[Albedo]

Yes, it's all about the obsession over something minute that's different from the linear thinking as I've somewhat tried to explain in post #1006. To further elaborate.. it's about what we all are trying to latch on to and to some degree fail at.
 
 
 
Quote: http://asadl.org/jasa/resource/1/jasman/v88/iS1/pS49_s2?bypassSSO=1

The role of attention in detection under conditions of frequency uncertainty has been extensively studied in a masking paradigm using auditory cues to alleviate the uncertainty. Studies of cues requiring higher‐order processing, for example, cues related to the signal in a fixed musical relation, show that such cues can be as effective as iconic (same frequency) cues. Subjects with absolute pitch detected pure tones in noise in conditions with (1) no cues, (2) iconic cues, or (3) visual cues intended to operate through auditory memory. Visual cues consisted of various symbols corresponding to the frequency of the signal presented in musical notation on a display terminal. In all cases, these were better than no cues and in the most effective cases, the visual cues were as effective as the iconic (auditory) cues.

 
In the everyday life music sometimes sounds dreadful, sometimes not... timbreeeeeeeeee!

 

[drez]

Quote:

 
Blind ABX testing can work well for verifying the existence of an alleged concrete difference, which can be previously discovered in any way, sighted or not. There is usually no shortage of those, I see many claims of gear having "better sound stage", "more impactful bass", "improved details", "less sibilance", and more being made with great confidence. Statements like these would imply already knowing exactly what to listen for ?
 
What reliability advantage does sighted listening have over an otherwise identical, but blind test (assuming that the "not knowing what to listen for" issue is already dealt with), other than being more likely to produce the - more desirable to audiophiles - positive result due to its own flaws ? Two wrongs do not make a right.

 
But to me blind testing, within the domain of discussing and selecting audio gear, is for mo mostly superfluous in that it mostly focuses on the production of evidence which can be used by others, so that they are excused from gathering their own evidence and experience.  It is a fairly blunt and mute instrument, and more often than not it is applied with the intention of mythbusting and is held to be more conclusive than it deserves to be.  It is like taking an eye test while squinting one's eyes.  It is both more controlled, and yet it yields a lower resolution of information.
 
Objective testing with equipment, although having problems of it's own, is much more reliable in my opinion.  It can reliably show many differences that blind testing cannot, and this information can be used to challenge or corroborate sighted testing.  

 

[Tilpo]

But to me blind testing, within the domain of discussing and selecting audio gear, is for mo mostly superfluous in that it mostly focuses on the production of evidence which can be used by others, so that they are excused from gathering their own evidence and experience.  It is a fairly blunt and mute instrument, and more often than not it is applied with the intention of mythbusting and is held to be more conclusive than it deserves to be.  It is like taking an eye test while squinting one's eyes.  It is both more controlled, and yet it yields a lower resolution of information.

Objective testing with equipment, although having problems of it's own, is much more reliable in my opinion.  It can reliably show many differences that blind testing cannot, and this information can be used to challenge or corroborate sighted testing.  

Isn't that the very definition of empirical evidence, which forms the foundation for the natural sciences?

 

[drez]

Quote:

Isn't that the very definition of empirical evidence, which forms the foundation for the natural sciences?

 
For the first part yes, for the second part hopefully not.  I should hope that the results and data should be continually challenged both from within the discipline and across disciplines rather than excusing the collection of further data or development of different methods of research.

 

[liamstrain]

Quote:

 
Using schizophrenics in examples over and over isn't very useful.  
 

 
The point is that these are not schizophrenics. This is a fundamental issue with being human, as "playing backwards/lyrics" the example you yourself posted, demonstrated. 

 

[Tilpo]

For the first part yes, for the second part hopefully not.  I should hope that the results and data should be continually challenged both from within the discipline and across disciplines rather than excusing the collection of further data or development of different methods of research.

But it's not like there has been only one double blind test. There have been countless as far as I know. Not for every single component, but I don't find it such a strange thing to extrapolate data from comparisons between other components.

 

[drez]

Quote:

But it's not like there has been only one double blind test. There have been countless as far as I know. Not for every single component, but I don't find it such a strange thing to extrapolate data from comparisons between other components.

 
My point is that blind tests are not the right kind of testing methodology to discern very small changes between audio samples - if you measure the differences and they are below -90dB in level then there is not much point doing another blind test in the same methodology - it is just beating a dead horse, even though thoroughness can be a good thing.
 
If you measure the difference parameter, and prior blind testing has shown that this magnitude cannot be detected by blind testing, then there is small chance that further blind tests will prove successful.  The testing methodology, in said hypothetical configuration, is more than likely not capable of producing a meaningful result and probably need to be reviewed.
 
Don't get me wrong I think blind testing has it uses just it also has very evident limitations in what it can demonstrate also.  For example if DBT test between two components shows no difference - this does not mean that they are the same - it just means that in the test participants could not tell them apart.  There could be a lot of important/useful performance differences that are not demonstrated by the blind test.  Of course one can argue that differences that cannot be shown in blind tests are not tangible and therefore not important, but I do not think this philosophy/attitude is universally/exclusively valid nor self-evident.

 

[Tilpo]

My point is that blind tests are not the right kind of testing methodology to discern very small changes between audio samples - if you measure the differences and they are below -90dB in level then there is not much point doing another blind test in the same methodology - it is just beating a dead horse, even though thoroughness can be a good thing.

If you measure the difference parameter, and prior blind testing has shown that this magnitude cannot be detected by blind testing, then there is small chance that further blind tests will prove successful.  The testing methodology, in said hypothetical configuration, is more than likely not capable of producing a meaningful result and probably need to be reviewed.

Don't get me wrong I think blind testing has it uses just it also has very evident limitations in what it can demonstrate also.  For example if DBT test between two components shows no difference - this does not mean that they are the same - it just means that in the test participants could not tell them apart.  There could be a lot of important/useful performance differences that are not demonstrated by the blind test.  Of course one can argue that differences that cannot be shown in blind tests are not tangible and therefore not important, but I do not think this philosophy/attitude is universally/exclusively valid nor self-evident.

In most cases if differences are not discerned in a proper blind test (decent sample size, good equipment, double blind), then it does have significance: it means that the difference, if it exist, is either below audible threshold or on the border of it.
If it does not show up on a blind test, the difference simply cannot be huge, or else the test wasn't done properly.

Over the past blind tests have as far as I know more or less proven that most CD players, amps and DAC's have very similar performance, meaning boutique components in these three categories are probably not worth it over decent mid-fi. It does not necessarily mean they are the same, no. But in my opinion it's more than enough reason to stick with mid-fi components and not feel the need to upgrade them.
Tests have also shown that cable myths are false. But to be honest, I think that was done by electrical measurements and comparing them to established psychoacoustic thresholds as well, it's just that people didn't believe such tests.

 

[drez]

Quote:

In most cases if differences are not discerned in a proper blind test (decent sample size, good equipment, double blind), then it does have significance: it means that the difference, if it exist, is either below audible threshold or on the border of it.
If it does not show up on a blind test, the difference simply cannot be huge, or else the test wasn't done properly.
Over the past blind tests have as far as I know more or less proven that most CD players, amps and DAC's have very similar performance, meaning boutique components in these three categories are probably not worth it over decent mid-fi. It does not necessarily mean they are the same, no. But in my opinion it's more than enough reason to stick with mid-fi components and not feel the need to upgrade them.
Tests have also shown that cable myths are false. But to be honest, I think that was done by electrical measurements and comparing them to established psychoacoustic thresholds as well, it's just that people didn't believe such tests.

 
To me the fact that many boutique CD players are not worth the money does not require blind testing - it is pretty obvious when some of them are measured with test equipment!  But that is probably missing your main point.
 
To me I interpret your main point to be that double blind testing shows the limits of what is audible - is this really the case?  In other words when the individial parameters are isolated in their own blind test (like on this website), can people detect the same differences in music playback blind tests?
 
I think it is a fair conclusion that if differences don't show up in blind tests that they are not huge or even particularly substantial for most people, but to me there are definite limits in terms of what can be inferred from the current body of tests.
 
On a sidenote I have an idea for a DBT I wish to perform with WAVE vs FLAC CD rips ( I have two files that to me sound very different) but I need to check the level of the two files first - is there a good (free) program to do this with?

 

[maverickronin]

You should be able to null them with Audio Diffmaker.

 

[Tilpo]

To me the fact that many boutique CD players are not worth the money does not require blind testing - it is pretty obvious when some of them are measured with test equipment!  But that is probably missing your main point.

It gives us an idea of how the measurements correlate to physical world. Additionally, it makes it easier to believe there isn't an audible difference, for those who don't accept the measurements as enough proof.

To me I interpret your main point to be that double blind testing shows the limits of what is audible - is this really the case?  In other words when the individial parameters are isolated in their own blind test (like on this website), can people detect the same differences in music playback blind tests?

That's not my main point.
Audible thresholds are determined in psycho acoustical research.

My point is that if it doesn't show up in DBT, then it is going to be a very subtle difference at best. Not enough for me to care.

I think it is a fair conclusion that if differences don't show up in blind tests that they are not huge or even particularly substantial for most people, but to me there are definite limits in terms of what can be inferred from the current body of tests.

Of course, but in general it does provide evidence for the view that most modern electronics perform at or really close to transparency.

On a sidenote I have an idea for a DBT I wish to perform with WAVE vs FLAC CD rips ( I have two files that to me sound very different) but I need to check the level of the two files first - is there a good (free) program to do this with?

Foobar2000 has a very good component for doing DBT's with different files.

 

[drez]

Complains that I have no recording device and shuts down

 
I will do some google research.

 

[maverickronin]

I thought it could do it with files you already had. Sorry about that.

 

[kiteki]

Audio diffmaker, while I haven't researched it yet, my immediate thoughts are
 
- The A/DC / recording process is vital
 
- when you compile, subtract and listen to the residual, it could lead to some very flawed conclusions
 
 
For example, some guy was telling me recently information above 11kHz is unnecessary, I told him I can very easily ABX a 16kHz+ cut-off and he thinks that's chasing fairy tales or hallucinations.
 
He came to that conclusion by using an equalizer to cut off everything under 15kHz, (in which case you can hardly hear anything at all), so it would seem like when you're playing the 15kHz and lower information it would very easily mask the 16kHz+ information, so the conclusion is  that extremely faint information is unnecessary.
 
That just isn't how it works in real playback, and now he's calling the redbook standard too high.
 
The diffmaker would easily lead to similar conclusions if you don't take the ADC or real playback into account, which it clearly doesn't.
 
The column on the right hand side with notes like "Paints and lacquers used on cables" or "EMI control devices" seem to indicate it's cynical research anyway, seeking not looking.
 
Just a random example, how am I supposed to record, compile, subtract and listen to a ferrite clamp on a USB cable?  I don't think that's possible with that program, so does that mean ferrite clamps are useless?  The exact same thing would happen with every single capacitor out there, if we all followed cynical research like that audio systems would all be built at radioshack.

 

[kiteki]

In most cases if differences are not discerned in a proper blind test (decent sample size, good equipment, double blind), then it does have significance: it means that the difference, if it exist, is either below audible threshold or on the border of it.
If it does not show up on a blind test, the difference simply cannot be huge, or else the test wasn't done properly.

 
Usually the last part.  I'm sure a standard RealTek soundcard versus a high-end DAC is borderline in a lot of cases.  That doesn't mean we should all use RealTek, even if it sounds "okay". =) 
 
Can you tell the difference between $2 and $50 shampoo in a DBT?
 
Let's all use...