[artpiggo]

Hi All, First time using this forum section.

I want to ask for everyone's opinion.

When the debates are about digital / bits-are-bits or anythings that most people called as voodoo / mystical stuff that can't find any single scientific / graphical / measurement detail to support their claim.

But if the claimer him/herself can do 100 out of 100 times of correct ABX test with judge (who is outsider,non-related person to him/her or the judge can be you yourself in case you are not sure if the judge is being paid or not) when you also are witness in that room and see by your own eyes that he/she didn't do any tricks and still totally answer all of 100 times correctly.

Methodology and validation
- You also check bit-bit of the tested files with the reliable and accurate program.
- you check the room walls, judge, computer, source anything.
- You can ask whenever you think experiment is biased.
- you also makes the file by yourself in same methodology of their claim and let them retest

You found nothing wrong there. Claimer still answers correctly. He/she just cant prove supporting scientific research.

Then:
Do you think that by doing this and still get 100% correct in stastically large-enough times of trial, does it convince you to believe that his claim is right and no explanation needed anymore.

Or you still doubt it unless he/she can explains in scientific way.

Thanks for your comments.

 

[Strangelove424]

Science doesn't deal well with the purely hypothetical. I think this is more of a personal question than anything else, about levels of proof required, and what the purpose of the proof is. If David Copperfield was standing next to two identical amps, and each time he touched one amp the sound magically had better tone and richness than the other, but only while he touched it, and I was listening to this myself, I still wouldn't believe it without a scientific explanation. If this was at a magic show, I would say to myself, "wow, there's no way to explain it to myself, but hey good show and I had a decent dinner."

If, however, someone was trying to sell me David Copperfield as a sound-improvement item for $2,000, I would become extremely suspicious, even if I heard it with my own ears. I would want to see lots of data about David Copperfield's physical effects on the amp, x-ray and mri scans of his body, electrical readings of his finger tips, certification of these measurements, and an explanation of how the phenomenon I witnessed took place. That's not to say David Copperfield was lying to me, it's just that he never gave me anything fundamentally true.

 

[castleofargh]

if somebody can pass an ABX test(no need for 100% to pass), we can conclude that the subject can tell there is a difference between 2 tracks playing on his system. but to draw any conclusion about the cause of the audible difference, we need to be able to assume that the variable tested is the only difference between the 2 files. for example if we test bit depth variations and one file starts half a second after the other, or is clearly louder, we're not testing bit depth anymore.
so some people might wish to know how the files were created and maybe get them to try and replicate the test themselves. as is customary of any good scientific experiments, yet not done nearly enough. most of the time nobody will bother checking all that, but of course the more unexpected a result will be, the more skeptical people will become. that's only natural.

 

[pinnahertz]

The first thing to do would be to verify that the test is a true ABX, and that all the bias controls are in place and working. So many people think they're doing an ABX but it has built in tells that bias the results. And you have to be very careful about exactly what you're comparing. For example, comparing different sampling frequencies is actually very difficult because what you are really comparing is the sampling frequencies and the performance of a DAC and possibly an ADC running at those frequencies along with the associated filtering.

It is possible to score very highly on an ABX test and still be wrong without the proper controls and understanding. I can't accept that score as conclusive without a thorough examination of the entire test setup and methodology. Just saying it's ABX is not sufficient.

 

[artpiggo]

The first thing to do would be to verify that the test is a true ABX, and that all the bias controls are in place and working. So many people think they're doing an ABX but it has built in tells that bias the results. And you have to be very careful about exactly what you're comparing. For example, comparing different sampling frequencies is actually very difficult because what you are really comparing is the sampling frequencies and the performance of a DAC and possibly an ADC running at those frequencies along with the associated filtering.

It is possible to score very highly on an ABX test and still be wrong without the proper controls and understanding. I can't accept that score as conclusive without a thorough examination of the entire test setup and methodology. Just saying it's ABX is not sufficient.

What if ABX test is verified with methodology, system set up and all evidence. You can also be QAQC guy asking for defects of experiment and let them retest in whatever way you want too. Then what's next? Is it suffice or not yet in your opinion.

 

[pinnahertz]

What if ABX test is verified with methodology, system set up and all evidence. You can also be QAQC guy asking for defects of experiment and let them retest in whatever way you want too. Then what's next? Is it suffice or not yet in your opinion.

In general then I would agree with castleofargh, it should suffice, as long as conclusions are specifically related to those exact conditions.

 

[bfreedma]

What if ABX test is verified with methodology, system set up and all evidence. You can also be QAQC guy asking for defects of experiment and let them retest in whatever way you want too. Then what's next? Is it suffice or not yet in your opinion.

Has you scenario actually occurred, or is this a theoretical discussion?

 

[artpiggo]

Has you scenario actually occurred, or is this a theoretical discussion?

Theory discussion.

 

[Speedskater]

Good measurement is all about measuring differences. If ABX differences are heard, then there are meaningful measurement differences. But it can take skill and good test equipment to make the correct measurements.

 

[sonitus mirus]

Good measurement is all about measuring differences. If ABX differences are heard, then there are meaningful measurement differences. But it can take skill and good test equipment to make the correct measurements.

Agreed.

At some level of detail, nothing is identical and there is always some difference. For many of these tiny differences, we could probably devise a means to measure them if it didn't already exist.

Additionally, there could be measurable differences identified, though they may generally be considered to be inaudible to humans. This is another area where an ABX could be utilized to help verify. This is not to suggest that a simple ABX test someone like myself may perform at home is the authority on the matter, but a more rigorously controlled ABX is currently the best method that is available at this time. The personal ABX, when attempts are made to remove bias as best as possible, can be an excellent tool for quashing those ideas where obvious differences are supposedly being heard where no measurements would indicate that such a difference would be possible to hear.

ABX is critical to the advancement in many areas of science and is an effective means for removing most bias in helping to determine if an audible difference is actually being heard, though there are detractors.

I've thought I heard an obvious difference, proved to myself that I was not actually able to hear a difference through ABX, and then I was not able to perceive that difference anymore. Clearly I was capable of overcoming the illusion and bias I was experiencing.

As complex as our cognitive behavior might be, this same result may not be possible with much finer details, but I'm personally not interested in identifying to that level for differences. I'm not going to listen in some unusual way or create some other pathological criteria to try and barely eek out some statistical evidence that some difference may be likely under certain conditions. It's music, I want to listen and enjoy it.

 

[Guidostrunk]

I feel that if you have to torture yourself to hear an audible difference in something. You're probably better off not even bothering. It completely overshadows the enjoyment of listening to music.

Unless of course that person enjoys torturing themselves more than just listening to the music.

Seems like a lot of wasted effort to do a "professional" controlled abx test that's perfectly implemented.

Like the above post. IT'S MUSIC. lol