone; it's done as a matter of course on a daily basis in the day press. So what's the big deal here for everyone? A taboo topic or or a taboo personage or why is the boat rocking so heavily?
5. "Sneering at people who claim to hear differences": I'm not going to be baited and move the discussion into something vague and inflammatory like this. Note that nothing even remotely related to your claim has been broached in the course of the current thread.
6. If you want to discuss that paper I referred to, start another thread and I'll say what I think of it. This thread is about something else.
Your agenda is clear, to debunk anyone who thinks phase audibility with music is real and to defame JA for making an erroneous statement.
One question for you though since I don't have the Lipshitz paper: Does Lipshitz in his paper actually break down the results and separate test tones from music? Or does he lump all the results together to get the 99% confidence limit? Are his conclusions separated between tones and music or is that your own doing? In other words, are you putting forward their conclusions of the work or your own interpretation of the results??
I have plenty of my own experiments with phase effects while designing speakers to prove that it is real. Using a digital xover with readily adjustable phase allows for some interesting and highly audible results. You should try these things yourself and find out what is audible and what is not. I recommend a Behringer DCX 2496, its cheap and very flexible because you can also do long time delays as well.
"Playing a policeman": One poster got already banned and had his account deleted a few days ago for pointing out this very same "error" of Atkinson's."
Look, its not the pointing out of the error, its the insistence that it was a malicious error intended to mislead readers. You are being intentionally obtuse and trying obfuscate the situation by claiming otherwise. This tells me you are not playing "on the level" and have an agenda to promote.
I don't think it was intentional and most others don't either. It is slanderous to accuse someone of something like that without at least some proof of how it was intentional. That is why TSP was asked to recant or be banned not for pointing out an error. Your way of twisting the reality reminds me of few other chumps here.
"Who are the policement? Those who post the facts for everyone to see, or those who go about deleting them where they become inconvenient and eliminate those who post them? Keep watching what happens to people like him"
Should I just call you Teflon man? Rather than addressing the issue headon you deflect it to the moderators who, BTW could care less about whether JA made a mistake or not but they care very much about slanderous remarks...and rightly so. TSP got what was coming to him for making comments without substantiation and you do too.
" Atkinson himself has suggested that any errors of fact in Stereophile be forwarded to his attention so he can issue a public correction of them. I have done this as suggested. Look at the reactions, including your own. The public may draw its own conclusions.
"
Did you issue them in private first? Give him a chance to address you personally and make the correction? If not then you were just baiting him in a public forum...not cool.
"If you want to discuss that paper I referred to, start another thread and I'll say what I think of it. This thread is about something else.
"
What a copout!! YOU posted that link in THIS thread! So why should I have to start an new thread for you to address it? If you introduced it you should be prepared to discuss it.
1. "...to debunk anyone who thinks phase audibility with music is real and to defame JA for making an erroneous statement."You are misunderstanding. I am not debunking anyone; Lipshitz is. Or, to be precise, he debunked that myth almost thirty years ago already.
Defame JA? Exactly how have I "defamed" him? By pointing out an error of fact that gives rise to a false claim? Here on AA he has invited people to point out errors of fact so that he can correct them on Sphile. So here's one. But instead, Atkinson has resorted to attacking my person (not the content of my post) in an effort to cast doubt on my credentials, hence also the credibility of the numbers by Lipshitz that I've posted.
Sorry if you for some reason take it personally, but I think it has more to do with your worldview than any of the numbers I've posted. As with all religions, when an article of faith is contradicted by facts, these facts will be first relativized and then suppressed or swept under the carpet. Look what's going on in this thread and above in the "Wits and Wisdom of Lipshitz" thread... It's even more obvious than one would expect, and even sanctioned by you know whom, which probably means that the article of faith now questioned is even more central in the belief system than what we'd have thought.
2. The Lipshitz papers: Read the reply to you by Andy_C below.
3. Audibility of polarity: I am simply pasting here what Lipshitz has found: his numbers. Argue with him if you think his results lie. I don't really care enough about this subject.
4. "...don't think it was intentional" :
Fine, then he can correct this error if it was merely accidental.
Has he done so? No.
Has he even admitted his error: Not really. In effect he has, but look at the extent of the evasions.
Has he been informed of this error in the past, and by others as well? Yes on both accounts (search the Critics forum -- this has been pointed out to him by at least 2 other posters early last spring).
Has he been offered opportunities to correct this error in the past: Several times.
If this wasn't Atkinson we are talking about, would you then say the error was more likely intentional than not?
And if he wasn't truthful in his claims (remember there is a number of such statements), and if he keeps insisting that there is nothing wrong with such statements, what will you say about the moderators' actions in promptly banning the poster Truthseekerprime and deleting his AA account for pointing this out?
5. "Did you issue them in private first? Give him a chance to address you personally and make the correction?"
When you detect a factual error in statements posted here on AA, do you notify the poster in private or do you note the error in one of your own posts made in reply? (This is how it happened with Atkinson's error -- search for "The Atkinson Diet" over at Critics.) You certainly haven't done so with me, not once, when believing I have committed an error. And once more, JA has publicly invited people to bring up factual errors published in Sphile so he can publish a correction regarding them (he has proudly claimed here on AA that this -- publicly correcting errors -- is what separates Sphile from its competitors).
Also to repeat once more, Atkinson has been offered numerous chances to make this correction, and what has done? Nothing.
So... merely an accidental error? What is it doing, still up there? ;-)
6. "YOU posted that link in THIS thread!" Yes and for the very specific reason related to this thread topic that I already mentioned. If you want to change the topic, start another thread. I'll be glad to participate.
Howdy
You have been posting phrases like "... the moderators' actions in promptly banning the poster Truthseekerprime and deleting his AA account for pointing this out?"
As has been explained before to you ( http://www.audioasylum.com/forums/prophead/messages/4/41389.html ) Truthseekerprime deleted his own account, by his words, in protest of having his posts deleted. I for one believe him that that was his reason for doing so. I don't think he saw "The Bored"'s post explaining what was going on before he deleted his account. That's why I posted http://www.audioasylum.com/forums/whine/messages/1/11884.html , just in case.
-Ted
See link: I have been assuming that your proclaiming to TSP that "you are banned till you acknowledge and recant" and The Bored's proclaiming immediately afterwards that "User 43577: account closed" meant what the posters stated.
If by this something else was meant than what those words seem to mean, I stand corrected (if rather perplexed).
- "Your transgression was explained to you, you are banned till you acknowledge and recant" (Open in New Window)
Howdy
It's not too complicated:
"The Bored" was noting for others that TSP's apparent response to being asked to support his allegations or to retract them was to close his own account.
I thought that there was a chance that there was a misunderstanding and hence my post just above the one you cite http://www.audioasylum.com/forums/whine/messages/1/11884.html
Note also that as I said this was explained directly to you (if tersely) in http://www.audioasylum.com/forums/prophead/messages/4/41389.html
I expect no more comments about this.
-Ted
This message has been moved to a more appropriate venue .
> Atkinson has resorted to attacking my person...
No. I pointed out 1) that the mistake you made in repeated postings
(confusing confidence level with the actual score) demonstrated an
unfamiliarity with statistical analysis of experimental results, and 2)
that you had incorrectly stated that the phrase "subtly audible" meant
the same as "inaudible." I made no other comments about you personally.
> Atkinson has been offered numerous chances to make this correction, and
> what has done? Nothing.
Really? That's terrible. :-)
John Atkinson
Editor, Stereophile
"...an unfamiliarity with statistical analysis of experimental results..."
What are we to think about your ability to judge statistics skills when you read Lipshitz' results to mean that polarity reversal is audible "on music to a 99% confidence limit!"? ;-)
"...you had incorrectly stated that the phrase "subtly audible" meant
the same as "inaudible."
Indeed, so "subtly audible" on music that it couldn't be detected at all. (It could of course be measurable, but the effect seems to be below the hearing level of humans.) Or so says Lipshitz based on his own research on the subject. ;-)
Could you please finally point out the locus in Lipshitz where he claims differently?
TIA! :-)
> > ...an unfamiliarity with statistical analysis of experimental > > results..."
>
> What are we to think about your ability to judge statistics skills when
> you read Lipshitz' results to mean that polarity reversal is sudible
> "on music to a 99% confidence limit!"? ;-)
I believe this is know as an IKIABWAY or somesuch. That the poster
excuses his own error on the grounds that others have also made errors.
You admitted making the error multiple times but declared it "no biggie."
I thought it was a "biggie" as you had made your misunderstanding the
basis of many postings criticizing me. I politely pointed out your
misunderstanding without making it personal, but without any
acknowledgment on our part (unless it was buried somewhere in the
voluminous postings) that you were wrong to make that specific
criticism.> > ...you had incorrectly stated that the phrase "subtly audible" meant
> > the same as "inaudible.
>
> Indeed, so "subtly audible" on music that it couldn't be detected at
> all.
That is your intepretation, but I and others disagree based on Stanley's
own comments. You may feel that your opinions should have the status of
fact but I don't. And I certainly dont see the point of see arguing with
your opinion.
Look. You have posted a large number of messages saying that I was
wrong to write what I wrote 20 years ago. I told you that it was
possible that I had made an error and that that I would look up the
original article in the BAS Speaker when I could. In the meantime,
I amended the text that had you so angry so that it was literally
correct. Yet here you are, still going on and on and on. Does your
behavior not strike you as being just a tiny bit obsessive? Not totally
rational? Possibly an over-reaction? Perhaps due to an underlying
agenda on your part, given that you have now raised the matter in
another thread just a couple of hours back?
I have said all that I intend to. I doubt that that will be so in your case, however. :-(
John Atkinson
Editor, Stereophile
(Edited to correct a couple of typos)
"I believe this is know as an IKIABWAY or somesuch. That the poster excuses his own error on the grounds that others have also made errors."
Oh, like the thing that you are now doing right now? I understand.
I admitted my own error and as you know it anyway has nothing to do with the issue on hand. If you disagree, show me the locus in Lipshitz where he supports your claims, please. (But you will wisely avoid doing so because you know that there is none.)
What I don't understand is why you want to keep posting such a large number of messages trying to deny it all. Errors (of facts, even of judgements) are normal and usually no one things too much of them. Moreover this is the kind of a factual error that you have requested be forwarded to your attention if detected in Stereophile. I have done just that, and I have done that most politely. But by now your avoidance behavior has made it clear that there must be more at stake than an innocuous error or a defective statistical analysis of test scores. Instead of being a man about it, you keep trying to somehow deflect attention away from your misrepresentation by proposing that I have an "agenda" because I pointed it out. Yet the only agenda in evidence seems to be that of yours; otherwise why not acknowledge the thing and get it over and done with? Is the polarity issue really that crucial for your purposes? If audio journalism is to be about investigation of facts, not imposition of fiction, it needs to be based on the best available information, not driven by ideologies and interests other than those in truth alone.
"That is your intepretation, but I and others disagree based on Stanley's own comments. You may feel that your opinions should have the status of fact but I don't."
You are falsifying my point. Where is the "interpretation"? I posted the score from Lipshitz' own research results. Numbers are not an opinion, numbers are facts. If you think they lie, you may wish to address Lipshitz directly to correct his scores for him so they better suit your purposes.
"...obsessive...not rational...over-reaction...underlying agenda", etc. :
Once again, you try to avoid dealing with the message by attacking the messenger.
That message, as shown by Lipshitz' own test scores, was:
- Stan Lipshitz never confirmed audibility of polarity inversion on music with his DBTs.
- Stan Lipshitz never confirmed audibility of polarity inversion on music to a 99% confidence limit.
- None of this has ever been done in any other DBTs, either.
Contrary to what you claim in each case.
TL
This classic needs to stay in the record (particularly the part about acting "most politely," then telling JA to be "a man about it..."
ILSHIFMCB!
(I'm Laughing So Hard I've Filled My Colostomy Bag)
One question for you though since I don't have the Lipshitz paper: Does Lipshitz in his paper actually break down the results and separate test tones from music? Or does he lump all the results together to get the 99% confidence limit?
You might want to try actually reading the thread.
Here is a quote you posted:
"The authors have demonstrated the two-tone experiment described above to numerous people on different systems. No one has ever failed to hear the timbral change with phase, and discern the polarity reversal on this signal with unvarying accuracy. Indeed, in a double-blind demonstration to eleven members of the SMWTMS audio group [13], the accuracy score was 100% on the summed 200-Hz and 400-Hz tones over loudspeakers, and overall, including musical excerpts, the results on the audibility of the polarity inversion of both loudspeaker channels were 84 correct responses out of 137, this representing confidence of more than 99% in the thesis that acoustic polarity reversal is audible."
And reference 13 is this:
[13] B. F. Muller, "Third World: The Scientific Subjectivists," Audio Amateur, vol. 11, p. 64 (1980 Jan.).
So apparently then Lipshitz did NOT separate the results between test tones and music, lumping them together and drawing his conclusion for both tones AND music. Or am I missing something? If that is the case, then Tlyyra's objections don't even agree with Lipshitz's own words and JA, while perhaps not strictly accurate comments were also not grossly off target with regard to these remarks.
Quite a bit, in fact.
Read the sources cited.
And then read also what you yourself post. "In a double-blind demonstration to eleven members of the SMWTMS audio group...":
- What does it say the 99% confidence level applies to? The overall score (combining both test tones and music).
(This besides the point that the figure here is completely meaningless anyway.)
- What does it say about audibility on music? Nothing.
- How does Atkinson interpret this information? "Work by Stanley Lipshitz in the late '70s [footnote to the above citation], using carefully organized double-blind testing, confirmed that a reversal of absolute signal polarity will be subtly audible on music to a 99% confidence limit! (Indeed, it is one of the few things that can be reliably detected with double-blind testing.)
Key claims misrepresenting the Lipshitz quote you yourself cited are underlined, to assist your reading.
Pay attention also to the false claim in the last sentence: there are zero DBTs showing this at all, and zero showing that this can be reliably detected. Trust you will let me know if I'm wrong...
If you are really interested in this test with the eleven members of the SMWTMS audio group, I am attaching the scores from it once more (maybe the 10th time now?). The row heading "Van der Kooy / Lipshitz Training Signal" means "the test tone designed by Lipshitz & Vanderkooy" and the row heading "Sheffield & Other Phonograph Records" means "music." Ask me if you can't figure it out.
"What does it say about audibility on music? Nothing"
Not true or they wouldn't have grouped the data together. Did you ask yourself why they did that? Was it Lipshitz then who had the agenda to show polarity was audible? One might wonder that now or is there something YOU are missing?
"How does Atkinson interpret this information? "Work by Stanley Lipshitz in the late '70s [footnote to the above citation], using carefully organized double-blind testing, confirmed that a reversal of absolute signal polarity will be subtly audible on music to a 99% confidence limit! (Indeed, it is one of the few things that can be reliably detected with double-blind testing.)
"
Once again, your implication that he made a misinterpretation on purpose or with malicious intent is clearly your attempt to defame his character. However; your own "interpretations" of Lipshitz's results are no more factual as you are separating data that they in fact used together. Perhaps there are things in the test about which you have no idea and they therefore had a good reason to lump them together. Lipshitz was no dummy and if he lumped the data together there must have been a good reason (unless you think he was misleading people, in which case JA is only guilty of being misled by Lipshitz).
"Key claims misrepresenting the Lipshitz quote you yourself cited are underlined, to assist your reading"
No assistence necessary because the conclusion from Lipshitz himself was not so different.
""The authors have demonstrated the two-tone experiment described above to numerous people on different systems. No one has ever failed to hear the timbral change with phase, and discern the polarity reversal on this signal with unvarying accuracy. Indeed, in a double-blind demonstration to eleven members of the SMWTMS audio group [13], the accuracy score was 100% on the summed 200-Hz and 400-Hz tones over loudspeakers, and overall, including musical excerpts, the results on the audibility of the polarity inversion of both loudspeaker channels were 84 correct responses out of 137, this representing confidence of more than 99% in the thesis that acoustic polarity reversal is audible.""
Now, do you see a breakdown in their conclusions such as you have done? No. Why is that do you think?? Is it a conspiracy?? Doubtful. One thing is clear they are equating audibility with both tones AND music. Now you can take issue with Lipshitz for lumping the data together if you like but you cannot say that JA's interpretation of their conclusion is false. They DO mention polarity reversal being audible with music.
It is YOU who is reinterpreting their results to declare JA to be a maliciously misrepresenting the results. JA, 20 years ago, was just paraphrasing their conclusions, which do include polarity reversal as being audible on music. Maybe JA's paraphrasing was a bit selective but you are completely rewriting their conclusions!!
> > "why they did that [grouped data together in narrative reports]" :Not true. Even in the quote you have given they give the score for test tones. But in this instance, not for music; probably because the results went against the expectations and Lipshitz wasn't too comfortable spelling out the results loud and clear. (A little bit like with the totally unanticipated results from the recent test MikeL, Ted & co. did with the Transparent Opus MM $30K+ cable vs. off-the-shelf Monsters, which are not exactly trumpeted around here...) Why else try and brush aside the score on music but not with the test tones? Yet music was kind of the key, don't you think? We're not interested in playing test tones in our systems.
> > "your implication..." :
Let's stick with what I am actually saying for a change, not what in your opinion I must be implying, OK?
> > "[Atkinson] made a misinterpretation on purpose or with malicious intent" :
As you like to phrase it.
> > "[my] own "interpretations" of Lipshitz's results" :
Score on music: 60 / 113 = 53%, p less than .7
[Conclusion:] "The audibility of polarity on music was not confirmed whether the sound source was vinyl or a 1/4" 2-track master tape I recorded of a soprano recital with piano accompaniment."
[source: http://www.provide.net/~djcarlst/abx_pola.htm]
What's the part with the "interpretation"? 60? 113? 53%? p less than .7?
> > "you are separating data that they in fact used together" ":
First of all, they didn't do that. They tested music and test tones separately and kept scores separately, too. Why are you claiming something like this?
Secondly, even you will understand that had they actually not distinguished between music & test tones, the whole experiment would have been meaningless to everyone who's not spending time listening to test tones. And John Atkinson's claim that it's "audible on music to a 99% confidence limit!" would be just as false as before.
If they articulated the results only reluctantly or imperfectly, without separately listing the score on music in each of their reports (I've referenced for you at least two where this is listed separately -- they were not tested as "lumped together" but separately), that's another matter which has everything to do with reporting style but nothing to do with the test results themselves.
> > "do you see a breakdown in their conclusions such as you have done?"
The passage quoted doesn't give the score on music, but it does give the score on test tones, and it does give the overall score combining the results on both test tones and music. What is it that you cannot see? Have you checked any of their actual discussions included in those sources, or you think this is the only existing reference in the world to this experiment by Lipshitz? Actually it's quite well known and widely discussed, even included in curricular materials in schools, theses works, etc., etc., so your strange attitude about it is a bit misguided I'd say. As I've said before: check the references. I too have provided you with some and there are more. Run a google search at the least. I don't personally own a monopoly to the information on their research.
> > "One thing is clear they are equating audibility with both tones AND music" :
That's pretty evident wouldn't you say? But the point is that audibility could only be statistically observed *if* the 24/24 score on test tones was included as part of the *overall* score. On music, audibility could *not* be established *with any confidence at all.* Go back and review their results once again (or see link for a more detailed analysis of them if that helps).
> > "you cannot say that JA's interpretation of their conclusion is false" "
Again -- John Atkinson: "Work by Stanley Lipshitz in the late '70s (footnote 9), using carefully organized double-blind testing, confirmed that a reversal of absolute signal polarity will be subtly audible on music to a 99% confidence limit! (Indeed, it is one of the few things that can be reliably detected with double-blind testing.)"
Now tell me: This statement by Atkinson, true or false?
> > "It is YOU who is reinterpreting their results" :
Would you care to point out where and how I'm doing this? 60 / 113 = 53%. In the "60" or in the "113" or in the "53%"?
> > "JA...was just paraphrasing their conclusions" :
Some paraphrasing: "not audible" is now "is audible"; "never been detected" is now "can be reliably detected," etc. etc.
> > "you are completely rewriting their conclusions" :
60 / 113 = 53%. Would you care to point out where I'm doing the rewriting: In the "60" or in the "113" or in the "53%"?
> > "Is it a conspiracy?"
Too strong an expression most likely, but what anyone can see is a sustained effort in this thread by especially those with an industry affiliation affixed to their name to camouflage, deny, or suppress through other means the fact that JA has misrepresented important research results in his Stereophile column.
You quoted me as writing back in 1988: John Atkinson: "Work by Stanley
Lipshitz in the late '70s (footnote 9), using carefully organized
double-blind testing, confirmed that a reversal of absolute signal
polarity will be subtly audible on music to a 99% confidence limit!"
Please note that I did amend the text of this article a while back to
more accurately reflect the Lipshitz findings. That you continue to
quote the original text is somewhat dishonest on your part, I feel.
I also wrote in that 1988 article "(Indeed, it [ie, absolute polarity] is
one of the few things that can be reliably detected with double-blind
testing.)"
This statement of mine is absolutely correct. So why do you keep
quoting it as a "gotcha."
John Atkinson
Editor, Stereophile
1. "Please note that I did amend the text of this article a while back to more accurately reflect the Lipshitz findings."See link. Are you trying to become a public jester or something?
2. "That you continue to quote the original text is somewhat dishonest on your part, I feel."
Sorry if you feel wounded about it, but, first of all, the point is and has been about the truthfulness of what was said in that column that you originally published.
Secondly, the point that you have now added -- "[and definitely audible on test signals]" -- is a masterly example of your evasiveness, though, I have to add, quite clever in its own way: it seems like it's a modification of some sort but in the reader it only reinforces the false understanding of what Lipshitz did (your misrepresentation of his research results). Bingo! You think you now got it all -- is that it? No need to own up to what you've falsely claimed, and still keep your fans happy?
It's no problem for me to include those 6 new words in my quotations from now on.
On a more serious note, please explain how you think that totally tangential note corrects anything about the false statements about Lipshitz' results that are still there, unmodified? Yes, the ones we have been talking about... If you are forgetting you can consult the link.
3. "I also wrote in that 1988 article '(Indeed, it [ie, absolute polarity] is one of the few things that can be reliably detected with double-blind testing.)'
This statement of mine is absolutely correct. So why do you keep quoting it as a 'gotcha'."Sorry, what??? Well, if this is "absolutely correct" it should be very quick and easy for you to provide a couple of references to such DBTs with which audibility of absolute polarity signal reversal has been detected on music (this is what you claim, not your newest little white "misquotation" above, and you claim it still today: See last 2 sentences at the end of 3rd paragraph in "Absolute Polarity" ). If it has been "reliably" detected in DBTs, surely there must be at least two DBTs with which it has been detected, since otherwise you couldn't even vaguely suggest "reliability," which by any acccount should indicate greater than 50% success rate in a series (plurality of tests).
Right, the point about statistics and you :-)
So looking forward to learning about those DBTs! Thanks in advance!
But since we both know there are in fact zero such DBTs, you better start thinking of some new evasion for your reply.
You think someone still believes you are simply making innocent errors upon innocent errors? What a careless editor-in-chief they have in that case picked for the most prestigious audio journal in the country... ;-)
Can you help thinking that if only you had acknowledged and corrected these "errors" already, say, last winter, there'd be no need for you to push further and further forward on this increasingly tortuous path of more and more complicated evasions and more and more obvious denials that indeed are busy becoming most unflattering to you and your case?
> It's no problem for me to include those 6 new words in my quotations
> from now on.
Thank you. If you are to continue criticizing others for their supposed
lack of honesty, it behoves you to behave honestly yourself.
John Atkinson
Editor, Stereophile
No prob. You know yourself what a joke that "amendment" is. Do you expect even your staunchest supporters to take you seriously after that? Even if you yourself can't, I'm sure many among them can read plain numbers. Do remember that more people read these discussions than participate in them, and that's worldwide.I can see that instead of opting to provide supporting material for your absurd claims, you keep jumping at the messenger to avoid attention to the message.
We still remain waiting for the two references that could minimally support your following claim and show that you are not simply a cynical lier in advancing that false claim over and over again (your latest falsehood in boldface):
John Atkinson in Stereophile: "Work by Stanley Lipshitz in the late '70s (footnote 9), using carefully organized double-blind testing, confirmed that a reversal of absolute signal polarity will be subtly audible on music [and definitely audible on test signals] to a 99% confidence limit! ( Indeed, it is one of the few things that can be reliably detected with double-blind testing. )"
John Atkinson above: " This statement of mine is absolutely correct. "
So: Two references please to show that audibility of polarity reversal can be reliably detected with DBTs?
Even one??
Otherwise the conclusion will be evident even to the casual readers here that you are lying, since this has been pointed out to you already many times before, and so it is not like you are unaware of the facts by now and are simply making one more "innocuous error" as your apologists would have it.
When will you realize that more falsehoods will not get you out of the previous falsehoods?
I feel almost sorry for you. You have to spin such a web of falsehoods in support of your previous falsehoods that you are busy becoming something like a caricature of your own self.
> > Please note that I did amend the text of this article a while
> > back to more accurately reflect the Lipshitz findings. That you
> > continue to quote the original text is somewhat dishonest on
> > your part, I feel.
>
> No prob. You know yourself what a joke that "amendment" is.
I am sorry that I cannot claim the same mindreading powers that you
possess, tlyyra. No, I don't know that it was a "joke." I amended the
text in the usual way to indicate an editorial change (brackets +
italics). And regardless of your opinion of the amendation, I thought it
apparent that my having done so gives the lie to the claims you and
TruthSeekerPrime have made on this newsgroup that I took _no_ action
after the issue had been raised.
> Do you expect even your staunchest supporters to take you
> seriously after that? Even if you yourself can't...
More mindreading claims!
> I can see that instead of opting to provide supporting material
> for your absurd claims, you keep jumping at the messenger to
> avoid attention to the message.
Let's see. I have not called you a liar (or "lier"), I have not
cast doubt on your honesty, not have I gratuitously called you
names, tlyyra. What I have done is respond to you what you have
incorrectly written (ie, your erroneous statement that "subtly
audible" means the same as "inaudible") and your own
misunderstanding of statistical analysis, where you repeatedly
declared that a 99% confidence limit means the same as the
listeners getting 99% of identifications correct.
> We still remain waiting for the two references that could
> minimally support your following claim and show that you are not
> simply a cynical lier in advancing that false claim over and over
> again.
"We"? Who else are you speaking for, tlyyra? Or do you have a mouse in
your pocket? :-)
> > > John Atkinson in Stereophile: "Work by Stanley Lipshitz in the
> > > late '70s (footnote 9), using carefully organized double-blind
> > > testing, confirmed that a reversal of absolute signal polarity will
> > > be subtly audible on music [and definitely audible on test signals]
> > > to a 99% confidence limit! (Indeed, it is one of the few things that
> > > can be reliably detected with double-blind testing.)"
> >
> > John Atkinson above: "This statement of mine is absolutely correct."
>
> So: Two references please to show that audibility of polarity
> reversal can be reliably detected with DBTs? Even one?? Otherwise
> the conclusion will be evident even to the casual readers here
> that you are lying, since this has been pointed out to you already
> many times before, and so it is not like you are unaware of the
> facts by now and are simply making one more "innocuous error" as
> your apologists would have it.
Sigh. I really think that your anger arises from your apparent
inability either to quote people correctly and from your
willingness to make an assumption about what was _really_ meant
by the words on the page, tlyyra, rather than their literal meaning.
Why you are so ready to dismiss the first-order meaning of what someone
has written, I have no idea.
For example, please note that my saying earlier in the thread that "This
statement of mine is absolutely correct" was clearly referring to my
parenthetical sentence on my 1988 essay that "Indeed, it is one of the
few things that can be reliably detected with double-blind testing,"
_not_ the entire quoted passage as you misleadingly present above. As
I was introducing the subject of absolute polarity, the antecedent for
the word "it" was just plain "absolute polarity." And yes, absolute
polarity _is_ conclusively audible in DBTs. You yourself, tlyrra, have
given references to such tests where the listeners scored 24 correct
identifications out of 24 trials, and in a report I published in HFN in
the mid-1980s, I write about a test I took in Boston in 1984 where,
using an ABX comparator, I scored 19 correct identifications out of 20.
Now, it is fair to clarify that such success was with asymmetrical test
tones. But if you read my parenthetical sentence carefully, you will
note that I did not qualify it in any way, by saying "on test tones
only" or "not on music." This is because I was using this literally
correct statement to poke fun at the whole concept of double-blind
testing. I had thought the humor self-evident; obviously I was wrong to
have thought that, given that I have had to explain the joke to you. :-(
Earlier in this thread, you repeatedly stated that what I actually meant
with this parenthetical sentence was that "it" (ie, absolute
polarity) "could be reliably detected _with music_." That is _not_ what
I actually wrote back in 1988, and I don't see the point in arguing with
your incorrect interpolation of what you think I _should_ have written
instead of what I actually wrote. You are welcome to wander alone in
_that_ semantic forest, tlyyra. :-)
> When will you realize that more falsehoods will not get you out
> of the previous falsehoods? I feel almost sorry for you. You have
> to spin such a web of falsehoods in support of your previous
> falsehoods that you are busy becoming something like a caricature
> of your own self.
Falsehoods? I correctly reported back in 1988 what Stanley Lipshitz had
written, that with a mixture of music and test tones, absolute polarity
was detectable under blind conditions with a confidence limit of 99%.
Subsequent breakdown of the test results _after_ my 1988 essay had been
published showed total identification with test tones but not with music
to an acceptable degree of statistical confidence. But as Stanley did
say back in the 1970s, this is something that is "subtly audible" on
music and may not even be detectable by all listeners. These factors, as
has been pointed by others, make experimental design and statistical
analysis more difficult than had been appreciated back in the 1970s.
And please note that, despite your mindreading claims earlier in this
thread, I don't believe absolute polarity to be very important. While I
do preserve it in my own recordings, almost all commercial recordings
are compromised in this respect, as I pointed out in that 1988 essay.
But if there is a recording that I feel might benefit, I flip the preamp
polarity with my remote control. Most often it makes no difference; very
occasionally it does. It's "no biggie," to use your own phrase, and
certainly not worth the thousands of words you have now posted on this
subject.
Happy Holidays.
John Atkinson
Editor, Stereophile
So you are unable to provide any references to support your claim that reversal of absolute polarity will be audible on music, and that this audibility "indeed can be reliably detected with DBTs"?Not even one.
What are we supposed to think of your sincerety here?
What's a bit more of a surprise, however, is the degree of your arrogance, which seems to lead you to thinking that you can get away with anything with those audiophiles reading your obfuscations.
Hmmm.
Let's see:
> "...the claims you and TruthSeekerPrime have made on this newsgroup that I took _no_ action after the issue had been raised..." "
Sorry, where have I made such claims? What I've pointed out is that you have not corrected your misrepresentation of Lipshitz' claims, something which holds still at the time of writing this post.
But I will abstain from saying that you are lying again since pointing that out that seems to be punishable by banishment from this forum.
> "Who else are you speaking for, tlyyra?" :
Do you think Robert Young, bjh, and Morricab are the only ones following this discussion?
> "...your apparent inability either to quote people correctly and from your willingness to make an assumption about what was _really_ meant by the words on the page...rather than their literal meaning. Why you are so ready to dismiss the first-order meaning of what someone has written?" :
Hmmm. Let's see: In one of the sources you yourself cite, the following words by Lipshitz appear on the very first page:
"On normal music or speech signals phase distortion appears not to be generally audible." (Lipshitz & Vanderkooy, "On the Audibility of Midrange Phase Distortion in Audio Systems," JAES Vol.30 No.9, September 1982.)
Your take on the "first-order meaning" of what Lipshitz just stated, as presented in your Stereophile column:
"Work by Stanley Lipshitz in the late '70s (footnote 9[above article])...confirmed that a reversal of absolute signal polarity will be subtly audible on music..." (http://www.stereophile.com/asweseeit/988awsi/index1.html)
Talk about "inability to quote people correctly"... :-(
Tell me, are you really this cynical, or just not bothering to pay any attention?
> "Falsehoods? I correctly reported back in 1988 what Stanley Lipshitz had written, that with a mixture of music and test tones, absolute polarity was detectable under blind conditions with a confidence limit of 99%." :
No. You state "audible on music." So here we have yet another...well, I'm not going to say lie , since for some reason that seems to get people banned if it's about these test results; instead, let's just say "absolutely correct statement" of what Lipshitz "really meant"... How's that? ;-)
> "...to poke fun at the whole concept of double-blind testing." :
Oh. So you've turned your coat since that article, since in it you still praise Lipshitz, his "formidable mind" and "carefully organized" DBTs, so highly? Poke fun about them now just because it turned out those pesky DBTs in the end actually didn't provide any support for your agenda?
> "I don't believe absolute polarity to be very important." :
No? Another interesting twist in the plot then? In that same Stereophile column you go as far as giving people the mailing address where to send in their checks to get a copy of Clark Johnsen's self-published pamphlet promoting the subject...
The rest of your post doesn't deserve a word in reply. If I may paraphrase your own words, it's just an attempt to create a semantic swamp.
> > > > > John Atkinson in Stereophile: "Work by Stanley Lipshitz in the
> > > > > late '70s (footnote 9), using carefully organized double-blind
> > > > > testing, confirmed that a reversal of absolute signal polarity will
> > > > > be subtly audible on music [and definitely audible on test signals]
> > > > > to a 99% confidence limit! (Indeed, it is one of the few things that
> > > > > can be reliably detected with double-blind testing.)"
> > > >
> > > > John Atkinson above: "This statement of mine is absolutely correct."
> > >
> > > So: Two references please to show that audibility of polarity
> > > reversal can be reliably detected with DBTs? Even one?? Otherwise
> > > the conclusion will be evident even to the casual readers here
> > > that you are lying, since this has been pointed out to you already
> > > many times before, and so it is not like you are unaware of the
> > > facts by now and are simply making one more "innocuous error" as
> > > your apologists would have it.
> >
> > Sigh. I really think that your anger arises from your apparent
> > inability either to quote people correctly and from your
> > willingness to make an assumption about what was _really_ meant
> > by the words on the page, tlyyra, rather than their literal
> > meaning. Why you are so ready to dismiss the first-order meaning
> > of what someone has written, I have no idea.
> >
> > For example, please note that my saying earlier in the thread
> > that "This statement of mine is absolutely correct" was clearly
> > referring to my parenthetical sentence on my 1988 essay that
> > "Indeed, it is one of the few things that can be reliably
> > detected with double-blind testing," _not_ the entire quoted
> > passage as you misleadingly present above. As I was introducing
> > the subject of absolute polarity, the antecedent for the word
> > "it" was just plain "absolute polarity." And yes, absolute
> > polarity _is_ conclusively audible in DBTs. You yourself,
> > tlyrra, have given references to such tests where the listeners
> > scored 24 correct identifications out of 24 trials, and in a
> > report I published in HFN in the mid-1980s, I write about a test
> > I took in Boston in 1984 where, using an ABX comparator, I
> > scored 19 correct identifications out of 20.
> >
> > Now, it is fair to clarify that such success was with
> > asymmetrical test tones. But if you read my parenthetical
> > sentence carefully, you will note that I did not qualify it
> > in any way, by saying "on test tones only" or "not on music."
> > This is because I was using this literally correct statement
> > to poke fun at the whole concept of double-blind testing. I
> > had thought the humor self-evident; obviously I was wrong to
> > have thought that, given that I have had to explain the joke
> > to you. :-(
>
> So you are unable to provide any references to support your claim
> that reversal of absolute polarity will be audible on music, and
> that this audibility "indeed can be reliably detected with DBTs"?
I have inserted the full text of mine to which you were responding,
tlyyra, so others can clearly see how you twist my words in order
to score debating points. I don't see the point of continuing a
discussion with someone like yourself who misquotes, misinterprets,
and misrepresents what others write, then claims he knows what others
_really_ meant to have written, even though they didn't actually write
the words he "quotes."
> I will abstain from saying that you are lying again since pointing
> that out that seems to be punishable by banishment from this forum.
Thank you, tlyrra. That is very gracious of you. In return, I will
refrain from asking whether English is your first language, given
how difficult you find it to comprehend what others write.
> In one of the sources you yourself cite, the following words
> by Lipshitz appear on the very first page:"On normal music or
> speech signals phase distortion appears not to be generally audible."
> (Lipshitz & Vanderkooy, "On the Audibility of Midrange Phase
> Distortion in Audio Systems," JAES Vol.30 No.9, September 1982.)
"Not generally audible" does not mean the same as inaudible under all
circumstances, so I fail to see this as the "gotcha" you believe it
to be, tlyrra.
> Your take on the "first-order meaning" of what Lipshitz just stated,
> as presented in your Stereophile column: "Work by Stanley Lipshitz
> in the late '70s (footnote 9 [above article])...
No, the complete footnote 9 in my 1988 essay referred to _two_
articles, vide: 'A little understood factor in A/B testing,'
The BAS Speaker, March 1979, followed by 'On the Audibility of
Midrange Phase Distortion in Audio Systems' (with John Vanderkooy
and Mark Pocock), JAES, Vol.30 No.9, September 1982."
> ...confirmed that a reversal of absolute signal polarity will be
> subtly audible on music..."
> (http://www.stereophile.com/asweseeit/988awsi/index1.html)
> Talk about "inability to quote people correctly"... :-(
Forgive me, but I fail to see how how does your quoting from
the 1982 AES paper proves that my 1988 paraphrase was incorrect.
"Not generally audible" allows for audibility under some
circumstances; Stanley had earlier used the phrase "subtly audible"
on music. None of this equates to "inaudible under all circumstances,"
as you appear to believe.> > "Falsehoods? I correctly reported back in 1988 what Stanley
> > Lipshitz had written, that with a mixture of music and test tones,
> > absolute polarity was detectable under blind conditions with a
> > confidence limit of 99%.":
>
> No.
No? That is what Stanley wrote. I know that in your opinion he was
wrong to state the combined results, but your opinions are not my
concern.
> You state "audible on music."
When did I write that it was "audible on music" without qualification?
Again, tlyrra, your willingness to misrepresent what others write makes
discussion fruitless.
> > please note that, despite your mindreading claims earlier in this
> > thread, I don't believe absolute polarity to be very important.
>
> No? Another interesting twist in the plot then?
When have I said the opposite, that absolute polarity _is_ very
important? More mind-reading on your part, tlyyra?
> In that same Stereophile column you go as far as giving people the
> mailing address where to send in their checks to get a copy of Clark
> Johnsen's self-published pamphlet promoting the subject...
Yes. It's an excellent read, passionately written. I still recommend it.
> > While I do preserve it in my own recordings, almost all commercial
> > recordings are compromised in this respect, as I pointed out in that
> > 1988 essay. But if there is a recording that I feel might benefit, I
> > flip the preamp polarity with my remote control. Most often it makes no
> > difference; very occasionally it does. It's "no biggie," to use your
> > own phrase, and certainly not worth the thousands of words you have
> > now posted on this subject.
>
> The rest of your post doesn't deserve a word in reply.
Sorry. I thought you would interested in my opinion. Oh well,
have a nice day, tlyrra.
John Atkinson
Editor, StereophileEdited to correct a typo
"But in this instance, not for music; probably because the results went against the expectations and Lipshitz wasn't too comfortable spelling out the results loud and clear."
Reading Lipshitz mind now are we? You know, of course that mind reading has never been satisfactorily proven with a DBT; therefore I don't believe you are capable of doing so and THEREFORE your WAG (wild assed guess) is duly noted.
You have serious diarrhoea of the keyboard. Can't you simply say the same things over and over in a more compact form? Sheesh!
"> > "[my] own "interpretations" of Lipshitz's results":
Score on music: 60 / 113 = 53%, p less than .7
[Conclusion:] "The audibility of polarity on music was not confirmed whether the sound source was vinyl or a 1/4" 2-track master tape I recorded of a soprano recital with piano accompaniment."
[source: http://www.provide.net/~djcarlst/abx_pola.htm]
What's the part with the "interpretation"? 60? 113? 53%? p less than .7?
"
None of this was from the Lipshitz paper. NONE! JA was taking his paraphrase from that paper and not from the ABX website that DID NOT EXIST in 1988 when JA made the reference!! DUH! So while you MAY be right about the conclusions (thus making Lipshitz look a little bit stupid) JA was paraphrasing conclusions from the paper. If you should be mad at anybody it should be Lipshitz!!
I honestly don't care if Lipshitz was right or wrong, I care about your attack on JA's character without a shred of evidence of wrong doing. A misquote is not evidence of wrong doing!!
The sad truth is that you can't even make sense out of the Lipshitz quote, which doesn't distinguish in the final analysis between tones and music. That was done later on the ABX website. I see now that it is not even your own work you are flashing around.
"First of all, they didn't do that. They tested music and test tones separately and kept scores separately, too. Why are you claiming something like this?
"
Because in the quote the clearly lump them together (84/137), so yes they DID do that!
"The passage quoted doesn't give the score on music, but it does give the score on test tones, and it does give the overall score combining the results on both test tones and music"
Exactly they didn't do that and they gave a quote about achieving a 99% confidence level for the COMBINED data. No mention was made about music alone, saying to the effect that polarity reversal was inaudible on the music. Nowhere is that said!! Lipshitz own conclusions don't support your accusations, unless you want to accuse Lipshitz of wrongdoing now.
"the whole experiment would have been meaningless to everyone who's not spending time listening to test tones."
That is YOUR conclusion and not the one coming from the Lipshitz quote. Again, JA was merely paraphrasing that quote from that paper, he did not have the benefit of the ABX websites breakdown on things and the message implied by Lipshitz's quote is certainly not the one you are trying to make.
"That's pretty evident wouldn't you say? But the point is that audibility could only be statistically observed *if* the 24/24 score on test tones was included as part of the *overall* score. On music, audibility could *not* be established *with any confidence at all.* Go back and review their results once again (or see link for a more detailed analysis of them if that helps).
"
I said its evident from their quote not necessarily in reality. As I was not there for the tests, nor have I talked with Lipshitz personally on the results I cannot comment too much on what they really mean. Apparently the quotes from CJ show that Lipshitz continued to believe in the audibility of these things with music.
The study you posted also be interpreted in different ways. From the pure lumping together of statistics one can conclude that with music polarity reversal is in audible; however, they note that it is VERY listener dependent, in other words there were SOME individuals who COULD readily identify phase distortion on music while most could not. The fact that some individuals could means that the sampling of individuals for testing is CRITICAL to get a meaningful test. Either you have to select very carefully (thus probably not a random sample) or you have to have a LARGE number of samples. The conclusion might very well be that normal descriptive statistics are NOT suitable for listening tests because they cover over significant differences in individuals.
In analytical sciences they teach that sampling is just as important as the measurements themselves for having an accurate and precise result.
For example, if 100 people are tested on 10 trials and 90% can't hear a difference on Jazz and end up guessing randomly we get 50%. So that's 450 right and 450 wrong. However; 10% get 90% right (so 90 out of the remaining 100 trials). Now we get 540/1000 = 54%, which you would claim is not statistically significant and therefore inaudible, right?
But these 10 out 100 people can reliably hear the difference so at least with some people it IS audible and the statistics fail to show this.
This is what the authors in the paper you posted meant by "highly listener dependent". I am sure the same kind of misrepresentation is in the data by Lipshitz and all the other blind tests where normal statistics were implemented.
However; you may not know this but one of the main criteria of conventional statistics is that the data adheres to a Gaussian or "normal" distribution. If it does not, ie. the data is bimodal or even multimodal, has excessive skewing or non-gaussian fit then conventional statistics are not a good choice for meaningful results. Non-parametric statistics must then be used. For example, during my Ph.D. we studied atmospheric aerosols, looking at both the size and chemical composition. The resulting size distributions were Log-normal in shape and bi or even trimodal. It was meaningless to calculate a "mean" aerodynamic diameter and we relied mostly on counting and histograms.
Listening results can be compiled the same way. You make a histogram of the number of listeners who got 1 right the number who got 2 right etc. Then you make a histogram and look at the distribution of correct responses. If this result is not gaussian then I would argue that calculating meaningful statistics (mean, std. deviation, confidence limits etc.) with conventional statistics is pointless. A histogram will tell you if you have a small but significant number of people getting more than a random number of responses correct. Retesting will tell you if it was that 1 in 100 random result of if they can do it reliably. If yes, then at that point you can no longer say it is inaudible but only audible to very good listeners or "special" listeners. Listeners who get 0 right can also imply statistical significance because the chance to get 0 right from guessing is as low as getting 10 right from guessing.
"60 / 113 = 53%. Would you care to point out where I'm doing the rewriting: In the "60" or in the "113" or in the "53%"?
"
Lipshitz never wrote this they wrote 84/137 did he not? Anyway that number, as I have pointed out above is meaningless unless we see how the responses break down by individuals. If the resulting histogram is not gaussian then it is meaningless because it covers over individuals with highly correlated responses with those who really don't hear it and guess randomly. This simple calculation of correct to total responses is a simpleminded treatment of the data and most likely not where the relevant information lies.
"Some paraphrasing: "not audible" is now "is audible"; "never been detected" is now "can be reliably detected," etc. etc."
No that is not what he did. AGAIN the quote from Lipshitz:
"The authors have demonstrated the two-tone experiment described above to numerous people on different systems. No one has ever failed to hear the timbral change with phase, and discern the polarity reversal on this signal with unvarying accuracy. Indeed, in a double-blind demonstration to eleven members of the SMWTMS audio group [13], the accuracy score was 100% on the summed 200-Hz and 400-Hz tones over loudspeakers, and overall, including musical excerpts, the results on the audibility of the polarity inversion of both loudspeaker channels were 84 correct responses out of 137, this representing confidence of more than 99% in the thesis that acoustic polarity reversal is audible.""
They never say its inaudible on music now do they? No, in fact the imply that it is audible on music, having satisfied a 99% confidence limit. If you want to question anything then question this conclusion not JA referral to it.
"Therefore it is indeed possible to say that overall, polarity reversal is audible with 99% confidence; and that it's audible on test tones with better than .000001 confidence; BUT it is NOT possible to say (like you do) that it is audible on music with any confidence at all.
"
No see they say overall and that implies music as well. So, JA's quote is not inaccurate. The conclusion from Lipshitz is suspect; however, plus the simplistic statistical treatment is rather sad given the nature of the tests and a counting statistics would likely yield much more information (particularly of a few individuals who likely scored quite well on tones and music like in this other study you posted the link to, strangely, they also don't look at the individual responses, which makes their (subtle) audibility conclusion suspect as well).
"but what anyone can see is a sustained effort in this thread by especially those with an industry affiliation affixed to their name to camouflage, deny, or suppress through other means the fact that JA has misrepresented important research results in his Stereophile column.
"
Complete BS.
"intentionally obtuse"?
I think at this point we can safely drop the "intentionally"!
ROTFL
Everything matters, don't forget to tweak your placebos!
...so bring in the tauts and the jeers!
Good point!
| FAQ |
Post a Message! |
Forgot Password? |
|
||||||||||||||
|
||||||||||||||
This post is made possible by the generous support of people like you and our sponsors:
Top of Page ]
[ Contact Us ]
[ Support/Wish List ]
[ Copyright © 2025, Audio Asylum, all rights reserved ]
