|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
In Reply to: Well, if you run one subject, and they control the switching... posted by jj on December 02, 2002 at 20:59:20:
"I haven't seen the problem with trained subjects."That doesn't mean it doesn't happen, or is always detected, does it.
Follow Ups:
And if you think that a lack of performance on the part of a subject can't be detected, well...It's time for some specifics from you, thank you.
JJ - Philalethist and Annoyer of Bullies
What about you. Seems to me the few published litening test have been picked apart by several people for various reason, and yet you persist on doing the same thing over again, which will just be picked apart again....
Seems to me the few published litening test have been picked apart by several people for various reason...Isn't that what the scientific method is all about? Provide the details by which a given test can be evaluated? You posted a link about a guy testing the audibility of phase who used some mediocre pro speakers as his "reference". Further, he stated that he assumed for the purposes of the test that they were phase linear. Interestingly, he stated that the audibility varied greatly with the listener. Duh.
Bogus tests prove nothing.
"You posted a link about a guy testing the audibility of phase who used some mediocre pro speakers as his "reference""Huh? Please refresh my memory on this. The only link coming close is one by a fellow(Scot's Guide) who discovered a flaw in another person's test, where they used a different load impedance to change the power level, instead of increasing the amplifier output.
He never ran any test, but rather did a mathematical analysis to show what was measured by the other party.
"Facts are stubborn things; and whatever may be our wishes, our inclinations, or the dictates of our passions, they cannot alter the state of facts and evidence." --John AdamsI agree with this quote found on another site you referenced, "The Scientific Art of Audio". True, except for when the evidence does not support the alleged "facts". I particularly like the section on "The truth about esoteric power cords". Gee, those pics do offer some compelling "facts"! The phase test is linked from here as well.
http://www.fortunecity.com/skyscraper/motorola/145/
rw
If you have a problem with the work, write to the author/school and see what response you get.
I have far better ways to spend my time than to debate with an inept grad student. Like enjoying music.Daisuke needs to spend a couple of years listening to music on what are truly state of the art audio reproduction systems before he has any understanding of what he is trying to prove. He is at present light years away from that reality.
rw
Yeah, right.
you have not had the good fortune to hear a system where such claims of inaudibility are rendered moot.Time to fall back on the "scientific mantle" that cites no sources nor reveals any experience with truly SOTA equipment. Let the untrained ear guide your way. What does Motorola have to do with high end audio anyway? The rigors of family radio reception?
rw
It's truely amazing that you think anecdotal evidence is credible.
Not to mention subjective opinion.
Since no component is perfect, one necessarily has to make choices based on preference. My preferences have themselves evolved over a long period of time. The fact that differences exist is not the question for experienced listeners using high resolution systems. The question is which differences are most important to any one listener. Is image specificity more important than harmonic integrity? Is stellar bass response more important than dynamic accuracy? There is certainly no one answer.So when one speaks of the clearly audible differences among components (using musical content, not static test tones) where the numbers are not supported by the observational data, then which conclusion do you draw?
1. The consistently perceived differences among a large sample of folks (thousands of high end component customers) are completely attributable to merely psychological factors.
OR
2. The testing methodology does not mimic the musical reproduction process closely enough to be relevant. You pointed out the challenge requiring powerful supercomputers to handle testing using complex musical content (which, by the way is how audio equipment is used by the public). Likewise, you have identified the other challenges of conducting DBT tests offered by jj. Your solution? Dumb down the tests to make them more convenient . The result of the simplification process is to make the tests largely meaningless.
Perhaps there are some valid DBT tests that do not contain the flagrant flaws evidenced in the links posted thus far. We're all waiting to see them!
rw
"Your solution? Dumb down the tests to make them more convenient. "I certainly don't know where you get this. There is never any convienient method of handling the human factor in these types of tests.
As for the the link, the phase response of the speakers would have been nice to know, but every speaker, has some phase/group delay anomoly. As long as the test was conducted with the exact same set of speakers in the exact same environment, then the "flaws" are consistent throught the testing and therefore have a minimal impact. Had either the testing environment or the speakers changed, even swapping them from left to right, then you would have intoduced a variable that would have to be seperately accounted for.
As to the quality of the speakers, is mearly a subjective opinion on your part, not to mention your comments regarding what the author should be doing with his time. Whether you want to believe it or not, those "mediocre" speakers have very likely been used in many, many studios to make mixdown choices.
Here are four such examples quoted from your posts.DUMB DOWN # 1 - Make the tests easier to conduct, not more accurate.
When asked about running multiple subjects simultaneouly, you said:
You'll never get satisfactoy statistical result without enough participants. You'll never be able to run enough participants in your lifetime without some other methodology.
So let's not bother with considering the problems with this approach, shall we? We'll go with quantity instead of quality.
DUMB DOWN # 2 - Utilize simplistic static tones as test criteria
You also have to understand that the simpler a test the better... It is quite possible to make a test signal so complex that a room full of Cray computers can't properly analyse it.
Fine. Those tests are entirely relevant for those souls who spend their leisure time listening to test tones. Does that fully characterize the performance differences when reproducing musical content on high resolution gear. I think not.
DUMB DOWN # 3 - Use testers unable to discern fine differences
This one is simply amazing:
All the tests I've participated in did not require training.
I guess it all depends upon what you are trying to prove. If you are trying to determine the audible characteristics of a family radio with the general populace, then this probably works fine. If, on the other hand, you are trying to establish ultimate audible effects of the highest resolution audio equipment, then you must have trained listeners.
Do you know how the suspension designs of exotic performance cars are finalized? While the initial modeling is done by computer they then create prototypes and perform exhaustive experiments using professional test drivers. I can imagine the laughter by Ferrari or Porsche, etc. if you were to suggest that these changes be dialed in by some guy off the street.
DUMB DOWN # 3 - Use completely unfamiliar musical material as the comparative reference
This one came as a surprise:
I suppose you also think that it is okay for the participant to be familiar with the test material,
Absolutely! Here again, what is your motive? It takes me a couple of weeks to fully grasp the musical content of a new album and explore all the nuances of a fine recording. Flash uncontrolled sections of new material at someone and indeed they are not going to hear a lot of diferences.
DUMB DOWN # 4 - Use mediocre equipment that is not "the most accurate audio equipment available"
As long as the test was conducted with the exact same set of speakers in the exact same environment, then the "flaws" are consistent throught the testing and therefore have a minimal impact.
Those flaws present are capable of completely masking the results. If you try to measure the ultimate cornering capability of a Dodge Viper on a set of $100 Pep Boys tires, you are not going to get an accurate picture of that vehicle's performance envelope.
It is easy to understand why you or your testers believe that there are few audible differences based upon how you've crippled the tests!
You're taking things completely out of context to try and build yourself a case. Sorry, I won't bite. Next time you're at Disney, say hi to Bugs.
While at Epcot, the wife and I stayed for the big 9:00 music and fireworks celebration around the lake. While the fireworks were impressive, the sound was dreadful. You'd think Mickey could afford better.The response certainly clarified your opinion advanced in previous posts on testing precepts and methodology.
rw
Yes, those pictures do. Most likely that wire is hot-dipped galvanized steel. A far cry from cryo treated OFC copper.....blah, blah, blah.So applying Ocams Razor to the situation and it isn't hard to realize what affect an esoteric power cord is going to do to improve power transfer into your equpment. Or is that conclusion to hard to derive?
Thanks for the other reminder. When the opportunity presents itself, I'll have to re-read the paper. In general, it would be a mistake to not characterize the equipment being used for the test, but again, I need to see what the context is here that you're complaining about.
Yes, those pictures do.Do what? They certainly do not prove that the use of esoteric power cords with high resolution music playback systems is inaudible.
Or is that conclusion to hard to derive?
You seem to treat this subject as though it was theoretical astrophysics where experiential proof is impossible. Here on planet earth, it is quite easy to test the efficacy of different power cords on high resolution music playback systems.
I need to see what the context is here that you're complaining about.
Here are some quite amusing excerpts from the study:
"It is paramount in conducting listening tests with the most accurate audio equipment available if the test subject is to hear any audible effects."
Agreed. Ok, so why use a dinky low end "pro" mini monitor as the choice for "the most accurate audio equipment available"? Already the guy has invalidated the study using his own criterion. The title of the study is "Audibility of Phase Distortion in Audio Signals". Did the tester carefully conduct phase tests on the test equipment itself to verify that the results were independent of the equipment used?
"Phase characteristics for this loudspeaker were not investigated...Again, the phase response for the headphones was not investigated."
I sure hope the tests that you rely upon do not possess such obvious flaws. Back to the drawing board Daisuke!
rw
"Yes, those pictures do.Do what? They certainly do not prove that the use of esoteric power cords with high resolution music playback systems is inaudible."
All credible audio reviewers (unlike you) fully state all of the equipment and musical content used for an evaluation. Any trained listener can tell the difference on a high resolution system. We're talking more than a Bose radio here Bruce.
There are no peer reviewed objective tests anywhere that bear this out. Only subjective ones that fall under the category of anecdotal evidence.
But you won't provide evidence.You have come up with "problems" in the "normal" way that people run things but you won't provide evidence for that, either.
What's more, you seem unable to understand how one can evaluate the performance of a subject relative to known thresholds.
Why don't you submit a paper to the next US AES convention?
JJ - Philalethist and Annoyer of Bullies
Ya know, you're beginning to sound like some of the others around here with all the guessing you're doing.Starting with "my methods" - never, ever said that, made it clear numerous times. Start over.
You're putting forth the test, and without providing a single detail, dismissing everything everyone says to you. So you've taken on responsiblity personally for these ideas, and you can live with it.Now, you have yet to offer any evidence about subjects having trouble with familiar material. On the converse, there are reams of data showing than unfamiliarity with the entire setup, music, etc, desensitize test.
You claim that there is no problem with running multiple subjects. On the converse, there are reams of data showing:
1) That multiple subjects affect each others' scoring.
2) That an inability to do a clean, fast, listener-controlled switch desensitizes a test.
3) That primary auditory memory starts to fade inside of 200 milliseconds.How, then, can you run fixed-switched tests that do not allow time-proximate comparison (where "proximate" means 200 milliseconds) and have sensitivity.
You have made some very strong claims, and you won't provide any evidence, yet you whine incessantly when others point this out.
If you want to convince anyone that this test you talk about does anything beyond run a desensitized test that will not get to already-measured JND's you have to do better.
Lots better.
Like I said, submit a paper to the 115th AES on this.
JJ - Philalethist and Annoyer of Bullies
" and you can live with it."I'm not the one with the problem, so it's just fine by me. And furthermore, my employment is worth far more than a speaker cable test.
"You have made some very strong claims, and you won't provide any evidence, yet you whine incessantly when others point this out."
Oh bullshit. You likewise have made some strong claims and haven't backed up anything you propose either. You haven't even answered the one question of bias I asked.
"If you want to convince anyone that this test you talk about does anything beyond run a desensitized test that will not get to already-measured JND's you have to do better."
So far, you have no hard evidence either, the published cable listening tests have all been criticized in one way or another, inculuding not having enough statistical evidence. (Even if I don't agree with their arguments against the results) Yet you have just persisted on touting that we do the same thing. Some say that's the definition of insanity. Repeating the same thing and expecting different results.
Been nice jerking your chain jj. Throughout this, you've shown to be stuck in the same dogmatic rut that the "Annoyer of Bullies" likes to poke at.
Maybe now I know where those stereotypical "ABX'er" comments originated.I've stated the common understanding. The evidence is visible in the AES, ASA, various psychology journals and texts, psychometry journals, and the like. It's all over the place.
Your claim about statistical size in cable tests begs the actual hypothesis entirely. It only takes one repeat performer to push that performer's p1 and p2 so far down as to be sure they detected something.
I haven't CLAIMED that the tests represent the entire population, and neither have many other people, so making that criticism is a waste of time.
The evidence is clear: Untrained listeners don't hear as much.
The evidence is clear: Tests that are not time-proxmiate are not as sensitive.
The evidence is clear: Multiple subjects cross-influence and add noise to the test.
Those are well-demonstrated in the literature. You must show that the test you espouse can do as well, despite the long list of demonstrations otherwise, before you have any credibility.
WHERE IS YOUR EVIDENCE?
JJ - Philalethist and Annoyer of Bullies
Still won't admit that no current test is perfect, will you?
As you and everyone else knows, you have materially and unquestionably misrepresented my position.Your statement constitues a full and explicit accusation of professional misconduct on my part.
I require a full and complete retraction of the false position you have laid at my door.
Your trolling is starting look quite familiar, and your willingness to create false positions speaks ill of you.
JJ - Philalethist and Annoyer of Bullies
Oh bullshit. Like you haven't done the same to me. Get over yourself.
Your claims go very strongly against many people's experience.In particular the length of auditory memory simply destroys any chance of having a good result without proximate switching that is listener controlled.
And, of course, the "mob aspect" will rule as well.
If you want credibility, put up a paper for review.
You're starting to look like a straw man working for the anti-DBT camp in my mind.
JJ - Philalethist and Annoyer of Bullies
"Your claims go very strongly against many people's experience."Sorry, it's fact. Probably been done other places too, but again, because it is proprietary you will find no publically available documentation.
"And, of course, the "mob aspect" will rule as well."
Mob? Whose mob? Yours for not wanting to explore the possibility that in order to satisfy all the DBT critics, you'll have to better than what you are proposing now? The fact that there may be labs which HAVE overcome the difficulties with the published data? The only fact here is that your mind is closed to the possibility that some of the published difficulties have been overcome and you have no interest in persuing it.
"In particular the length of auditory memory simply destroys any chance of having a good result without proximate switching that is listener controlled."
Okay, so you can EXACTLY remember a complex auditory input for 10 minutes or more?
"If you want credibility, put up a paper for review."
Are you really that dense? One, I have made it abundantly clear multiple time that this is NOT my doing. Two, it is PROPRIETARY, I have also made this abundantly clear multiple times. Does someone need to knock it into your head with a hammer and chisel?
"You're starting to look like a straw man working for the anti-DBT camp in my mind."
-Mob? Whose mob? Yours for not wanting to explore the possibility
that in order to satisfy all the DBT critics, you'll have to better than what you are proposing now? The fact that there may be labs which HAVE overcome the difficulties with the published data? The only fact here is that your mind is closed to the possibility that some of the published difficulties have been overcome and you have no interest in persuing it.-Persuing what? That's a lie. You haven't offered anything TO persue, only a bunch of unsubstantiated, extremely extraordinary claims. You have contradicted some of the most strongly demonstrated principles of auditory testing, but you won't even offer anecdotal evidence, let alone any extraordinary proof of your extraordinary claims.
First, it's you who claims that the test you report on, but won't
take responsibility for (even here) is perfect. I'm not making novel claims, you are. You claim an advance, but you won't provide any evidence to evaluate.Second, the mob aspect of multiple subjects is not a conjecture, it's
a known, done deal. If you're going to use multiple subjects you're
going to have to show some major evidence that you've come up with a way to completely avoid that.Then you deceptively state - Okay, so you can EXACTLY remember a
complex auditory input for 10 minutes or more? -As you are well and truly aware, unless you don't even read what I
write, I am the one asserting that 200 milliseconds is the far limit
for comparisons of small acoustic differences. I have no idea where your "10 minutes" comes from, nor why you have deceptively implied that I claim any such thing.No, you can't recall that long. That's why any test that does not
allow each and every listener to switch AT WILL is extremely suspect.
You've indicted your own test. It's your test that has a delay between the similar parts of the same presentation, not mine. Couldn't you at least get your story straight?Finally, you ARE making claims here, and extraordinary ones. Claiming
that it's not your test, that it's proprietary, and the like, are
lame, weak excuses. You've described the test, so it's not proprietary OR a trade secret. You've made claims about its sensitivity, so you've obligated yourself to support those claims,
so:Submit a paper, and we'll see what comes of it.
You made the claims, now deliver the evidence.
You've made extraordinary claims, and used multiple instances of
extremely deceptive "logic" in your defenses. To wit: You imply,
completely without justification, that somehow 10 minutes of detailed
partial loudness memory are required for time proximate testing, when
that is just the opposite of the truth. You state baldly that I am
claiming that "tests are perfect" when in fact you are the only one
here implying that, and for your test. You won't say when or where these tests happen, so we can't confirm your claims that way. You won't even use your real name, and deny yourself even that bit of credibility.You may know something, but your presentation suggests only that
you're talking (if your own description is correct) about a seriously
insensitive test. You will put up no evidence, but you repeat the same empty claims over and over again.You appear to have nothing to offer. Write the paper, and get it peer-reviewed. Until you do, I'm afraid that I'll have to regard your claims as specious.
Frankly, you read like a straw-man instantiation of the subjectivist's claims about people who do DBT's.
JJ - Philalethist and Annoyer of Bullies
Maybe you should just go revisit some recent history:http://www.stereophile.com/showarchives.cgi?141
And then maybe you'll come to realize that the same old same old ain't gonna cut it. That's all I've been trying to do JJ, is get you to realize this, but you are so damned entrenched you can't see it.
Put up or ...
JJ - Philalethist and Annoyer of Bullies
This post is made possible by the generous support of people like you and our sponsors: