I don't believe cables make a difference, any input?

[john curl]

Bud, is Mille' still alive? He designed the MAC 3500 that we used with the Grateful Dead. It is a wonderful power amp. Much better than standard MAC's. We met years ago. The best amp of its kind, that I know about.

[SY]

He is already convinced that no difference is audible, would you really argue that his body language can´t be a confounder?

Wouldn´t that contradict nearly everything you´ve said about nonverbal communication before?

No, not at all. As long as the experimenter doesn't know the answer, there can be no Clever Hans effect. Again, you have to postulate some mysterious mind-control causing the participant to suddenly not be able to hear differences when blinded- remember, at the start of the tests with the experimenter present, Tiefenbrun and Lavigne and Zipser agreed that the differences were clearly audible. Apparently, the mind-control was switched off then...

[Jakob2]

Originally Posted by SY

And if memory serves, I've already critiqued the Sturm paper. It was tough reading- my technical German is weak- but if I understand correctly, the technical controls of the experiments were not impressive. Amusingly, this was exactly the sort of test that the faith-based criticize: group sessions, unfamiliar system, short passages, no listener control of presentation (a strong point of ABX).

AFAIR you´ve only stated that is wasn´t significant, without any analysis or argument presented.

Regarding technical controls you have to be more specific, because he uses several and even a negative control.

It may be amusing, but in difference to ABX he uses a paired preference test which is much more familiar to the normal user.
It´s intentionally a group test and so it is different compared to a test of a single listener, and he did, according to his description, some preliminary tests and refined the test procedure based on these preliminary tests.

But it is as usual, if a test result is positive one has to look if it was the EUT that made that result or if other reasons were possible.

In contrary, if a test result is negative, one has to look if some confounders prevent the participants from detecting a difference.

Sturm did use a negative control, that was nice to see.

Wishes

[john curl]

It is difficult, to 'impress' SY. Me, I'm easily impressed, I just listen for myself. Sure, a few errors squeak by, one in a while, BUT the overall result is PROGRESS! ;-)

[Jakob2]

Originally Posted by SY

No, not at all. As long as the experimenter doesn't know the answer, there can be no Clever Hans effect. Again, you have to postulate some mysterious mind-control causing the participant to suddenly not be able to hear differences when blinded- remember, at the start of the tests with the experimenter present, Tiefenbrun and Lavigne and Zipser agreed that the differences were clearly audible. Apparently, the mind-control was switched off then...

Is that so? If the experimentator thinks that no difference is audible he can´t express that nonverbal?
It wasn´t reported but there is nowhere stated that Tiefenbrunn for example was familiar with ABX tests and of course no positive control was used.

BTW, remember that Lipshitz himself confirmed by his description that no one was reaching the sufficient sensitivity level.

Wishes

[SY]

Sturm did use a negative control, that was nice to see.

That was the strong point. Unfortunately, there seemed to be very little information on the controls implemented to insure absence of non-auditory cuing. If I'm mistaken about this because of my mediocre German, please point me to the proper sections.

[SY]

Is that so? If the experimentator thinks that no difference is audible he can´t express that nonverbally?

He can even express it verbally. But as long as the subject still claims to hear the differences with the experimenter present (but with non-auditory cues), you have to postulate not just mind-control, but mind-control that can be switched on and off!

[boconnor]

Originally Posted by Jakob2

@ terry j,…I know that blind testing of loudspeakers is difficult, but in fact, if you were really convinced that you were not able to deal with the various bias mechanisms (at least to a certain degree) than consequently you would have to quit any sighted listening evaluation.

That’s a bit unfair for two reasons.

First, as terry j pointed out, dbt for loudspeakers is simply not feasible in a home environment, even if you wanted to do it. You would have to make sure each loudspeaker was placed in its respective “sweet spot” each time you listened. And what if one speaker system was passive and the other active? Imagine the complications in amps and switching. That’s one reason why loudspeaker manufacturers like Harmon Karman built special facilities to mechanically move different speakers during listening tests. In contrast, dbt of cables can be done straightforwardly at home with a person’s own equipment.

Second, there are clear, measurable differences between loudspeakers. And loudspeakers with different designs clearly sound different from each other, in particular in the way they interact with the listening space. It would not be very useful to do a dbt between say a Quad electrostatic and a horn loaded infinite baffle design in order to say “Can we distinguish A from B?”. Of course we can, more or less instantly.

The best use of dbt is, for instance, when engineering theory says a priori the differences should be subtle or non-existent. Then you need dbt to remove the non-audio biases that can come into play so that we can determine if there really is a difference between A and B. That is rarely the case with loudspeakers but is quite germane for cables.

[john curl]

It doesn't work for cables, either.

[Jakob2]

Originally Posted by boconnor

That’s a bit unfair for two reasons.

First, as terry j pointed out, dbt for loudspeakers is simply not feasible in a home environment, even if you wanted to do it.
<snip>

Feasible or not, doing a sighted evalution does only make sense if you think you are able to deal with your bias (at least to a certain degree);
Otherwise if you can´t control your bias, a sighted comparison is pointless as you´ll mostly hear what you want to hear.

It is not about fairness or not, more to think about the various arguments; if it´s black or white, that any sighted comparison is pointless (Toole/Olive have shown that rating _and_ ranking changed sighted compared to blind), if there is some grey you must admit that one might be able to learn to deal with his bias (something that Toole/Olive haven´t investigated as far as i know)

You would have to make sure each loudspeaker was placed in its respective “sweet spot” each time you listened. And what if one speaker system was passive and the other active? Imagine the complications in amps and switching. That’s one reason why loudspeaker manufacturers like Harmon Karman built special facilities to mechanically move different speakers during listening tests. In contrast, dbt of cables can be done straightforwardly at home with a person’s own equipment.

In short, hey it is pointless, but i do it anyway as it is difficult to do it otherwise.

Really not wanting to be offensive, but the argument was not "difficulty of test" but "uncontrollable bias" .

Second, there are clear, measurable differences between loudspeakers. And loudspeakers with different designs clearly sound different from each other, in particular in the way they interact with the listening space. It would not be very useful to do a dbt between say a Quad electrostatic and a horn loaded infinite baffle design in order to say “Can we distinguish A from B?”. Of course we can, more or less instantly.

You really should try a blind test on that one; you might be in for a shock (assumed that you were not used to blind testing)

The best use of dbt is, for instance, when engineering theory says a priori the differences should be subtle or non-existent. Then you need dbt to remove the non-audio biases that can come into play so that we can determine if there really is a difference between A and B. That is rarely the case with loudspeakers but is quite germane for cables.

I´m sorry, but that is not true; dbt is part of scientific methodology and is used if you were searching for reliable results according to scientific standards; everything else is just experimentator bias.

And please remember, a dbt itself doesn´t remove _the_ non-audio bias, it just removes a bias that is introduced due to peeking.
All other bias mechanism have to be addressed too.
Look at wikipedia for example for an impressive list of possible bias mechanism and try to figure out which of these are affected just by "blinding" .

@ terry j,

sorry i forgot to address one point. You did ask, why the sighted listener did not need any sensitivity control or something else.

My answer would be simply, that if someone stated that he hears something, it should be clear that you can´t trust in him; it is just a subjective statement and if you don´t know him and further if you haven´t shared some listening experience you can´t know whether his statement has any meaning or if it is just true.

A DBT usually claims a lot more (at least it is notoriously used so some time later) and therefore - according to scientific standards- it has to be shown that it is an objective, reliable and valid test.

Wishes