Re: Christer, This Makes You Look Stupid, And Saying So Makes You Look Even Stupider.....
I know. Didn't I say so? I just questioned that the possible
directedness would make much difference here, since I
got the impression you and Frank do after all consider this
a more subtle effect than the difference between cable types.
Might have got you wrong there. Anyway, if I actually did hear
a difference, and not just imagined it, it would reasonably be
much more apparent when doing the test the right way, as
was intended, right?
Point taken. 🙂
However, I was cautious to at least not claim the test to have
any validity, even when I thought I had done it in the right way.
The main reason for posting it was not so much to report my
possible findings, but rather to show that I am willing to try
things out. Besides, since I suddenly rememberd that the
interconnects actually were made from directed cables, it seemed
tempting to give it a quick try. I will do further test, I promise,
although perhaps not in the near future. My most immediate
plans are to build a better headphone amp, which I am working
on, and to tweak my CDP a bit or two. I think it might be better
to do these upgrades before going further into cable issues.
I do further think this incident illustrated the concept of
psychological bias. I am by nature and profession, usually
careful and sceptic. I don't usually believe somethings just
because I want to believe it. Even if I hear a difference, I
question whether it is real or imagined. Yet, despite there
most likely not being any difference due to my error, I managed
to convince myself that I possibly heard a difference.
BTW, you should not trust anybody you just know over the
internet, unless you have some additonal information/reason
to back up their credibility. I don't believe much in the concept
of authorities. Basically everybody errs once in a while.
Remember, even Einstein cheated with his equations in the
50's since he couldn't believe the results. Incidentally, the
results were proven correct after his death. (Sorry physics
guys if details are not correct here).
BTW 2, what is your and others take on doing the inteconnects
in the way I described, using dual coaxial for signal and ground
respectively?
Edit: I had missed the subject line of your post. I hope I am
misunderstanding you. As I read it you seem to suggest I
should have shut up about my error, to avoid making a fool
of myself. I rather admit an error and make a fool of myself,
than delude other people who might have believed in what
I falsely claimed. I confess I used to have a problem with
admitting errors when I was younger, but I have been working
on that and try to always admit my errors, provided I am
convinced they are, or possibly could be, errors.
mrfeedback said:
I know. Didn't I say so? I just questioned that the possible
directedness would make much difference here, since I
got the impression you and Frank do after all consider this
a more subtle effect than the difference between cable types.
Might have got you wrong there. Anyway, if I actually did hear
a difference, and not just imagined it, it would reasonably be
much more apparent when doing the test the right way, as
was intended, right?
Point taken. 🙂
However, I was cautious to at least not claim the test to have
any validity, even when I thought I had done it in the right way.
The main reason for posting it was not so much to report my
possible findings, but rather to show that I am willing to try
things out. Besides, since I suddenly rememberd that the
interconnects actually were made from directed cables, it seemed
tempting to give it a quick try. I will do further test, I promise,
although perhaps not in the near future. My most immediate
plans are to build a better headphone amp, which I am working
on, and to tweak my CDP a bit or two. I think it might be better
to do these upgrades before going further into cable issues.
I do further think this incident illustrated the concept of
psychological bias. I am by nature and profession, usually
careful and sceptic. I don't usually believe somethings just
because I want to believe it. Even if I hear a difference, I
question whether it is real or imagined. Yet, despite there
most likely not being any difference due to my error, I managed
to convince myself that I possibly heard a difference.
BTW, you should not trust anybody you just know over the
internet, unless you have some additonal information/reason
to back up their credibility. I don't believe much in the concept
of authorities. Basically everybody errs once in a while.
Remember, even Einstein cheated with his equations in the
50's since he couldn't believe the results. Incidentally, the
results were proven correct after his death. (Sorry physics
guys if details are not correct here).
BTW 2, what is your and others take on doing the inteconnects
in the way I described, using dual coaxial for signal and ground
respectively?
Edit: I had missed the subject line of your post. I hope I am
misunderstanding you. As I read it you seem to suggest I
should have shut up about my error, to avoid making a fool
of myself. I rather admit an error and make a fool of myself,
than delude other people who might have believed in what
I falsely claimed. I confess I used to have a problem with
admitting errors when I was younger, but I have been working
on that and try to always admit my errors, provided I am
convinced they are, or possibly could be, errors.
Welcome To Club Idiot.............
LOL, LOL, LOL .............
Christer, be assured that my post is intended in the sincerest good light.
Also rest assured that if you dish it out like you have done here, and then make a mistake like you have admitted, you will be drawn and quartered, and cruelly and mercilessly so.
You have actually performed a very worthwhile experiment in that (with further practice) you should in future be able to divorce the 'expectation' factor that I have spoken of and instead concentrate on actual sonics.
Questioning one's own ears and repeated A/B experiments are the ingredients to growing "golden, or even platinum ears?" - your quote.
Intelligent practice will hone your listening skills, and empower you to further trust your own ears, and for that matter understand other listeners ears.
Repetition and variation build experience and diffrentiation.
Eric.
PS - There is an old saying about how it is sometimes better to keep ones mouth shut and and allow others to think that one is an idiot, rather than open one's mouth and remove all doubts.
LOL, LOL, LOL .............
Christer, be assured that my post is intended in the sincerest good light.
Also rest assured that if you dish it out like you have done here, and then make a mistake like you have admitted, you will be drawn and quartered, and cruelly and mercilessly so.
You have actually performed a very worthwhile experiment in that (with further practice) you should in future be able to divorce the 'expectation' factor that I have spoken of and instead concentrate on actual sonics.
Questioning one's own ears and repeated A/B experiments are the ingredients to growing "golden, or even platinum ears?" - your quote.
Intelligent practice will hone your listening skills, and empower you to further trust your own ears, and for that matter understand other listeners ears.
Repetition and variation build experience and diffrentiation.
Eric.
PS - There is an old saying about how it is sometimes better to keep ones mouth shut and and allow others to think that one is an idiot, rather than open one's mouth and remove all doubts.
The keymaster
Eric anticipated my answer: in tests like this, the key sheet should be left in the hands of a third party. I'd be happy to do that, but since I live near Steve and have expressed a position here, I'd rather have someone more neutral hold the key and announce the scoring. It's not that I'd cheat (I wouldn't), but I don't want there to be any question.
What I would like, however, is a couple of the wire lengths, with the "direction" marked, so I can examine them for the possibility of non-auditory clues (e.g., die marks, witness lines...).
Eric anticipated my answer: in tests like this, the key sheet should be left in the hands of a third party. I'd be happy to do that, but since I live near Steve and have expressed a position here, I'd rather have someone more neutral hold the key and announce the scoring. It's not that I'd cheat (I wouldn't), but I don't want there to be any question.
What I would like, however, is a couple of the wire lengths, with the "direction" marked, so I can examine them for the possibility of non-auditory clues (e.g., die marks, witness lines...).
The 'expectation effect' is not volitional, and it is universal among humans. The idea that one can learn to tune it out is a construct of dishonest audio writers. This concept is unknown in any other areas of sensory research.
Hi,
Alright then, we need a volunteer. Anyone?
Suggestions?
Or shall we appoint one, as in the army...😉
No objection to that from me.
If you do this kind tests long enough the expectation factor does wear off rapidly though.
Cheers,😉
Alright then, we need a volunteer. Anyone?
Suggestions?
Or shall we appoint one, as in the army...😉
No objection to that from me.
If you do this kind tests long enough the expectation factor does wear off rapidly though.
Cheers,😉
Peer Group Reviewers..........
"Alright then, we need a volunteer. Anyone?
Suggestions?
Or shall we appoint one, as in the army..."
Yeah, Jocko, Fred, John Curl..... and any other volunteers.
"No objection to that from me.''
Yup from me too.
"If you do this kind tests long enough the expectation factor does wear off rapidly though."
Yeah, diminishes to zero eventually.
Eric.
"Alright then, we need a volunteer. Anyone?
Suggestions?
Or shall we appoint one, as in the army..."
Yeah, Jocko, Fred, John Curl..... and any other volunteers.
"No objection to that from me.''
Yup from me too.
"If you do this kind tests long enough the expectation factor does wear off rapidly though."
Yeah, diminishes to zero eventually.
Eric.
fdegrove said:
Oh, could be anything from simple malice to ego and insecurity.
Let's say I had some deeply held belief that there could not possibly be anything to directionality. Well, I might not be all that secure in such a belief and not willing to accept possible truths which would conflict with that belief and sabotage the test to keep from having to face up to such a possible truth.
I've no malice nor any particular beliefs one way or the other on this matter other than my belief that I haven't seen any convincing objective evidence of actual audibility. However you and Eric can't possibly know with any absolute certainty what might be going on in my mind.
Because I'm wanting to do everything I can to make this test as thorough as possible in order to avoid as many possible pitfalls as can reasonably be foreseen, I anticipated this possibility and wanted to give you and Eric the opportunity to opt for a third party.
The statistics should be fine as long as you're confident in the performance of the reference pair, i.e. that the effect is significant enough that you haven't any trouble making the distinction.
se
fdegrove said:
This is a claim that could easily be tested. Take some difference which everyone agrees is just barely audible -- maybe make a bunch of copies of a CD, with some sent through some subtle but audible filter. Label the CDs, some correctly and some not, and let the test subjects see the labels. Then run whatever test you want with them so that the test subjects classify them as filtered or raw, and check to see whether the incorrectly labeled ones are classified accurately as often as the correctly labeled ones.
If someone could demonstrate just this, I'd be a lot happier to believe some of the other claims being made about particular differences in the absence of DBT-ing.
mrfeedback said:
Thank you. I have faith in that as well, but I want to do this test as thoroughly as possible and so I tried to anticipate as many possible pitfalls as possible and thought it was only reasonable to give you and Frank the option of a third party.
Yeah, that'll take a lot more work. But the most significant and exciting of all will be actually establishing audibility. It would be, to put it mildly, groundbreaking. And while I have no particular beliefs on this issue one way or the other, I do get a bit bigger thrill when the underdog comes out on top. Which is why I want to give this test every honest chance of succeeding as possible.
Of course not.
That's what we're here to find out.
That directionality manifests itself best under the conditions you noted, i.e. one channel wired opposite the other, the beauty part of this test is that it can be done under the exact same conditions that you would normally experience it and best of all without the need for any sort of A/B switching so we eliminate any issues of memory, or switching speeds, or switches in the signal path, etc.
The ONLY difference will be that you won't know how each pair is wired so that we eliminate any psychological bias brought about by sighted listening. In other words, it will establish that if you're able to accurately identify each pair, it will mean that you did so based on sound alone.
No, null results are null results. They're inconclusive.
What's the difference? You said that you've heard the difference when one channel is wired opposite the other. That's precisely what you'll be presented with here. The only difference is that you won't know how the test pairs are wired. If the effect is actually audible, you won't need to know.
If you can only perceive directional differences under specific circumstances, then you'd better lay out clearly what those specific circumstances are before we go any farther.
Yes. Actually I was already thinking of a way that even though I'm in California and you're in Australia and I won't be present during the actual listening, we can make this truly double blind.
It would work like this:
I would randomly number the test pairs and notate which numbers corresponded to which pairs (i.e. whether the pair was wired the same or oppositely). I would then send the test pairs to a second party who has no knowledge of how I had numbered them and they would in turn randomly number my random numbers, notating which of their numbers corresponded to my numbers.
Myself and the second party would seal our notations in envelopes and send them a third party who would keep hold of them while the tests were being done. After which time the third party would open the envelopes and post the notations publically and we could then analyze the results.
That way NONE OF US will know how the cables you're using for the tests were ultimately encoded.
se
Re: The keymaster
See my post to Eric about actually doing a double key system.
Certainly.
se
SY said:
See my post to Eric about actually doing a double key system.
Certainly.
se
By the way, Frank, how do you want to do your listening tests? Do you want to use terminated cables or do you want bare ends that you can use with your phono stage as you described previously?
se
se
Maybe you have mentioned it already and I have missed it,
but how many cables will the listeners get and has anyone
tried to calculate what statistical significance is possible to
get. Depends on how many rights they get, of course, but
it might be useful to know in advance what significance level
can be ascribed to
various percentages of right answers. You seem to
actually do real research here guys, so it would be a pity to
spoil a positive result by having used too few cables despite
the percentage being high. If the result turns out positive with
a sufficient significance level it should be properly scientifically
published.
PS, I am lousy at statistics and being a theoretical researcher
I don't do empirical studies, but AFAIK the significance level
for a certain percentage of correct answers will increase with
the number of samples, which is why I posted this.
but how many cables will the listeners get and has anyone
tried to calculate what statistical significance is possible to
get. Depends on how many rights they get, of course, but
it might be useful to know in advance what significance level
can be ascribed to
various percentages of right answers. You seem to
actually do real research here guys, so it would be a pity to
spoil a positive result by having used too few cables despite
the percentage being high. If the result turns out positive with
a sufficient significance level it should be properly scientifically
published.
PS, I am lousy at statistics and being a theoretical researcher
I don't do empirical studies, but AFAIK the significance level
for a certain percentage of correct answers will increase with
the number of samples, which is why I posted this.
Christer said:
I was planning on doing 5 trials with 20 pairs for each trial. So we're talking a total of 100 evaluations.
Statistics isn't my strong suit but I would think that out of 100 evaluations we should be able to get a reasonably high level of confidence.
However before this is all over I'm going to try and enlist the help of a statistician to give an analysis of the results, unless it comes up 50/50.
Agreed. As I'd stated before, if the results turn out positive, I'd be more than happy to contribute to writing up a paper to submit to the AES.
Exactly. Which is why I wanted to shoot for 100 evaluations total. I mean, getting 10 out of 10 is statisticially significant, but doesn't have all that high a level of confidence. And any positive result will be highly scrutinized so if we're going to put forth the effort, we should take it as far as practical.
se
Hi,
ASAIK, bare wire is good enough.
Keeping in mind statistical evidence targets, I'd rather you'd ask Eric as well.
We should both run test in a similar way to keep things ...errr...even slightly remotely scientific.
I don't want you to go through too much expense but if terminated it should be terminated the same way for both particiants or the test would be void.
Cheers,😉
ASAIK, bare wire is good enough.
Keeping in mind statistical evidence targets, I'd rather you'd ask Eric as well.
We should both run test in a similar way to keep things ...errr...even slightly remotely scientific.
I don't want you to go through too much expense but if terminated it should be terminated the same way for both particiants or the test would be void.
Cheers,😉
Hi,
Christer, would you mind playng keyholder?
Reading your post, it just struck me...this kind of tests is just your academic turf isn't it?
Cheers,😉
EDIT:
And here again Christer's academic experience may help us on how to set u a valid test...
Assuming he's interested that is?
Say yes? Please? On his knees now...Yes please???
O.K....now you all have solid proof of me proposing to a man....
Kidding,😉
Christer, would you mind playng keyholder?
Reading your post, it just struck me...this kind of tests is just your academic turf isn't it?
Cheers,😉
EDIT:
And here again Christer's academic experience may help us on how to set u a valid test...
Assuming he's interested that is?
Say yes? Please? On his knees now...Yes please???
O.K....now you all have solid proof of me proposing to a man....
Kidding,😉
fdegrove said:
Let's, were I then supposed to do the renumbering or just keep
the records of the renumbering process? Yes, I guess I could
do that, if Eric doesn't mind (I have had some mail contact
with him and think we have a peace agreement, even if we
disagree about ccertain things).
It is not exactly my type of professional stuff. I am a theoretical
researcher, so empirical studies is not my strong side. I construct
mathematical proofs. However, that should hardly matter for
this function in the experiment. The important thing is that
someone who does know how to do empirical research
thinks all this through, so we are not missing anything that
might spoil the result. SY, don't you know this stuff?
Steve Eddy said:
If all evaluations were independent (which they won't be), the chance of getting 10 out 10 right strictly by chance is 0.1% -- that'd make me fairly confident. If you're just guessing and have a 50-50 chance of guessing right on each evaluation, then there's only a 5% chance of getting 8 or more out of 10 right, or 58 or more out of 100.
Since the evaluations won't be independent, I'd think that a logistic regression might be a good way to go. I'm not a statistician, but if you can't fine one I'd be happy to chew on the data.
Hi,
Eric, won't mind for sure...he's a very nice fellow, one I'd like to have as a neighbour... 😉
Sorry, my bad then...
Let's call on Red Chief SY to set up the academically acceptable environment then ...
With only two particapants, I'm not even sure we could even meet anything even remotely academic...
Cheers,😉
Eric, won't mind for sure...he's a very nice fellow, one I'd like to have as a neighbour... 😉
Sorry, my bad then...
Let's call on Red Chief SY to set up the academically acceptable environment then ...
With only two particapants, I'm not even sure we could even meet anything even remotely academic...
Cheers,😉
fdegrove said:
I don't think he would mind either, but he still seems not to
understand my point of view or why I question what you
guys claim. Still want him to say it's OK.
Oh indeed we can. We may be able to prove that there are
at least two individuals who are able to hear the direction of
a cable under certain circumstances, at a certain
significance level. That would probably be major breaktrhough,
as Steve has said. It won't say anything about what percentage
of the population can hear it, or if women are better at hearing
it (too obvious in this case, of course 🙂 ), etc. But proving it
can be heard at all by some individual is IMO the most
important result. Then, of course, one may conduct an endless
number of studies, testing various types of cables, use them
for different functions in the system etc. etc.
An interesting thing with this experiment is that you and Eric
cannot cheat, since it is in you interest to get a positive result.
If, on the other hand, we were to test an arbitrary group of
people, some may fake a negative result due to psychological
bias. That is, positive results are free from psychological bias,
while negative results aren't, as far as I can see.
fdegrove said:
Ok, but the type of wire will need to be the same for both tests. Will the Vampire magnet wire I mentioned do for you? It has a solderable polyurethane enameling.
Mmmm. Well yes, if the wire's going to be terminated for one it should be terminated for the other and vice versa. Good point.
Ok, so we're going to need to get you and Eric to agree on terminated or unterminated.
However keep in mind that we only need to establish audibility for one person and my goal has been to have you each do the listening under conditions most favorable to each of you, which may not be the same.
Agreed. Thanks for bringing that up.
Perhaps Eric would be amenable to using unterminated wire as well as the RCAs would constitute another variable and leave some ambiguity as to whether it was the wire or the terminations that were responsible.
Though even with that ambiguity, actually establishing audibility of directionality whether with or without terminations would be just as profound. You just wouldn't be able to safely conclude that it was due to the wire.
Again, my goal is to have you each do your listening under the most favorable conditions.
se
- Status
- Not open for further replies.