Is cable sound real? A more holistic approach trying to track it down.

[SIY]

Are you familiar with Pano’s potato and banana tests?

 

[pma]

(off for dinner for a while)

You might put the cables against the original file. As the people don't succeed in original x tube preamp ABX, I strongly doubt they will get anything from different cables. You will get some sighted impressions (just impressions) and NO valid positive ABX result.

 

[watchnerd]

I'd like to intervene here and mention one important thing: subjectivists hold their position firmly, because they care that they hear the difference when the test is sighted and they don't care about blind testing. Objectivist - mostly on ASR - position holds, that unless gear measures differently, it makes a difference.

However, in real use cases, we'll probably not encounter unsighted listening (i.e. at home) that often, so perhaps to abstain from being either an objectivist or subjectivist, we should intend that it is recognized both that subjectivist approach (that sighted listening might cause different hearing due to biases) and that objectivist approach hold at the same time (that the same measuring gear that is heard different when sighted and not heard the same when blind).

This part I understand:

"that sighted listening might cause different hearing due to biases"

But this part:

"that the same measuring gear that is heard different when sighted and not heard the same when blind"

Isn't that just making the same point as the first one and thus the commonly accepted reason why we do ABX tests?

Or am I missing something?

 

[watchnerd]

Hi everyone,

We all know the myriads of testimonials about cable sound obtained from sighted listening and completely uncontrolled conditions (level matching etc), which makes them basically moot. Then again, the sheer number of reports would statistically allow for at least a handful of setups where some form of cable sound actually did happen and could have been proven with the right means at hand.

Generally, I'd like to divide potential cable sound -- talking line level interconnects here as for now -- in two categories:

Category A, trivial cable sound: this is what comes from using a cable simply not up to the task, and/or from side effects not actually contributed by the cable itself, apart from its basic properties. Some examples would be:

• using a very long high capacitance cable on a very high impedance output (passive preamps and such).
  
• using long unbalanced cables, or ones with high shield resistance, or even unshielded „return wire“ types in setups with „ground loop“ problems. Same with cables that have open shield connection on one end.
  
• using badly shielded, or even unshielded cables in general picking up external electrical and magnetic fields, and notably in setups where the cable connects to devices that suffer from EMI problems (RF demodulation)
  
• ...

Category B, what I would like to coin „intrinsic“ cable sound. This refers to the claims that the construction of the cable is making a difference even when the simple lumped parameters relevant for interconnects -- capacitance, inner conductor(s) and shield resistances -- are about the same (if that even matters, that is). Some examples would be:

• silver conductors sound different than copper conductors
  
• high purity copper / silver conductors sound different than off-the-mill copper / silver
  
• single-strand (solid-core) conductors sound different than multi-strand
  
• plated conductors sound different than non-plated
• isolation materials matter (dielectrica)
  
• geometry details matter
  
• ...

We would only be interested in Category B here, the „intrinsic cable sound“ claims.

Attempts have already been made trying to verify cable sound in more controlled manners, like the required level-matched blind comparison, and occassionally some positive results may actually have been found. Alas, the setups usually were not prepared for a technical investigation, on-site and in-situ, as to what causes the differences, technically. More simply worded, it is a futile attempt to measure the cables under completely different conditions than those that applied during the blind tesing (assuming it created positive results). You might say when we are after intrinsic cable sound the circumstances should not matter much and the „characteristic“ of the cable should manifest in the lab, and of course this should be true. Then again, when a BT yields positive results wrt to differences percieved this sort of guarantees -- within the limits of the statistical significance the test allows for -- that there must be actual differences in the signal and it would be very nice and insightful to peel out these difference in the actual music material that was used for the test.

Cable Blind Tests in real life, using people's actual HiFi rigs, are a complicated matter. But in this age and day of the internet, there is a way to remote the tests to a large group of participants by using loop-back recordings. While not exactly recreating the situation of a real cable change in one's system, chances are there that the captured differences can still be identified. Plus we have the great opportunity that the circumstances are extremely well defined and replicable and everything is in one hand, allowing for meaningful analysis as the setup for the recordings is exactly the same for the technical investigation and the recording for the blind testing.

So what I have in mind is the following:

• I will make precision and well-documented loopback recordings with all the necessary care, using different cables (I have some „higher quality“ XLR and RCA interconnects here on loan for this, besides my bog-standard interconnects). The RME ADI-2 Pro, USB connected, is an obvious choice for a DAC-->cable-->ADC loop-back setup, for making the recordings, plus I actually happen to have two of those interfaces so that more „real-world“ scenarios than simple loop-back within a single device can be tested as well if that is deemed necessary for any sort of reason.
  
• Phase 1, Getting Familiar: I will offer the recordings for a sighted subjective comparison first, to establish some sort of baseline if people think they can hear some differences in the first place. Open discussion and sharing of impressions and hints what details to listen for are welcomed. This of course opens the door for priming (the first published testimonials, and any "reputation" of the person stating it, will subtly influence others in their subsequent sighted findings) but this is fully intentional. I want to recreate the normal end user situation as seen in the more subjectivist forums.
  
• Phase 2, Assertion: Once enough feedback has accumulated that at least some are certain to hear differences, the blind-testing phase proper will start, otherwise the specific test will be over and we have to start from new, using different test tracks, change some details of the setup and whatnot, of course it will be a limited amount of cycles. Preferred format for the blind testing is ABX log made with foobar2000 and associated ABX plugin but we might have other options, too. I'm aware that cheating is possible but lets take that aside for now. ABX'ing requires patience and training, so it would be best everyone gets familiar with it early on, right in Phase 1. Preferably, we will have several rounds of Phase 2, occassionally using hidden positive and negative controls to increase the robustness of any results.
  
• Phase 3, Investigation: If some evidence can be accumulated that in the specific setup used to make the recording there are perceived differences, we have layed the grounds to actually start an in-depth technical investigation to search for the differences in the signals that are deemed responsible for the changes. Otherwise, the test is over for obvious reasons, though it still could be worthwile to check the recordings. The technical investigation is basically simple, just look for differences in the signals, but actually the details can get quite complicated, and of course baseline stuff is required to establish limits of resolution and uncertainity, all the bells and whistles needed for a proper technical analysis. I have some years of experiences in this field, as will other have, so I am really confident that once cable differences are nailed down in blind-testing, one can find the dominant root cause.

Important note: We all know this is a loaded topic, and with the majority of ASR members probably leaning to the objectivist side of things it is clear that many of us will be thinking like „c'mon, it's already been proven in many measurements -- like Amir's -- that proper cables make no technical difference for the transmitted signals whatsoever, so don't waste your/our time on this“. If this is your stance which is completely viable of course, then please do NOT post in this thread (and any associated threads that might follow). I would like to encourage a meaningful discussion on technical and procedural details under the premise that it does actually make sense to try such an investigation as outlined above. Some of you also might participate in the more subjectivist forums and could gather some comments, or even actual participation, from the membership there, maybe acting as a man-in-the-middle to relay discussion and results of those who don't want to register here.

Best, Klaus

I don't understand the underlying hypothesis you're trying to test.

Is it:

"That it is possible, with radical cable construction methodologies, to make a cable with unusual LCR values that will be audible?"

I think physics already tells us the answer to that....

 

[DSJR]

I'm so sorry to be a sour-puss, but I'm damned glad I'm out of all this bullshit long ago now. The cables I '*really* heard a repeatable difference with were all RCA single ended types (balanced connection removes a multitude of sins if done properly I discovered ) and the cables that 'made a difference' totally unsuited to audio use primarily. I do wish you well in this investigation though. Having had my 'wire epiphany' a few years back, I just can't be arsed now, as it's all sighted subjective and price-cachet driven to me now. Again, sincere apologies and of course I'll look in to see what others with better ears than mine can discover

 

[watchnerd]

I'm so sorry to be a sour-puss, but I'm damned glad I'm out of all this bullshit long ago now. The cables I '*really* heard a repeatable difference with were all RCA single ended types (balanced connection removes a multitude of sins if done properly I discovered ) and the cables that 'made a difference' totally unsuited to audio use primarily. I do wish you well in this investigation though. Having had my 'wire epiphany' a few years back, I just can't be arsed now, as it's all sighted subjective and price-cachet driven to me now. Again, sincere apologies and of course I'll look in to see what others with better ears than mine can discover

Ditto.

Here is the issue I have:

Even if one engineers cables to have a particular sound on purpose (yes, of course that's possible with highly usual LCR factors), why the hell would you?

EQing your system via cables is just incredibly inefficient and complicated when you can just....you know....use EQ.

 

[Wps998]

Hmm interesting and awesome effort..Whatever the results tho, i will still buy nice cables. I want to, i can afford them, my house has 3 phase power, my audioroom has dedicated 3 phase line, and buying cables does not make me sacrifice expense to my other audio gear. Buy what you want, what makes you happy and can afford. Its a hobby, there is no right and wrong, smart or stupid. If blingy things can cause bias and make you hear more, so can reading an article or chart can make you settle for what is not satisfaction. Enjoy it to the fullest.

 

[watchnerd]

Its a hobby, there is no right and wrong, smart or stupid. If blingy things can cause bias and make you hear more, so can reading an article or chart can make you settle for what is not satisfaction. Enjoy it to the fullest.

I like buying nice looking things, too.

And, yes, I enjoy listening to gear I think is good looking more than ugly gear.

But I think we already know this and I'm not sure what more research on sighted bias would tell us that we don't already know.

 

[Feelas]

Isn't that just making the same point as the first one and thus the commonly accepted reason why we do ABX tests??

Yes, but that is the problem: we can know scientifically and objectively that two pieces of gear are the same and yet hear different, that's the point. I believe that subjectivists seem to feel that everyone in here completely disregards the second part, yet I'd be worried that nobody on ASR submits to the DBT protocol when actually listening to anything

It should be a primary matter to understand what is actually differently perceived when sighted testing, not merely assess that people are biased. I guess it'd be very entertaining to somehow try to translate the difference in sighted testing into measurable features. Like, "person X percepts 2dB more treble when sighted testing on the same measuring amplifiers".

In other words: why bother doing DBT, if you don't do it on daily listening? That'd be a good engineering feat to assess what is really the better piece of hardware; albeit you might not be able to rid yourself of the bias, and thus hear the worse as better... Somehow. Not that I agree with that approach, but I can see how that one appeals to some.

 

[Objectivist01]

All cables are same. End of it.

 

[KSTR]

Are you familiar with Pano’s potato and banana tests?

Yes I am, and made an Apple test myself. When it is hooked to a high input impedance amp (Tek AM502, DC bias applied through the apple) there was surprisingly little damage to the signal, but shielding the apple was a nightmare. With an 100k load (AP SYS2322 input) the degradation in level and distortion was quite high, though.

 

[KSTR]

You might put the cables against the original file. As the people don't succeed in original x tube preamp ABX, I strongly doubt they will get anything from different cables. You will get some sighted impressions (just impressions) and NO valid positive ABX result.

Original vs. loopback poses a problem for later analysis, the contribution of the DAC and ADC itself. To get a baseline, one would need to make also a shorted loopback test. Then we can just as well compare the shorted loopback vs. the cable loopback to minimize the unknowns and have better chances in the analysis, which will mostly based on subtraction techniques, looking into the residual, trying to remove the trivial linear stuff in the residual first to hopefully expose what really is interesting, some nonlinear effects, stochastic phenomena, whatever.

 

[watchnerd]

Like, "person X percepts 2dB more treble when sighted testing on the same measuring amplifiers".

I think this would be incredibly hard to quantify down to the level of single digit dB values and specific Hz.

I usually hear things like "it sounds brighter" or "it has more detail".

If the device in question is actually flat, I don't know how you reverse engineer that perception into a specific perceived decibel value and frequency.

 

[KSTR]

I don't understand the underlying hypothesis you're trying to test.

Is it:

"That it is possible, with radical cable construction methodologies, to make a cable with unusual LCR values that will be audible?"

I think physics already tells us the answer to that....

No, that's not what I'm after (I think I've already explained that), at least in a first step. Pathological setup effects don't need any further research, sure.

 

[watchnerd]

In other words: why bother doing DBT, if you don't do it on daily listening?

I do it to myself when I have a decision to make about something sighted where I think I maybe hear a difference, but I'm not sure.

Most recent example, posted on ASR, is testing if I really could hear a difference between two different DSP based phono stages I own, or was imagining it.

 

[watchnerd]

No, that's not what I'm after (I think I've already explained that), at least in a first step. Pathological setup effects don't need any further research, sure.

Then I guess I don't understand what unique hypothesis (as opposed to as set of test conditions) you're testing.

 

[KSTR]

In other words: why bother doing DBT, if you don't do it on daily listening?

Seems you missed the point. I think we all agree to the following: According to the whole rationale of this forum, we will want to apply high standards to the technical analysis which will be a lot of work. Therefore, the premise to justify the technical investigation in the first place should follow the same high standards to avoid wasting the efforts. The only way to do this is the blind-test, which, as I've explained in the start post, then gives us some confidence "that there is something" worth digging in deeper.

 

[preload]

Then I guess I don't understand what unique hypothesis (as opposed to as set of test conditions) you're testing.

I'm not clear either.

OP, what is the specific question you're trying to answer? It feels like you've come up with the experimental methodology first.

 

[pma]

Original vs. loopback poses a problem for later analysis, the contribution of the DAC and ADC itself. To get a baseline, one would need to make also a shorted loopback test. Then we can just as well compare the shorted loopback vs. the cable loopback to minimize the unknowns and have better chances in the analysis, which will mostly based on subtraction techniques, looking into the residual, trying to remove the trivial linear stuff in the residual first to hopefully expose what really is interesting, some nonlinear effects, stochastic phenomena, whatever.

This is absolutely evident, however if people do not get original vs. tube (+inevitable DAC/ADC) with almost 3% THD, or signal recorded from power amp terminals loaded by speakers, then it is highly improbable they will get a cable "sound". I know there is DAC/ADC added in the test I speak about and that I repeatedly prepared. Over the years of diyaudio tests and even here I became absolutely sceptical to disclose something like "cable sound" in online abx test, if it is properly prepared with matched level and time aligned. I know you will do it properly. But how can you expect to disclose anything in a properly prepared "cable test"???

You know why I used the test against the original? To shut the mouth to those arguing that the DAC/ADC path used was not good enough.

 

[watchnerd]

I'm not clear either.

OP, what is the specific question you're trying to answer? It feels like you've come up with the experimental methodology first.

Right.

We already know:

--LCR values that are pathological in nature really can influence the sound, measurably so

--Sighted bias can influence what is heard, and the power of suggestion can influence specific perceptions (e.g. if you're told silver cables sound brighter, you're reasonably likely to hear it that way)

These are already well-established.