Conclusive "Proof" that higher resolution audio sounds different
- Thread starter amirm
- Start date
That's a very logical statement, Max but is that what's happening? Look at Amir's description of how he felt doing the ABX test. Do you think that any of these factors might influence the test? Do you think these were new factors not there during normal listening?
So did this test involve just removal of one variable & nothing else changed?
How many other blind tests are solely "removal of more variables" & how do you know - do you test for this?
So, yes, a laudable & logical statement but it needs implementation in any test otherwise it remains just an unfounded statement that doesn't apply to any particular test.
Last edited:
John, I asked the question and you replied with five more questions 
I'll ask it again in the hope that you might answer it. We can then analyse what's what.
But clearly the more variables you remove, the more valid the test. Why then would sighted tests be better than any form of blind tests in terms of identifying differences?
I'll ask it again in the hope that you might answer it. We can then analyse what's what.
But clearly the more variables you remove, the more valid the test. Why then would sighted tests be better than any form of blind tests in terms of identifying differences?
I'm sorry, Max, - if you don't understand how discussion works? If you don't understand my points, feel free to ask but simply stating your question again gets us nowhere.John, I asked the question and you replied with five more questions
I'll ask it again in the hope that you might answer it. We can then analyse what's what.
But clearly the more variables you remove, the more valid the test. Why then would sighted tests be better than any form of blind tests in terms of identifying differences?
I'll break it down for you:
Your first statement "But clearly the more variables you remove, the more valid the test" you have not proven operates on any test except probably fully controlled blind tests. Even ABX tests introduced new influencing factors - you may have thought you were "clearly removing variables" but your also introducing new variables. My examples of Amir's test description show this (you can look at others or examine your own experience).
So, to be logical, Max - your second statement "Why then would sighted tests be better than any form of blind tests in terms of identifying differences?" As I said already - half-arsed blind tests & we see ABX tests introduce uncontrolled factors - simply listening as we do naturally doesn't introduce unnatural listening/testing procedures. So, if you wanted me to restate your whole sentence it goes like this "Clearly we need to control all influencing variables so unless we can do this sighted tests are just the same as any other test, probably better because we haven't introduced any new, unknown, unnatural influencing variables"
John, I think it's easy to allow these kind of discussions to get unnecessarily over-complicated and confusing, so I'll try to keep to the point that's been our main focus of debate of late in this thread.
You state that only tests that account for every possible variable imaginable are valid.
You also state that long-term sighted testing can reveal differences that the aforementioned tests cannot, therefore these tests are also valid.
My argument is that your statements contradict each other. Both cannot be correct.
I'm interested in hearing you explain how you think both can be correct.
If you do this, I'm happy to explain why I disagree with both of your statements.
You state that only tests that account for every possible variable imaginable are valid.
You also state that long-term sighted testing can reveal differences that the aforementioned tests cannot, therefore these tests are also valid.
My argument is that your statements contradict each other. Both cannot be correct.
I'm interested in hearing you explain how you think both can be correct.
If you do this, I'm happy to explain why I disagree with both of your statements.
Last edited:
Yes if you are looking to prove something, it is necessary to have valid tests
No, I didn't say "are also valid" - I said have other strengths that blind testing doesn't - quick A/B switching is concerned with spotting differences that rapid switching can reveal. There are other differences that rapid switching will not pick up but these long term listening tests haven't been done properly, AFAIK (Nousaine's article on this had a flaw or two).
How many people do you know have picked their equipment using a valid blind test? More likely that many people use forum-based reports of blind tests, think they "prove" something like all XXXs sound the same & then stupidly buy any XXX.
So, here's my summary in bullet points:
- Valid blind tests are non-existent in audio forums
- ABX testing is much more controlled than the above but still leaves out control over some factors
- rapid A/B testing is good at revealing certain types of differences but not necessarily all
- bearing the above - long term sighted testing has advantages, IMO, - it's our natural way of listening (at home on our system) so we will have a better idea of how we will enjoy/live with the equipment being tested.
I liken this to buying beds or sofas - I would much prefer to have a service that allowed me to take the bed home & sleep on it for a couple of weeks than to try testing it in a shop -I just know that I'm gambling when testing beds by spending 5 (or 30) mins lying on one bed, 5 mins lying on another. Same applies to sofas. So can you tell me that living with a bed or sofa & being able to give it back wouldn't be a better option for deciding which you will purchase?
Absolutely, though our ideas of what's valid differ.
So long-term tests are not valid?
Same system, same music, same sound from speakers - the length of time one may sit in front of them does not affect the output of the speakers, and rapid-switching is only one element of short-term testing - you can also listen to sections of songs, or whole songs, during short-term testing. There is no reason why long-term testing will allow one to hear differences that short-term testing will not. To believe so is faith-based.
I'll pick up on these points later.
But you know all the different beds and sofas are different. Remember that audio testing such as being currently discussed is concerned only with differences, not preferences.
Does it control all variables, no! So therefore not valid
Is it just as useful as the blind tests that are bandied about on forums - more useful, IMO.
Does it give us a better handle on how we will like a device/system/sound over the long term - yes!
Is it better, in this regard, than A/B testing - yes, IMO
Indeed.
But it controls no variables, they do, so they're more valid, way more if knowledge is removed and levels matched. To hold a contrary view is faith-based, and irrational.
Agreed, though if a person wants to seek differences by way of comparisons, long-term is not the best way.
Agreed, if finding a preference is the goal, if not, see above.
answered - see Whatmore post
OK, you are saying that rapid A/B switching is not part of the testing & yet it is claimed that rapid A/B switching is the best way to identify differences. A bit of a problem, don't you think?
Ok, so now you want to change the stated advantage of A/B testing (rapid switching, which means eliminating short auditory memory issue) & change it so that it matches long term listening - just to prove a point. Kinda getting away from all the stated advantages of rapid A/B switching, isn't it?
Yes & I know that different DACs, Amps, etc are different & sound different!
As I said before - we don't make good human instruments so I do my audio testing usually to find out which I find better, which I prefer. I think most people are not involved in the form of audio testing (human measuring instruments) that you are talking about
I'm finished with this - I think you know what the answer is & don't need to ask - uncontrolled variables = what?
John, I'm simply pointing out how short-term testing can replicate long-term by way of being able to also listen to sections of music, or whole songs, as well as being able to rapid-switch. What else can long term cater for that short term cannot, in terms of looking for differences?
Bit of an umbrella statement there, and I think believe rather than know might be more apt, but fair enough
Cant argue with that, whatever works for you, etc.
The kind this thread is about, and we're all talking about, you mean
This.
John, you seem to be arguing broadly against blind listening, countering with undefined long-term listening, but all of your references to blind listening seem to be of the rapid-switiching, short listening segment ABX type. And of course there is no rule preventing you from comparing two things blind, under exactly the same conditions (except for the lack of knowledge of which item in the comparison is being played), for as long per items under comparison as you like. Is this the kind of blind listening being "bandied about" on internet forums? Which forums? This alternative has been suggested to you several times here, and you go on arguing against short-segment, rapid-switching ABX. You're right that rapid-switching ABX is best for detecting very subtle differences (differences on the cusp of audibility, I'd argue). That kind of listening bears little resemblance to long-term listening, so let's stop comparing them. Let's simply compare sighted listening to unsighted listening. Period. Let's assume, just for the sake of the discussion, that you're all set up to compare your DAC to DAC X, levels are matched, a good switching system is in place and you can push the button yourself, anytime you like, to change from A to B without knowing which is which. You can listen in short segments, whole albums, entire afternoons...switching and comparing for days, weeks, months...until you are absolutely satisfied you not only know the differences between A and B, you know which is which and have chosen your preference. Knowledge is the only variable, and we're not trying to prove anything, we're just trying to give John a better way to evaluate his own products.
How can that not be better than doing exactly the same thing, with full knowledge of which product under comparison is the one you put your blood, sweat, tears and money into?
Tim
I'm not, Tim. I was asked a specific question - "Is non-sighted long-term listening, VALID" I countered "I think you know what the answer is & don't need to ask - uncontrolled variables = what?"
Tim, you agreed to disagree on page 26 when you posted:
And as I replied then & would ask you to please accept again
Fair enough. I'll leave it to Max, who is doing fine work here. Did we agree that you were wrong? No...probably not.
Tim
Similar threads
