[PhilS]

Quote:

Originally Posted by bribassguy /img/forum/go_quote.gif You brain lies to you...

Always?

 

[mike1127]

Quote:

Originally Posted by bribassguy /img/forum/go_quote.gif You brain lies to you... it works with all the senses...

Yes, we are aware of that. What conclusion do you want us to draw from it?

When someone presents this general fact as though it somehow implicates all sighted tests (but for some reason not blind tests), well, that's silly.

I think it's far more useful to investigate under what conditions we are subject to illusions, and under what conditions we aren't. And to investigate what illusions are specific to comparing two audio devices. Most of the example optical (and some audio) illusions that get posted to this forum have nothing to do with comparing devices (or comparing anything).

 

[ClieOS]

I don't like the word 'lies', as it implies the brain is actively trying to deviate the person from seeing what he see, hear, or sense.

What I believe that is closer to the truth is, our brain is merely trying to 'fill in the blank' between what we see, hear, or sense. Rationalization is one of human's greatest gift and curse.

 

[mike1127]

Quote:

Originally Posted by Bullseye /img/forum/go_quote.gif What is the bad made when someone says this cable that costs $xxx and sounds excellent? That people who don't understand or know might think that that is actually the case.

Maybe it actually is the case.

Quote:

That is the reason why I bother arguing with other people that keep wanting to believe. So that other people don't waste their money on something that it is just good looking, whilst fooling themselves in order to justify it.

I get into these arguments too, but what I'm using arguing against is the idea that we have incontrovertible knowledge one way or the other. Some people consider a conventional DBT (let's say quick-switched, ABX box) to be solid knowledge. I don't. I think it's largely an invalid test protocol. Some people consider they can make a definitive conclusion about a cable by taking one listen. I don't think it's that easy or quick.

So I'm saying something similar to PhilS. My message is: please don't get distracted by people who are certain one way or the other, but rather make up your own mind by whatever method seems right for you.

 

[mike1127]

Quote:

Originally Posted by ClieOS /img/forum/go_quote.gif I don't like the word 'lies', as it implies the brain is actively trying to deviate the person from seeing what he see, hear, or sense.

In some optical illusions people see things that aren't there, or get events out of time order. I agree with you that your brain isn't "lying" to you, but brains definitely get some things wrong. Of course the question is how this relates to audio testing. It's not right to simply toss out this vague notion---"Your brain lies"---and think that means that from now on all amplifiers and cables should be evaluated by measurements and not listening, or something equally absurd.

 

[bribassguy]

Okay, so maybe it is not trying to be deceitful but it will try to fill in the gaps. It's a pretty simple trick really.

You hear two notes an octave away. You area told or shown something in this case the pitch will con't to drop as you reply it and then that's what you hear. Unless you critically listen or have perfect pitch you believe what you are told.

The more applicable question would be if the same holds true for or say sales men: These fancy looking high dollar cans sounds much better than these other one.

Would your brain fill in the gap there? It is fact people prefer food in fancy boxes, so do people prefer sound in a fancy package?

 

[TheAttorney]

It took some time, and endless trawling through scientific papers and this forum, but I can now declare that there are officially 1,478 different types of sound illusions that can fool our brains. And I’ve counted every one of them.

As already noted, many are positive, i.e. they help us turn a recorded signal into a brilliant, emotional music event. Amazing really. But the same illusions can be negative when it comes to comparing components. Many are subconscious, automatic, involuntary, not easy to control, and not easy for us to realise they are having an impact.

Recently, Mike1127 started a number of threads exploring various ideas about sound perception and, amongst other things, the best way of listening to overcome illusions. I followed these with interest because I would love to have a practical, foolproof way of choosing new components. Some of the responses to Mike’s threads were thoughtful, but too many followed a repetitive pattern which, to me, shows a serious blind spot exhibited by a certain kind of scientist. And some of this thread shows the same pattern. It stunts the exploration of alternative thoughts. Let me illustrate with the following discussion between an Enquiring Mind (EM) and a Scientist (SC):

EM: I think you can hear differences more reliably when listening to music as music instead of music as sound.

SC: The only thing that is truly reliable is a DBT. Look at this (spiral blue/green) illusion to show how easily our minds can be fooled.

EM: My original point was that that DBT or sighted doesn’t really come into it. Here’s one explanation as to why the spiral blue/green illusion could affect DBT as easily as a sighted test.

SC: The only thing that is truly reliable is a DBT. Oh and all cable companies are crooks.

EM: Hmm, my points have nothing to do with cables. I’m talking about sound perception. What about the monkey suit analogy? This shows that concentrating in a certain way can impact our ability to spot differences, even in controlled tests.

SC: The only thing that is truly reliable is a DBT. Look at this analogy (black/white spots) to show how easily our minds can be fooled.

EM: I’ve just given an explanation why the spiral blue/green analogy may apply to DBT as easily as any sighted method. You haven’t given a good counter argument to that, but now produce a new analogy. Here is one reason why the black/white spots analogy could affect DBT as easily as a sighted test.

SC: The only thing that is truly reliable is a DBT. And here is a new illusion to show how easily our minds can be fooled.

Enough. My point is that not once in any of these threads did I spot a counter argument to an EM idea that made me stop and think “Ah, the Scientist has good a point on this one, maybe I need to consider DBT (or other controlled method) more carefully” I truly do not understand how some scientists can be provide good critical analysis on many aspects of a debate and yet have such a blind spot that DBT is infallible. It seems so unscientific to me.

I’ll say this just one more time: DBT successfully address just 1 illusion, placebo, but what about the other 1,477 illusions? For the remainder, DBT (and other controlled tests) may be better than, equal to, or worse than sighted/uncontrolled tests, but until we have a more open minded debate we’re unlikely to ever find out.

 

[PhilS]

^ Hilarious.

 

[Dane]

This thread is pretty useless when it comes to exploring audio delusions.

First, the OP is an illusion, not a delusion.

Second, this visual illusion requires only the sense in question (sight) on order to work. If e.g. cables were a similar illusion (and not a delusion) that too would require only the sense in question (hearing), but as we all know audio delusions require sight as well in order to do their magic.

Third, this visual illusion consistently produces the same (or similar) impressions across multiple individuals.

So, what is an audio illusion? Stereo music is a pretty good example. It works consistently across individuals and only requires the sense in question in order to work.

You simply cannot compare illusions with delusions. Completely different things.

BTW, I'm all for delusions, whatever it takes to enjoy the music the most

 

[71877]

Quote:

Originally Posted by TheAttorney /img/forum/go_quote.gif I’ll say this just one more time: DBT successfully address just 1 illusion, placebo, but what about the other 1,477 illusions? For the remainder, DBT (and other controlled tests) may be better than, equal to, or worse than sighted/uncontrolled tests, but until we have a more open minded debate we’re unlikely to ever find out.

Do these 1,477 illusions occur in both DBT tests and uncontrolled/sighted tests (and not in a testing situation at all) (or: are they all relevant when comparing components)?

Yes, a DBT checks for a placebo effect (the bias about certain components), which is the whole point imo of a DBT: to see if the subject still can hear a difference when it doesnt know which component is which or which components are actually used. If there is a difference in results then (from the sighted test, so you might even say that a DBT on itself is rather meaningless without comparing it to a similair sighted test), it may be possible that the subject only thought there was a difference because it knew to which component it listened, which then created a bias. The DBT 'isolates' this problem of the bias in the test from other factors. Different controlled tests (or the same test with different kinds of components) need to be done for the other 'illusions'. Uncontrolled tests cant be used, since these in tests we are (more) unsure what is controlled, if anything is controlled at all. So yes, controlled is better if you can isolate the factor you want to research/test.

A problem however can be if the DBT introduced some new 'effects' to the subject over the sighted test, which significantly influenced the results of the DBT, so you can't be sure anymore if the difference in results is because of the placebo effect or the actual DBT. If an illusion occurs in both the DBT and sighted test (or any other test method for that matter) in the same way (so all tests are influenced the same way), then I dont really see the problem, as long it isolates the bias effect.

Another problem is that the new effects occur simply because of the testing, so that we cant be sure of any result, since the results can also be an effect of the actual test.

These problems can be tested on its own to determine if they significantly influence the test.

The 'discussion' was more like between a scientist (the EM) and a very silly stubborn person (which a scientist shouldnt be) btw ^^

 

[mike1127]

Quote:

Originally Posted by Gundogan /img/forum/go_quote.gif Do these 1,477 illusions occur in both DBT tests and uncontrolled/sighted tests (and not in a testing situation at all) (or: are they all relevant when comparing components)? Yes, a DBT checks for a placebo effect (the bias about certain components), which is the whole point imo of a DBT: to see if the subject still can hear a difference when it doesnt know which component is which or which components are actually used.

Common sense and introspecting on my listening process tells me that we have the following illusions in comparing components, which are present in both blind and sighted conditions.

BTW, you said that if an illusion is present in both blind and sighted conditions, it shouldn't matter because the variable doesn't need to be controlled. That's wrong. The point is that most current attempts to "control" a test assume that a human brain is a measuring instrument---that is, if an audible difference in sound is present, then the human brain will respond to that consistently and repeatedly. But we have these problems:

1. If your test track is just a short snippet and/or is looped and repeated many times, you are no longer testing for differences that are perceived in the music. (Because you can't perceive all the music under these conditions.) This is an illusion in which the music disappears.
2. If you use entire songs or some kind of long signal, the test music is very rich and it's normal to notice different things on each repetition, giving rise to the illusion that the signal has changed, when in reality you are just noticing something different.

In using a long signal, there is also the practical problem that test trials take a long time and it is hard to reach a statistically significant number of trials.

Quote:

The 'discussion' was more like between a scientist (the EM) and a very silly stubborn person (which a scientist shouldnt be) btw ^^

Believe me, every audio scientist I've met on the internet is not at all like EM and in fact dismisses or ridicules the statements by the EM, thereby greatly resembling the "silly stubborn" person.

 

[71877]

Quote:

Originally Posted by mike1127 /img/forum/go_quote.gif BTW, you said that if an illusion is present in both blind and sighted conditions, it shouldn't matter because the variable doesn't need to be controlled. That's wrong. The point is that most current attempts to "control" a test assume that a human brain is a measuring instrument---that is, if an audible difference in sound is present, then the human brain will respond to that consistently and repeatedly.

Well, I consider the human not only the measurement tool, but also the subject of testing, along with the tested component. Because imo, that is what you test: can the human testsubject hear the difference between 2 components, not if there's a difference between 2 components period. You do not only test the component, but the human too and yes, the human is flawed, not to mention that humans are far from 'the same'.
If there are factors that appear in all situations and do not directly influence the testing method (for example, the subject is ill, which affects his or her hearing), then yes, it shouldnt matter since you also test the human subject. You can seperatly test why the subject can or cannot hear a difference (which can be tested with a DBT ABX test for the placebo/bias effect, and you can argue that it can also be used for difference testing as a whole).
The great challenge is to find a testing method where there're no or few factors influencing the actual testing method (mainly keeping the circumstances the same when listening to component A or B) when testing both the human and components. Altho, a simple sighted test might even do the trick.....

Quote:

Believe me, every audio scientist I've met on the internet is not at all like EM and in fact dismisses or ridicules the statements by the EM, thereby greatly resembling the "silly stubborn" person.

I know what you mean. A scientist should be open minded in a way, although he or she can still dismiss things ofcourse with good argumentation.

 

[PhilS]

Quote:

Originally Posted by mike1127 /img/forum/go_quote.gif Believe me, every audio scientist I've met on the internet is not at all like EM and in fact dismisses or ridicules the statements by the EM, thereby greatly resembling the "silly stubborn" person.

That's because most of them are not really trained scientists, but are people who just read on the internet some reports of studies or, even worse, read what others have posted about the studies. Based on this, they adopt a broad dogmatic position about the issue of audible differences between certain audio components without fulling exploring the issue or considering alternative points of view.

If you read some of the threads in this forum, it's easy to pick out the people who really have knowledge in the relevant sciences -- and there are several on both sides of this issue -- and those who, while not being trained in the sciences, are at least able to have an intelligent, courteous, and open-minded dialogue on these issues. The rest are just trolls.

 

[Ed Seedhouse]

Quote:

Originally Posted by linuxworks /img/forum/go_quote.gif if your eyes can be fooled, why not your ears?

No need to argue by analogy in this case. Many types of aural illusions have been unequivocably demonstrated.

Our ears are often deceived. So are our minds, as can also easily be proven.

That does not mean, however, that either our eyes, our ears, or our brains are always deceived. In fact if we were always deceived we would have no way of knowing that we were deceived.

 

[royalcrown]

I'll try to answer the questions you've posed. My response is in bold.

Quote:

Originally Posted by TheAttorney /img/forum/go_quote.gif EM: I think you can hear differences more reliably when listening to music as music instead of music as sound. SC: I have never seen any reliable evidence that listening to "sound as sound" makes hearing differences any less reliable than listening to "sound as music" (or however it goes). In fact, I've seen a good number of blind test comparisons when listening to the entire track diminished one's ability to detect a difference, and in these same comparisons only after listening to sound purely as sound (that is, looking for some artifact in a specific passage) could they detect a difference. I don't have links to them right now, but they were in the mp3 test threads in the sound science subsection. I have also never seen any scientific test of any kind that has shown that listeners are able to distinguish a difference by listening to "sound as music" where previously they could not tell a difference by listening to "sound as sound." Finally, I have seen several blind tests where a listener, thinking that a cable swap has taken place, begins to describe all sorts of differences, only to later find out that the switch never happened. I think this is compelling evidence to believe that some form of expectation bias or placebo effect or whatever is at play rather than some other factor that has not shown up in any rigorous empirical testing. Of course, if evidence did arise showing that some other factor was at play, I'd be much more charitable to the notion. However, as it stands, we have substantial empirical evidence supporting the claim that some bias is at play, and zero empirical evidence that it's due to some other factor - as a scientist I have to default to the side that has the most compelling empirical evidence, even if the just-so story on the other hand sounds more interesting or compelling or persuasive. EM: Hmm, my points have nothing to do with cables. I’m talking about sound perception. What about the monkey suit analogy? This shows that concentrating in a certain way can impact our ability to spot differences, even in controlled tests. SC: In the monkey suit analogy, the subjects were specifically given directions to count how many times the basketball was passed. In the actual experiments (that I have gone over in great detail in many a psychology course), the group that were not given explicit instruction to focus on one aspect of the video over another picked out the monkey suit just fine. Furthermore, in a typical DBT for audio, the subjects are given as much time as they need to find a difference - they can listen to the track, or a specific section of the track as many times as possible, for as long as they need. I would suspect that, on the 3rd or 4th time watching the video, even given specific instructions, the subjects would find the gorilla.

I cut out a bit of the dialogue because they were responses to things I wrote over. If you want to continue this dialogue I'd be more than happy to, I'm just pointing out that it's somewhat unfair to make childish short assertions instead of the argumentation that's actually going on. Part of a fruitful discussion is to always take the most charitable interpretation of an opponent's argument when reconstituting it.

Quote:

Originally Posted by TheAttorney /img/forum/go_quote.gif Enough. My point is that not once in any of these threads did I spot a counter argument to an EM idea that made me stop and think “Ah, the Scientist has good a point on this one, maybe I need to consider DBT (or other controlled method) more carefully” I truly do not understand how some scientists can be provide good critical analysis on many aspects of a debate and yet have such a blind spot that DBT is infallible. It seems so unscientific to me.

I haven't gotten that kind of impression from either side - both sides make good arguments, and while I am persuaded one side over another, that doesn't mean I think DBT is infallible, and I don't think many people here think DBT is infallible regardless of side. Disagreeing with an idea (and indeed, arguing at length against that idea) doesn't necessarily equate to close-mindedness.

Quote:

Originally Posted by TheAttorney /img/forum/go_quote.gif I’ll say this just one more time: DBT successfully address just 1 illusion, placebo, but what about the other 1,477 illusions? For the remainder, DBT (and other controlled tests) may be better than, equal to, or worse than sighted/uncontrolled tests, but until we have a more open minded debate we’re unlikely to ever find out.

I don't think that DBT successfully mitigates all of those illusions. That being said, I also don't think that sighted tests mitigate any more - there's just no inherent advantage that a sighted test brings to the table over a blind test, and AFAIK nobody has argued this. An abundance of failed DBTs don't mean that cables don't make a difference, but that coupled with the lack of any positive DBT whatsoever for cables is far more compelling evidence in my eyes.

Given that people constantly claim to hear differences in sighted comparisons, why do those differences disappear under blind tests? If it is as simple of a matter as listening to "sound as music," how come we don't see anyone passing a DBT while using that tactic? Mike has gone a good way towards this, and I encourage his efforts, so long as he remains rigorous in his testing procedure (which I think he's doing sans some warranted critique outlined in that specific thread). However, I think the real question is: why not blind? Why not cover up that 1 illusion? Would you argue that sighted testing is more accurate than blind testing?