Beside ABX there are other test protocols like ABC/HR and MUSHRA for example. Well known and accepted by professionals. Even a paired preference test would be accepted if done double blind.
But of course the experimentator must be willing to put his own beliefs (about test protocols) under test to see what happens.
And, regarding your last sentence, you have to try sometimes even the unexplainable to find something new. As history of science taught us, a lot of the amazing new inventions was found by accident.
And funny enough, a lot of steps forward were made even if the theory was plain wrong at that time.
Wishes
What keeps amazing me is that there are audiophiles who, after many years in this hobby, still think that they would manage to prove differences between cables in DBT.
As I said before, I can hear differences when listening unblinded, but they simply disappear in DBT. I also made a few sets of cables based on what I believed should sound good but not all of them sounded good. The point is that I didn't have (at least consciously) preference for any of them before listening test, but I managed to hear difference when listening. 😕
As I said before, I can hear differences when listening unblinded, but they simply disappear in DBT. I also made a few sets of cables based on what I believed should sound good but not all of them sounded good. The point is that I didn't have (at least consciously) preference for any of them before listening test, but I managed to hear difference when listening. 😕
I've tried the 'random coin flip test' once with a cable directionality test, I don't like it. It happened to me that the cable stayed the same direction for four or five consecutive tests which tends to take away your 'reference'.
I would suggest marking the cables A and B, the listener ask for a change and then decide which cable is A (name of cable) and which is B. The A and B can be changed randomly. What do you think?
.
You just described the ABX test. You can listen to A, to B, but then you need to decide whether X is A or B.
jd
No, he described a paired (preference/identification) test; the assignment of "A" and "B" to each cable will be randomly choosen in each trial.
Wishes
It depends on the experience/training of the participant(s). It is possible to overcome the "detection of sameness problem" (at least that´s my experience) but it takes some time.
In any case i favour the A/B preference/identification test scheme.
But a positive control should be included as well. 🙂
Wishes
Last edited:
I like the 'Paired (preference identification)' test scheme. That is normally how we make our blind tests. Works for me.
There is no "positive control" for magic.
Remember, the claim is that factors where NO-one has shown ANY discrimination are claimed. Positive controls (e.g., frequency response) are plain and simple irrelevant; a positive control is a stimulus of the factor being tested but at a level where discrimination has already been demonstrated. You again miss the point of the test- these listeners claim that they ALREADY can hear the difference between A and B. The tests are NOT being done on an unknown system in an unknown environment, they are being done PRECISELY in the system and environment that the listener claims is amenable to hearing those differences.
That red herring is beginning to smell.
Steve: If TG1954 drops out for any reason and you're somewhere I can reach in a short hop on Southwest (where are you?), you're the First Alternate.
But I'd sure like it if John (whom I know, like, and trust) would be willing to step up to the plate instead of make excuses. And (to John), if you want to team up with someone else, since you use other people's ears from time to time, that would be fine, too, as long as everything is done double-blind.
But I'd sure like it if John (whom I know, like, and trust) would be willing to step up to the plate instead of make excuses. And (to John), if you want to team up with someone else, since you use other people's ears from time to time, that would be fine, too, as long as everything is done double-blind.
They're switched manually, same as when the claimant switches cables- except it's done out of sight.
The whole positive control thing is a red herring for what is being tested here. It's useful in lab tests where general thresholds are being tested; that's NOT what's being tested here.
The whole positive control thing is a red herring for what is being tested here. It's useful in lab tests where general thresholds are being tested; that's NOT what's being tested here.
I am in the Nashville area, but on the side opposite the airport. It is about a one hour drive, and expressway for all but about 2 miles.
It seems strange for me to see so many obviously informed people arguing about something that I have already experienced myself.
I wonder if the actual issue is not whether or not differences are detectable, but whether the perceived changes are for the better or worse?
Always nice to have some younger ears on hand as well, preferably some with quality audio experience and haven't damaged their ears yet 🙂
I wonder if the actual issue is not whether or not differences are detectable, but whether the perceived changes are for the better or worse?
Always nice to have some younger ears on hand as well, preferably some with quality audio experience and haven't damaged their ears yet 🙂
As long as I do not know which is A and which is B, when I listen, I will not be able to tell. This is normal for me, because the music is continuously changing. I can only catch minor differences when I can compare similar moments of music with both examples. To have a blind experimenter behind the scenes mixing the cables up, AFTER original identification is essentially an ABX type of test, not an AB test. Now, A can be arbitrarily assigned, by computer, if necessary, but once it is assigned, then it must always be defined as A when that cable is listened to. We found this out 30 years ago.
Gee John why do you (them/they/we the all knowing overlords of audio) think the differences go away unsighted?
First of all a positive control shows that the test (or the participants of the test) reach a sufficient sensitivity.
If you don´t know, which factor (please remember that in most case differences between two cables can be measured), it would be wise in general to control on the best level yet already confirmed.
Yeah, and your control should show, that the listener still is able to detect something _under_ _test_ _conditions_ .
You carefully avoided to address the first point regarding the "detection of sameness problem" . You know that listeners in blind tests answer this for ~70-75% wrong. You don´t believe that it could happen in a known system? Fine, nothing wrong with your belief, but put your belief to a test.
Please remember that TG doesn´t claim any "magic", just that he hears a difference between two cables in _sighted_ tests.
What is your problem with some control trials on different levels to see, what he is able to hear _under_ _test_ _conditions_ ?
Wishes
Instead of music, why don't you use constant tones. That would sure make it easier to determine what sort of differences there are. The only drawback being it wouldn't help you decide which is 'better' but as I understand it, that's subjective anyway so a moot point, no?
- Status
- Not open for further replies.