[Baxide]

   
If you are claiming on a science forum then absolutely you need to prove that the difference you hear is due to an actual audible difference rather than subjective bias. This can be done through ABX testing etc. Without that essential evidence there isn't even any point in putting forward any conjecture of what could cause such differences. Of course you are entitled to your opinion, just realise that it isn't a scientific opinion and therefore will not carry any weight on a science forum.
 
[]
 
Throughout human history science has had to battle commonly held beliefs, but it has always eventually won out. People used to believe that you could fall off the edge of the world; that the Earth was the centre of the universe with everything rotating around it; that the world was created by a mystical being 6000 years ago; etc.

If you do ABX testing, you are having to put your trust in another set of variables that connects up the different pieces of equipment. I was once confronted by that problem, which eventually turned out to be the amps. They were two identical amps, but one was placed in a different position in a rack, where it got slightly hotter due to the ventilation. That in itself changed the sound. But it did make everyone think that the ABX test was proven beyond any reasonable doubt, when it should not have.
 
As for the other case: it all depends on what one would consider to be the edge of the world. If you are on a mountain and stepped away from it, you could fall off. And the earth is at the center of the universe as seen from our perspective. By just looking up at the sky you cannot say where the centre of the universe is. You only have earth as a reference, and in science it is all about "with respect to"...

 

[Jabbah]

  If you do ABX testing, you are having to put your trust in another set of variables that connects up the different pieces of equipment. I was once confronted by that problem, which eventually turned out to be the amps. They were two identical amps, but one was placed in a different position in a rack, where it got slightly hotter due to the ventilation. That in itself changed the sound. But it did make everyone think that the ABX test was proven beyond any reasonable doubt, when it should not have.
 
As for the other case: it all depends on what one would consider to be the edge of the world. If you are on a mountain and stepped away from it, you could fall off. And the earth is at the center of the universe as seen from our perspective. By just looking up at the sky you cannot say where the centre of the universe is. You only have earth as a reference, and in science it is all about "with respect to"...

 
Another part to validating a theory is repeatability*. A couple of years ago a group thought they had detected faster than light neutrinos, turned out to be a faulty connection. This caused a massive commotion but nobody could reproduce and eventually the fault was found, as you found yours.
 
The mountain situation doesn't really apply as you can still see the ground and horizon. You can certainly make the the earth the centre of our solar system and the universe but geocentric models are massively more complex that heliocentric ones. The simplest theories that support all the observations win.
 
ETA: *Assuming someone is genuinely investigating audible artefacts through ABX testing then the test can be setup to minimise other variables through swapping components and ensuring that every combination is tested. The statistics of these tests can then be analysed to find the effect from the variable you are testing for.

 

[DrKC]

There are many fields for which the public has been loaded with untruths turned into facts through repetition by the media, their peers, by authorities, etc. Or at the least, their knowledge level is only up to the level where the argument with the best buzzword compliance (like the argument given for the effect of DAC lowpass filter on height perception) wins.

I don't know where this discussion will end - I've enjoyed reading it.  Joe's comments above seem to represent our current state with most internet discussions.  The more times and the more places something is repeated, the "truer" it becomes.  I do find it refreshing to see someone like Joe iterating this position.  His products are often those that are subjects of such limited-knowledge discussion.   Perhaps it's inappropriate, but cheers for you Joe.

 

[castleofargh]

 

Exactly. Just means some people need educating where the correct burden of proof lies. In this case all the scientist has to do is ask for the evidence of the cruisers.

 
This will also depends on what has been claimed. If I claimed I heard a difference in the cable, do I have to prove the difference exists or what caused the difference. It is the explanation that people refuse to accept. The burden of proof does not help in this case. Am I not entitled to an opinion without a scientific education? This is why the defense always claimed no evidence to contradict my belief. Bias is hard to overcome.
 
Take climate change; common defense of politician: "I am not a scientist so I don't believe climate change exist.
Evolution: There is no mathematical model to back up the claim, therefore evolution does not exist. This is an actual claim in a lawsuit against the education department.
Not only there is no burden of proof, it is basically I don't believe in science. It is I trust my pastor more than the liberal scientist.Oh the pope believes it. "Then he should shut up", said Rick Santorum.

(for the part on bold) you have to prove exactly what you claimed. when there is too much to prove, it's not a failure in the burden of proof but simply that you claimed too much ^_^.
 the burden of proof is a checkpoint into a conversation. if a claim can be substantiated, then we keep on. if not, then it stops right there until the person making the claim can bring some matter of evidence of what he claimed.
it's a weapon for not wasting time talking about something when nobody can even prove if it's real.
 
 
and of course everybody's entitled to an opinion with or without a scientific education. but an opinion, not a claim!
 
and of course the evolution bit, is a beautiful example of fallacy.

 

[RRod]

  I don't know where this discussion will end - I've enjoyed reading it.  Joe's comments above seem to represent our current state with most internet discussions.  The more times and the more places something is repeated, the "truer" it becomes.  I do find it refreshing to see someone like Joe iterating this position.  His products are often those that are subjects of such limited-knowledge discussion.   Perhaps it's inappropriate, but cheers for you Joe.

 
"A crime persevered in a thousand centuries ceases to be a crime, and becomes a virtue. This is the law of custom, and custom supersedes all other forms of law."
“A lie can travel half way around the world while the truth is putting on its shoes.”

 

[Joe Bloggs]

Speaking of which, one of the background facts given supporting the effect of DAC lowpass filters on height perception was that ringing from the lowpass is audible.


Since ringing is confined to the transition band of the lowpass filter (namely 20-22kHz for 44.1kHz audio, >40kHz for 96kHz audio, etc...), he'd need first need to demonstrate audibility of those ultrasonic frequencies in the listeners--at the volumes in which they occur in music (we'll ignore masking for now)

Something only the claimant can do through say an ABX test. If you think that there isn't going to be a difference then that bias could affect the ABX test and generate a negative result. If you think there is going to be a difference, that bias can't positively affect the ABX outcome.

Not talking about an ABX test--in the worst case (44.1kHz audio) a demonstration of audibility of a 20kHz pure tone *at levels present in the actual music at listening volume* would suffice for me.

 

[Joe Bloggs]

The purpose of this thread is not to bemoan the knowledge level of the general public, but to point out that as a practical matter, the burden is on you to prove to your audience that their hypotheses are unproved. Actually, sometimes what appears is actually two conflicting theories requiring proof from both sides.

Take the usual ABX flamewar. What you have here is actually one theory (that the equipment produces the perceived difference) vs another (that psychological factors produce the perceived difference). To your target audience, at least, you have just as much burden to prove your position--

1. level matching (and not just by hearing)
2. blinding

Evidence has to be given for the importance of both prerequisites for your position to be convincing.

As another example, I need to come up with evidence that the ringing phenomenon only occurs at the supersonic transition band (e.g. in the form of spectrograms) for my assertion in the posts above to be convincing to laymen.

 

[Jabbah]

It’s useful to differentiate observations from theories here. First you have to quantify what you can hear, then you can theorise about why you can hear a difference if it is there. So taking your ABX conflict “one theory (that the equipment produces the perceived difference) vs another (that psychological factors produce the perceived difference)”, before you even get to those theories you have to confirm the evidence, hence the ABX test. Seems that most people already have preconceived theories of why you can hear things hence this confusion.

Due to ABX testing being subject to negative influence but not positive influence, this would need to be performed by those that claim a difference can be heard. Also, you can’t prove a negative, so a false ABX test would not confirm that there is nobody that can hear the claimed difference whereas a positive one would confirm that some people can.

Your example for showing ringing phenomenon at supersonic transition bands would certainly be useful in convincing the layman that there should be some doubt cast on the claims, as would be your suggestion of demonstrating the audibility of a 20kHz signal at comparable levels to ringing phenomenon. In each case though you would need to show that those cues, even if considered inaudible*, do not influence the listener in any way. And how are you going to do that…? ABX? I’m sure you could do a much better test using an MRI scanner and a bunch of test tones, but that might be a little beyond our means.

*Our eyes are sensitive to a very narrow band of the light spectrum, however we can certainly feel infrared as heat. This is more akin to infrasonic, but there is evidence that high intensity ultrasound can be heard if fed directly to the skull, bypassing the middle ear. Maybe there are other vectors that would allow subtle effects of ultrasound to influence our perception? /DevilsAdvocate

 

[RRod]

*Our eyes are sensitive to a very narrow band of the light spectrum, however we can certainly feel infrared as heat. This is more akin to infrasonic, but there is evidence that high intensity ultrasound can be heard if fed directly to the skull, bypassing the middle ear. Maybe there are other vectors that would allow subtle effects of ultrasound to influence our perception? /DevilsAdvocate

 
Well you have things like Ōhashi's papers on non-aural perception of ultrasounds, which claim evidence for such phenomena. So then we get into the "which journal articles do we believe and don't believe" hullabaloo.

 

[Jabbah]

One of the requirements for a scientific theory to be accepted is repeatability. AFAIK no one was able to reproduce Ohashi’s results and this paper refutes the findings and details why:

http://www.aes.org/e-lib/browse.cfm?elib=10005

ie The additional ultrasound signal might have affected the sound produced in the audible range from a single speaker through non-linear interactions.

Of course, in headphones there is only one speaker per channel, so this may confirm that an audible difference may be heard on some headphones if there is an ultrasonic component to the signal. But then that would be due to the speaker unit rather than the ability to hear ultrasonics at those levels, and again you come back to the requirement for a double blind test to determine if audible differences are apparent.

 

[castleofargh]

The purpose of this thread is not to bemoan the knowledge level of the general public, but to point out that as a practical matter, the burden is on you to prove to your audience that their hypotheses are unproved. Actually, sometimes what appears is actually two conflicting theories requiring proof from both sides.

Take the usual ABX flamewar. What you have here is actually one theory (that the equipment produces the perceived difference) vs another (that psychological factors produce the perceived difference). To your target audience, at least, you have just as much burden to prove your position--

1. level matching (and not just by hearing)
2. blinding

Evidence has to be given for the importance of both prerequisites for your position to be convincing.

As another example, I need to come up with evidence that the ringing phenomenon only occurs at the supersonic transition band (e.g. in the form of spectrograms) for my assertion in the posts above to be convincing to laymen.

I really don't agree with that.
hypothesis are free for all to try and explain or deny, but claims are always the responsibility of the guy making them.
if I contest a claim by another claim, then it becomes my job to prove my own claim, but that in no circumstance removes any responsibility from the other guy proving his own claim.
here you're making an excuse for the ignorant to keep on making empty claims by shifting the responsibility. and that's really not ok. when I don't know something I ask questions. my sentences end with a question tag.
if I don't know stuff and pretend like I do I'm just a fool. I could be an arrogant fool, or a naive fool, but still a fool making a claim when I don't have a clue what I'm talking about.
 
obviously I'm ok with the naive one who comes genuinely thinking he knows something. as long as when others explain to him why he's wrong, he listens, and goes read on the subject. I'm that naive guy more often than I can count. you can't know what you don't know. but while the other guys are kind enough to explain stuff to me, and offer theories as to why I thought what I thought, the burden of proof never left my back. I made the claim, it's mine! and my ways out of that mess are to either provide evidence of my claim(then obviously we don't have any problem), or to withdraw my claim. at least until I find a valid reason to justify it.
obviously "my wife heard it in the kitchen" not being recognized as proof by science just yet.
 
each time we go to the third path where the guy keeps on spamming his point while ignoring everybody else, it's not a discussion anymore, it's stupidity and propaganda.

 

[RRod]

One of the requirements for a scientific theory to be accepted is repeatability. AFAIK no one was able to reproduce Ohashi’s results and this paper refutes the findings and details why:

http://www.aes.org/e-lib/browse.cfm?elib=10005

ie The additional ultrasound signal might have affected the sound produced in the audible range from a single speaker through non-linear interactions.

Of course, in headphones there is only one speaker per channel, so this may confirm that an audible difference may be heard on some headphones if there is an ultrasonic component to the signal. But then that would be due to the speaker unit rather than the ability to hear ultrasonics at those levels, and again you come back to the requirement for a double blind test to determine if audible differences are apparent.

 
My point was more that, in isolation, you could call that paper and Ōhashi's as just refuting each other. We can keep going, like if I posted this paper:
http://www.aes.org/e-lib/browse.cfm?elib=15398
yet another paper that generated much talk but whose methods seemed to generate some serious questions. You could then post Meyer & Moran, which was claimed to have been "debunked" by a paper by the Meridian people. It goes on and on. Both sides have their edifices to stand on, it just comes down to which one actually ends up being a garbage heap. Part of the problem on this board is that people want to reject the methods, like ABX, commonly used to determine if stuff is garbage.

 

[superjawes]

...What you have here is actually one theory (that the equipment produces the perceived difference)...

Without numbers to back this up, this is a hypothesis, not a theory, hence the burden of proof lies on the individual making the claim to either find the numbers or create them through proper testing.

... vs another (that psychological factors produce the perceived difference)...

This would be a response to the lack of data. "Skeptics" present this as a test that others can use to prove out the hypothesis mentioned above. It is presented because many claims have been made that various pieces of equipment sound different despite the numbers we have on hand. Basically, the theories we use to analyze the equipment say that each piece is audibly transparent, so when someone "hears" a difference, he/she should perform a test to isolate one's personal bias and expectations from the process overall.

Take cables for example. I design wire harnesses for a living, so I know a thing or two about conductors. They are very simple things, and as long as the correct gauge is chosen for the length and application (and the cable isn't broken in some other way), then the wires will do their job, and that means audio transparency. Because the specs vary so little between cables (and said variations are usually below audible perception), then I assume that anyone claiming to hear a difference has not isolated his/her expectation bias from the equation. So I would recommend a testing session to ensure that the expectation bias is eliminated.

 

[castleofargh]

 

One of the requirements for a scientific theory to be accepted is repeatability. AFAIK no one was able to reproduce Ohashi’s results and this paper refutes the findings and details why:

http://www.aes.org/e-lib/browse.cfm?elib=10005

ie The additional ultrasound signal might have affected the sound produced in the audible range from a single speaker through non-linear interactions.

Of course, in headphones there is only one speaker per channel, so this may confirm that an audible difference may be heard on some headphones if there is an ultrasonic component to the signal. But then that would be due to the speaker unit rather than the ability to hear ultrasonics at those levels, and again you come back to the requirement for a double blind test to determine if audible differences are apparent.

 
My point was more that, in isolation, you could call that paper and Ōhashi's as just refuting each other. We can keep going, like if I posted this paper:
http://www.aes.org/e-lib/browse.cfm?elib=15398
yet another paper that generated much talk but whose methods seemed to generate some serious questions. You could then post Meyer & Moran, which was claimed to have been "debunked" by a paper by the Meridian people. It goes on and on. Both sides have their edifices to stand on, it just comes down to which one actually ends up being a garbage heap. Part of the problem on this board is that people want to reject the methods, like ABX, commonly used to determine if stuff is garbage.

in such situations, the proper behavior is then not to make a claim. or to make claims that are more specific so that we can zero in on a particular test, or a particular set of abx.
it's very ok not to make a general claim that is bound to be untrue under specific conditions. we all know that with some gears, high res or content with a lot of ultrasound will indeed have audible changes in the 20hz-20khz range. we may decide to rule those devices out as defective because they just suck too much at keeping signal fidelity, but on those defective devices, ultrasounds do make audible change.
 
cables are the same, it's not hard to build a really stupid cable that will make audible differences. again to us, because we think that transparency is best choice, we would through that cable away. but it does make a case for cable differences being audible.
in both situations it's just as wrong to claim there will never be audible difference, just like it is wrong to claim that you can improve you music with cables and ultrasounds. both claims would be untrue and could be disproved.
now if instead we focus on gears that are not too far off of transparency, then claiming no audible difference once volumes are matched for those devices seems a lot more reasonable.
if something is too complicated we just have to reduce the external factors and make a claim within those factors if we can prove it.

 

[Jabbah]

My point was more that, in isolation, you could call that paper and Ōhashi's as just refuting each other.

No, you really couldn't. Kiryu et al cover a superset of tests including Ōhashi's and show conclusively that Ōhashi's conclusions were based on incomplete data and providing further evidence that contradicted those conclusions. There is no way that you could reverse this and say that Ōhashi refutes Kiryu.

We can keep going, like if I posted this paper:
http://www.aes.org/e-lib/browse.cfm?elib=15398
yet another paper that generated much talk but whose methods seemed to generate some serious questions. You could then post Meyer & Moran, which was claimed to have been "debunked" by a paper by the Meridian people. It goes on and on. Both sides have their edifices to stand on, it just comes down to which one actually ends up being a garbage heap. Part of the problem on this board is that people want to reject the methods, like ABX, commonly used to determine if stuff is garbage.

It's pretty clear if a paper debunks another, because it replicates the previous behaviour but then builds on that giving evidence that refutes the theory and conclusions of the previous paper. Going back to Popper:

"Popper's theory presents an asymmetry in that evidence can prove a theory wrong, by establishing facts that are inconsistent with the theory. In contrast, evidence cannot prove a theory correct because other evidence, yet to be discovered, may exist that is inconsistent with the theory."