I'm pretty sure that the posts to which he was responding were about that (including mine), but I was fascinated by his use of a false dichotomy which distinguished perception from results in tests of perception. If results in blind auditory perception tests are to be distinguished from auditory perception, then, once again, what are we talking about?
The usual tests have to rely on the answers of the participants. In this regard any test result depends on the subjective internal jugdement process.
Oohahsi et al. (experiments for the so-called Hypersonic Brain Effect) tried to get additional more (hopefully) objective informations by usage of EEG and PET/Scans, the more recent approach is to use fMRI.
Having that said, let´s remember that we only count hits (or misses) to the trial question but don´t know if that really reflects what is/was percepted, because we can´t look into the heads/brains of our listeners.
We know that testing the same sensory difference in different test protocols using the same participants leads to different results, so obviously the results are not just reflecting the "raw" auditory perception.
To keep it short, the assumption that "blind" test schemes somehow ensure that the results only depend on the auditory perception is most likely wrong.
In an another thread we already discussed the plethora of possible bias effect still at work in "blind" tests.
I hope you realize by now that it isn´t a "false dichotomy".....
Talking about the difficulties to do tests that are _really_ objective, valid and reliable, furthermore about which way to do better experiments, about what is considered to be sufficiental evidence and obviously about a lot more that would be better done in other threads, but that´s (forum)life.
Could it be that you´ve forgot for a moment about the last 40 years of research on HTRFs and binaural recording?
So, what we listen to is already a mechanically processed version of the orginal sound field and it is further processed by our auditory "apparatus".
Our brain is used to work with this individually preprocessed material and that´s the reason why binaural recordings done with individualized dummy heads/torsi are so stunningly "realistic" .
In the usual recording setup none of that is used (beside maybe the spacing) and it was imo one of the most surprising detections that similar perception is possible although the reproduced sound field (two channel stereophonic setup) was vastly different from the original sound field.
It works because our brain constantly tries to construct a perception from the cues it gets from the reproduction that is compatible to our experiences.
I gave an example of a real world ABX test on the other thread to show what's involved in such testing & bear in mind the participant was an audio engineer. He found it extremely difficult but well worth reading his experience which I'll reproduce just his summary post f his experience here:
Again, remember this guy is an sound engineer so works with & relies on his hearing every day - I presume he has more listening experience than the average audiophile.
He also achieved some positive ABX tests for differentiating jitter levels & had to train for a totally different audible cue
Thanks for that. I agree, listening for precision and extent (clarity) of 3D position/depth information is the first thing to listen for. Another is noting the groove/vibe/flow of the reproduced audio and how it affects oneself...self involvement/enjoyment/relaxedness etc. This is probably fairly difficult in a group situation.
Dan.
@scottjoplin,
As the more modern definition of the term "Psychoacoustics" coined by Blauert in the 1960s is to examine if a sound event leads to an auditory impression (in German it´s Hörereignis, so maybe audiotory event is a better translation?) and if different sound events lead to different audiotory impressions. That covers all from JNDs to DLs on every possible sound event.
So psychoacoustics is relevant especially as we are talking about the instruments to find out about listening impressions.
Traditionally psychoacoustic was more interested in the basic mechanisms and did not use multidimensional stimuli (aka music) to examine the hearing sense; today that has changed.
That´s another long ongoing discussion, as there is no universal agreement if the "accurate reproduction" should be related to the original event or to an arbitrarily choosen (a bit exaggerated of course) intermediate stadium (i.e. the recorded content that nobody really knows)
<snip> Anything else that is going on is in the listener's head, and I'm not interested in what goes on in people's heads, usually because they don't know, so how can I?
As the more modern definition of the term "Psychoacoustics" coined by Blauert in the 1960s is to examine if a sound event leads to an auditory impression (in German it´s Hörereignis, so maybe audiotory event is a better translation?) and if different sound events lead to different audiotory impressions. That covers all from JNDs to DLs on every possible sound event.
So psychoacoustics is relevant especially as we are talking about the instruments to find out about listening impressions.
Traditionally psychoacoustic was more interested in the basic mechanisms and did not use multidimensional stimuli (aka music) to examine the hearing sense; today that has changed.
That´s another long ongoing discussion, as there is no universal agreement if the "accurate reproduction" should be related to the original event or to an arbitrarily choosen (a bit exaggerated of course) intermediate stadium (i.e. the recorded content that nobody really knows)
Last edited:
As the more modern definition of the term "Psychoacoustics" coined by Blauert in the 1960s is to examine if a sound event leads to an auditory impression (in German it´s Hörereignis, so maybe audiotory event is a better translation?) and if different sound events lead to different audiotory impressions. That covers all from JNDs to DLs on every possible sound event.
So psychoacoustics is relevant especially as we are talking about the instruments to find out about listening impressions.
It always strikes me that what people mean who think psychoacoustics is irrelevant to this audio hobby, is that psychoacoustics is what people 'imagine' they hear
Thank youAs the more modern definition of the term "Psychoacoustics" coined by Blauert in the 1960s is to examine if a sound event leads to an auditory impression (in German it´s Hörereignis, so maybe audiotory event is a better translation?) and if different sound events lead to different audiotory impressions. That covers all from JNDs to DLs on every possible sound event.
So psychoacoustics is relevant especially as we are talking about the instruments to find out about listening impressions.
Traditionally psychoacoustic was more interested in the basic mechanisms and did not use multidimensional stimuli (aka music) to examine the hearing sense; today that has changed.
It is not possible to argue and debate the issues in this thread to an agreeable conclusion. Everybody knows that, right? And everybody should know that people on both sides who have already firmly made up their minds are not going to be persuaded to change here and now, perhaps not ever, perhaps no matter what experimental evidence were to be produced in the future. So, what is the point of arguing?
Nice picture of the horse not pulling but trying to shove.....
I think at least the poster who demand that "golden ears" should provide evidence by doing "blind" tests must know what they would accept as evidence. I have assumed to see the "extraordinary claims need extraordinary proof/evidence" argument for example.
In a broader sense it is imo worth to think about what could constitute sufficient evidence.
As you´ve mentioned Kahneman & Tversky, it´s a good example how important it is to remember that we are always more willing to accept confirming information than to accept contradictionary. Kahneman himself pointed out that he should have known better but was nevertheless trapped in the priming case.
That´s an interesting question as it points directly to bias effects. As usual there are already some experimental results (although again mainly from food sensory tests) where the participants where asked for their preferences when in reality testing the exact same stimulus twice or even more times.
If a "no preference/difference" answer option is included, the proportion of correct answers (means number of correct no preference/difference answers) is always the lowest when probing the exact same stimulus. Stable phenomenon over different product categories (and well known in audio tests) and even a crossnational effect.
But it is mainly noticed in multidimensional evaluations not so in one-dimensional experiments.
Yes, if a test setup is testing the difference between A & B, the participants naturally assume that there is a difference & they are trying to find it - so even if A is really the same as B, they will be psychologically primed to hear differences (psychologically, we want to be right) - when the option of no difference is included as a possible answer, it changes this stress/pressure
You seem to lack the overall system view to discuss the real issues and constantly end up in irrelevant details e.g. the one you refer to below.
Please educate me - what are the "real issues"?
I was imply pointing out Greisinger's research into using dummy heads for binaural recordings as being less than optimal for a realistic eperience
You quoted planet10's post in your post & seemed to be replying to this but maybe you weren't?
Regarding the kind of evidence it takes to change firmly held opinions in scientific fields, there is already much written about the subject. For example: Scientific revolutions or paradigm shifts | Deskarati
Thomas Kuhn: the man who changed the way the world looked at science | Science | The Guardian
Thomas Kuhn: the man who changed the way the world looked at science | Science | The Guardian
Last edited:
As usual it depends, as you don´t need to include a tie (i.e. the no difference/no preference option) option in an A/B test - but if you do you get another/an additional analysis choice.
Basically the ABX test is a pure test for difference so you can´t use it for preference tests.
I have to check if there were comparisons between ABX and A/B "same/different" test protocols already done.
- Status
- Not open for further replies.