oh dear... "Cable Pathways Between Audio Components Can Affect Perceived Sound Quality"

[CMOT]

I haven't read it yet, but urg. The skimming I did revealed this gem: "
In the hope that some audible difference will be detected, the cables in the two configurations were chosen to be as different as possible. With its much shorter length, bal- anced (versus single-ended) topology, and faster dielectric the higher-end cable A can be expected to have a more de- tailed and accurate sound. The question was whether inter- changing these interconnects would produce a recognizable and memorable timbral change that would be discernable in blind listening tests."

Bias to begin with anyone?

Also, it looks like there were only a three observations per subject, which is very bad to start with. No power (statistically speaking). The statistics that were run look a bit wonky. I haven't used Wilcoxon since grad school, but it might be okay as far as it goes. I can't tell. I wouldn't have analyzed things that way. Instead a simple between subjects anova or something....

In any case, have "fun" with this!

https://doi.org/10.17743/jaes.2021.0012

Cable Pathways Between Audio Components Can Affect Perceived Sound Quality
MILIND N. KUNCHUR
([email protected], [email protected])
University of South Carolina, Columbia, SC 29208, U.S.A.
The arena of highest fidelity in music reproduction, sometimes referred to as high-end audio, has many controversial claims and contentious issues. One such controversy is whether the cables and topology used to interlink components together make an audible difference. There seems to be a disparity between anecdotal experiences reported by audiophiles and published formal scientific research as to the minimal changes in system configuration that can be audibly distinguished. With the motivation of bridging this divide—which may originate from differ- ences in instrumentation and subject-listening conditions used by the two groups—this work utilized a high-performance audio system and extended-duration listening protocol that more closely resembles audiophile auditioning conditions. With these measures the present work was able to prove through direct psychoacoustic testing that two different analog-interconnect pathways can be audibly distinguished.

 

[pozz]

http://boson.physics.sc.edu/~kunchur//Acoustics-papers.htm

Paper available there. The contents are beyond me to assess.

 

[GGroch]

Thanks for posting this. I hope that AES members who have access to the paper (or non-members who are not cheapskates like me) will comment on it after reading it in full.

The bias of the researcher does not matter much to me. Cancer researchers who hope their research helps cure disease can still run scientifically valid tests. Non-members do have access to the comments, one indicates that the test just proves very long unbalanced cables may have audibly more noise than unbalanced ones, which is not a controversial conclusion. The devil is in the details.

UPDATE: pozz,,,thanks for the link. I will read it and update.

 

[Chrispy]

FWIW there are some free links posted up in the Audio Bullshit group on FB for some of Kunchur's stuff. Seems he's a professor that teaches some basic physics classes as well as audiobullshit classes .

ps the free links were to something he published for students to read for preparation for his classes....so might find it that way.

pps seems Pozz already has the link for the free stuff

 

[audio2design]

Already a thread on this.

 

[mansr]

KUNCHUR

You can stop reading there.

 

[BrEpBrEpBrEpBrEp]

This study compared a balanced to a single ended cable, with different amplifiers. They measured different noise levels. In other words, it says absolutely nothing at all about cables.

 

[mansr]

This study compared a balanced to a single ended cable, with different amplifiers. They measured different noise levels. In other words, it says absolutely nothing at all about cables.

FTFY

 

[Kal Rubinson]

This study compared a balanced to a single ended cable, with different amplifiers. They measured different noise levels. In other words, it says absolutely nothing at all about cables.

Same amplifier. OTOH, while it has both balanced and unbalanced inputs, it also has two different input stages.

 

[BrEpBrEpBrEpBrEp]

Same amplifier. OTOH, while it has both balanced and unbalanced inputs, it also has two different input stages.

Ah, I was just going off the conclusion:

So they seem to claim an extra/different amplifier in the signal path for the balanced cable. Maybe I misunderstood... regardless, the study doesn't seem to say very much useful, other than that balanced and single-ended signal paths can perform differently.

Edit: my mistake for not reading carefully enough - the additional amplifier was only for time-domain measurements with an oscilloscope.

 

[Kal Rubinson]

So they seem to claim an extra/different amplifier in the signal path for the balanced cable. Maybe I misunderstood... regardless, the study doesn't seem to say very much useful, other than that balanced and single-ended signal paths can perform differently.

The extra amplifier in the signal path would have been required for them to be able to measure time-domain with an oscilloscope. OTOH, the test amplifier's balanced input stage is, of necessity, different from the unbalanced input stage and that is another variable in the test paradigm.

 

[BrEpBrEpBrEpBrEp]

The extra amplifier in the signal path would have been required for them to be able to measure time-domain with an oscilloscope. OTOH, the test amplifier's balanced input stage is, of necessity, different from the unbalanced input state and that is another variable in the test paradigm.

Whoops... that's what I get for not reading carefully enough. Thanks.

 

[PaulD]

What a pity that the AES publishes this flawed material...
Similar the the POZZ link, I found another link to the paper on a forum, but this has a limited lifespan. https://wetransfer.com/downloads/49...c08c348a221306b92b8c4d7b20210616034436/affa3e

 

[amirm]

This study compared a balanced to a single ended cable, with different amplifiers. They measured different noise levels. In other words, it says absolutely nothing at all about cables.

It is a shocking error that got past the peer review. They should have caught this and not accepted it. There are other problems with the paper that I will post about when I get the time.

 

[amirm]

I haven't read it yet, but urg. The skimming I did revealed this gem: "
In the hope that some audible difference will be detected, the cables in the two configurations were chosen to be as different as possible. With its much shorter length, bal- anced (versus single-ended) topology, and faster dielectric the higher-end cable A can be expected to have a more de- tailed and accurate sound. The question was whether inter- changing these interconnects would produce a recognizable and memorable timbral change that would be discernable in blind listening tests."

This was another shocker to read in the Journal of AES. I think the paper may merit a formal complaint and response to AES for publication in future Journal.

 

[PaulD]

This was another shocker to read in the Journal of AES. I think the paper may merit a formal complaint and response to AES for publication in future Journal.

Please do complain, publishing this material is shockingly poor. I have complained about various poor papers several times and my complaints about the peer review process and what is published have been ignored. I review papers for several journals and conferences, and the AES has the most error prone system that I have observed. For the AES to regain credibility they must retract flawed papers publicly as other journals and disciplines do.

 

[DimitryZ]

This was another shocker to read in the Journal of AES. I think the paper may merit a formal complaint and response to AES for publication in future Journal.

I respectfully disagree.

The objective of the experiment is clearly stated. There is nothing wrong with that. Test methodology and test plan are then developed and executed to support the test objective.

I would note that the objections to "audible cable difference" are now more carefully phrased as "cables that measure similarly, sound the same," or something along similar lines.

If I wanted to demonstrate that the commercially available analog pathways *can* make an audible difference, it is methodologically correct to choose very differently measuring cables.

The entire reproduction chain remained unchanged, except the position of the input selector switch.

 

[pkane]

The entire reproduction chain remained unchanged, except the position of the input selector switch.

That's not the case. Different transmission, different topology, different active circuits are engaged when switching between these two cables. A better test would be to switch between two RCA or two XLR cables. Comparing apples to oranges and declaring they are perceived as different isn't a big revelation. Especially since balanced was created specifically to deal with noise over single-ended transmission.

 

[DimitryZ]

That's not the case. Different transmission, different topology, different active circuits are engaged when switching between these two cables. A better test would be to switch between two RCA or two XLR cables. Comparing apples to oranges and declaring they are perceived as different isn't a big revelation. Especially since balanced was created specifically to deal with noise over single-ended transmission.

I am sorry, but I do do a lot of test planning and execution for a living.

The first step in designing a test plan is to define your UUT (Unit Under Test). In this case the UUT is the "analogue pathway" from preamp to amp. There are many different variations in the design of such pathways, and they definitionally include different cable topologies and SE and balanced transmissions.

You may of course disagree that this test's UUT is defined too broadly - but there are no standards to apply here - the researcher gets to make this decision.

Now, if you start with the above definition of the UUT and have taken on a test objective to demonstrate that there could be an audible difference within commercially available solutions, it would in fact be imperative that you would use a "corner" case to prove your hypothesis. Therefore, cables of different topologies, transmission standards and lengths are perfectly valid experimental choices.

If one thinks of the analogue pathway design space as a three dimensional cube, with three axes being topology, length and transmission standard, the author set out to prove that the case of maximum variance is audible, and he had done so. This should not be surprising, especially for those of us familiar with Interlink 400 (Cable B), which has an odd sound signature in many systems.

The much harder to answer question is where this audibility threshold really is. It maybe that once you move away from the maximum variance, the audibility quickly disappears.

 

[pkane]

I am sorry, but I do do a lot of test planning and execution for a living.

The first step in designing a test plan is to define your UUT (Unit Under Test). In this case the UUT is the "analogue pathway" from preamp to amp. There are many different variations in the design of such pathways, and they definitionally include different cable topologies and SE and balanced transmissions.

You may of course disagree that this test's UUT is defined too broadly - but there are no standards to apply here - the researcher gets to make this decision.

Now, if you start with the above definition of the UUT and take on a test objective to demonstrate that there could be an audible difference within commercially available solutions, it would in fact be imperative that you would use a "corner" case to prove your hypothesis. Therefore, cables of different topologies, transmission standards and lengths are perfectly valid experimental choices.

If one thinks of the analogue pathway design space as a three dimensional cube, with three axes being topology, length and transmission standard, the author set out to prove that the case of maximum variance is audible, and he had done so.

The much harder to answer question is where this audibility threshold really is. It maybe that once you move away from the maximum variance, the audibility quickly disappears.

There's no value, at least in my opinion, to doing research into an area where the differences are not only expected, but exist by design. My comment wasn't about whether Kunchur muddled the experiment (he did), but that he didn't produce anything new or useful, except for a bunch of unsupported conjectures in his conclusion.