Side note, but does Gearslutz guy describe it as an easily sensed preference? Is it in another thread on Gearslutz? The one you linked only has 52 posts when I view it.
He seems to have gone to all the trouble of ABX to satisfy himself it was real, and actually goes out his way to reiterate how subtle it is, describing the difference as not remotely as important as moving "one microphone two inches". That's generally a small change already (it depends on the material, source and distance, admittedly - two inches across a close-miked guitar speaker is a different proposition from two inches on the ambience mic down the hallway...).
You are asking if he found it easy to come to a preference from his long term sighted listening? Really?
So what?
I believe he says this about the differences he hears which is much more like what audiophiles reports & what he focused on in his ABX listening
What's being discussed/examined here is how suitable ABX is for differentiating small differences. Are you saying it should only be used for gross differences?
Last edited:
Two of my audio friends and me took blind ABXXXXX inerconect cables test, cheap freeby vs. 20Euros one. We listened them first AB , and preference was made for 20E... Then we listened xxxxx, blind, trying to to guess which one is playing, and wrote our guessing on paper. I had 5/5 correct, my friend the same ! We repeted test once again with same result. Diferences are there if you want to find them, blind test or not...
Although maybe it could be used for small differences, its probably impractical for formal proctored testing to train large numbers of people in order to allow good statistical measurements.
Also, its probably not reasonable to ask volunteers to suffer through self-training to satisfy a few skeptics here.
ABX does seem to work find for gross differences so no reason not to continue to use it for that.
In addition, I have suggested many times it is probably possible to develop other tests, or modify existing ABX tests so that less training is necessary to demonstrate ability. Merely adding a "loop" checkbox to foobar ABX might be enough to drastically reduce a need for prolonged training to pass a test. That is, it might eliminate a prolonged period of practice for people who can already hear differences blind. For people who haven't learned how to hear some things and want to, more time might be needed for that.
I would go on to say most of the skeptics here could probably learn how to blind differentiate hi-res and CD fairly quickly. That isn't so hard. And with a low-practice blind test they could prove it quickly too.
I've said all along if one can't find a specific identifiable difference in sighted comparisons then you aren't ready to do ABX testing because that's what the test requires. But that doesn't mean that there are no differences - it's just means that you are not able to satisfy the type of listening/discipline required for ABX testing. I would suggest that not many are suitable so the casual use of ABX testing is bound to return a null result for what are consider small impairments.
Now here's the rub -what is a small impairment when forensically tested in close-up like this in ABX, can actually be a night & day difference in ordinary listening -that's a judgement call -do you consider a slightly out of focus picture just a small difference?
The night & day impression comes from the higher level emotional connection that such better focus brings to the listener - this can cascade into many different new aspects to one's listening - much more interest in the music, more realism, lack of fatigue (which the Gearslutz guy States), more believability in the sonic illusion being created by playback, etc.
Last edited:
I don't think many would disagree.
That's indeed a judgement call. Btw, "slightly out of focus" can be a small or a big difference depending how much out of focus it really is. That comparaison doesn't bring much on the table. But it has a nice emotional appeal in a debate.
That's a big claim and it would take more than one anonymous guy on gearslutz to take it as more than opinion at this point. It's an interesting hypothesis - I'll give you that - but it's hardly established.
Actually it's a lot of claims: cascading effect, link between higher fidelity and emotional impact, interest in music depending on fidelity, link in between emotional impact and lack of fatigue. All this would have to be substantiated one by one.
edit: btw, beyond the abx aspect of training, that thread on gearslutz is highly problematic. That guy could hear reliably cd-res material upsampled to higher-res. Was the resulting upsampled more or less faithful to the musical material ? It's very debatable. I'm also a bit worried by the lack of awareness of that guy on issues surrounding automatic src on windows.
Last edited:
Really, 69 pages, mostly of bickering against what I just said & you conclude "don't think many would disagree"
It's what the Gearslutz guy reported & what many who hear better sound report. A quick analysis
Readers can judge this for themselves - it's my opinion (take everything posted as IMO). Why does everything become a claim?
You miss the point - whether or not the upsampled sound was faithful or not, he had a preference for the upsampled version from long term sighted listening. He then put this preference to the test in Foobar ABX test & this is the document of that. It highlights all that I have been saying about ABX testing & its difficulty which you now seem to accept is correct.
If you agreed with this from the start why the constant push-back?
What I've stated, I've referenced with valid examples, evidence, research & my own experience so you can judge what has informed my opinion/conjecture.
AFAIR, nobody has countered what I said with evidence, research, valid examples - & people can judge their opinions likewise
Last edited:
+1
Not an unreasonable point of view, IMHO.
mmerrill99: Unfortunately, loosely mixing fact and other-than-fact is how confabulators tend to operate. Most likely better for your credibility to avoid any appearance of that.
Still plenty of opportunity to be persuasive.
This is one great thread! Thanks, Jon, for the work you've done on it. Personally the last (tiny) difference in DACs that I heard was between Sound Blaster Live! and Audigy 2ZS. Ever since that time every DAC I have had sounded about the same. I currently have two, can't tell one from the other. Say, in Audilab MDAC they allow you to choose a filter among several. Those are all very slightly different and if you switch to and fro many times you will eventually learn to hear the difference on certain music. But, if you enter a room with music playing and you will be asked to tell, what filter it is, no way you can tell. So why bother?
Nothing wrong with being sceptical, but that should still hold when analyzing/discussing experimental conditions, because methodologically flawed tests can´t corrobate a hypothesis even if the results are in accordance with the hypothesis/model.
Understood, but was the "long term assertion" meant in this way?
I think it is more related to the more "normal/casual" listening mode, we know (means there is scientific evidence) that specific conditions can be distractive. We know that it is quite easy to not see the forest due to the trees and we know that concentrating on a task usually reduces the number of other brain areas.
Working with short snippets of music to fulfill an explicit task in a test is something that surely can provoke the forrest/tree dilemma as even trying to force hearing something can.
Listening in "normal/casual" mode to music on a longer term (usually seen as state of "awareness") prevent this dilemma.
Not to mention that listening over let´s say a couple of weeks will most likely help to randomize the impact of other confounders. (remember the golden rule of testing)
Usually in these tests there is no clear defined hypothesis that will be examined, there is no systematic choice of samples, there are no controls embedded, statistical knowledge is sparse (therefore you´ll not find any calculation of statistical power before testing anything) and therefore one should be sceptical about any conclusions based on the results.
Understand, but imo everyone who has conducted controlled listening experiments with other people will have noticed that these problems exist before participants are getting used to the conditions. Science in general knows that too, therefore taking provisions to avoid it - see for example the cited ITU-R BS.1116-3 .
Understood, but - see above - there might have been a misunderstanding, as i´d think it is more related in finding a difference than to detect that difference. But even then, i assume that we all know the situation where we think "i don´t know what it is but it something is different" which is a unconscious reaction (gut feeling) and that it needs time to get a handle on it analytically.
Interesting. Of course you can refute what I am going to write, but when I switched from my Monitor Audio 2x5mm litz silvered copper to VdH The Magnum Hybrid between power amp and loudspeakers, there was such a huge difference that my wife without even having to ask her to listen, heard it immediately (more and better defined, easier to follow bass, better voices, and so on).
Of course this may have been due to the fact that I was using a classic Naim power amp at the time and they work much better when the cable has a much lower capacitance, but fact is that, to the very least, cables, as electrical interfaces between components, have properties that can affect how these work.
Of course, to some it is Naim power amps that are "defective by design" because of the way their output is implemented.
If we are comparing decent cables that are similarly designed, however, I tend to believe that spotting differences may be very difficult, if at all possible.
Roberto
I guess you are lucky. I am not pulling your leg. The moment you start hearing things your bank account will start to drain.
Roberto
I guess I'm lucky too, I tend to listen to the music, and I consider getting hung up on anything else an impediment
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.