Why pick on cables ?

[Clarkmc2]

Electrical measurements are great in that they have excellent repeatability and reproducibility. They however, are a surrogate for human listening and perceptions. A surrogate where no in-silico to in-vivo correlation to listening to my knowledge has been reliablely established. Try getting a statin today approved with simply cholesterol measurements. You need to demonstrate efficacy and safety with hard endpoints (reduction in heart attacks etc.).

Are you saying that if there is absolutely no change in the electrical signal that is the sole producer of a phenomenon, it can still actually be changed by virtue of someone listening to it? Statistical methods cannot study or verify a physical phenomenon that cannot possibly exist in the physical world. Again, this is not about what someone thinks they hear. Discussion of audio equipment is not about hearing, it is about changes in the behavior of the equipment. No change, no difference to hear. If there is a change, then feel free to determine its audibility. This is not related to the interactions of substances with complex biological entities. This stuff is relatively easy to measure.

Outside the domains of quantum mechanics, some things ARE cut and dried. I imagine misapplied quantum theory will be the last refuge of snake oil purveyors.

 

[iim7V7IM7]

Are you saying that if there is absolutely no change in the electrical signal that is the sole producer of a phenomenon, it can still actually be changed by virtue of someone listening to it? Statistical methods cannot study or verify a physical phenomenon that cannot possibly exist in the physical world. Again, this is not about what someone thinks they hear. Discussion of audio equipment is not about hearing, it is about changes in the behavior of the equipment. No change, no difference to hear. If there is a change, then feel free to determine its audibility. This is not related to the interactions of substances with complex biological entities. This stuff is relatively easy to measure.

No, I am not saying that. I am saying that conductivity, resistance, impedance, dielectric etc. are independent measurements and there is no established correlation to human psycho-acoustic perceptions. What electrical property correlates to the perception of instrument placement within the soundstage that we perceive? Human senses cannot be fully characterized by these surrogate measures alone.

 

[Clarkmc2]

No, I am not saying that. I am saying that conductivity, resistance, impedance, dielectric etc. are independent measurements and there is no established correlation to human psycho-acoustic perceptions. What electrical property correlates to the perception of instrument placement within the soundstage that we perceive? Human senses cannot be fully characterized by these surrogate measures alone.

Oh boy, an "X-factor" defense.

The measurements I am talking about are the output of the equipment while playing music or test signals. Either null tests or comparison of transducer output as percieved by calibrated microphones. Not electrical characteristics of equipment.

Soundstage is a minefield of psychoacoustic entanglement. A listener can perceive, modify or ignore soundstage. He can also create it from a mono presentation. Not meaning quantum entanglement, the only thing more baffling than audiophiles discussing soundstage.

 

[bigshot]

Quote:

My thought is that those who are relying on small A/B studies that demonstrate a ~ 50/50 chance of selecting one cable from another as evidence that there is no difference are not much different than the snake oil salesmen. Both may (key word here) have good intentions, but neither is based on sound science.

 
How about just measuring the minute differences and then comparing the figures to established thresholds of audibility? We know pretty well how much frequency response has to deviate to be audible, and we know how much distortion can be heard. It seems to me, even without listening tests, it should be pretty obvious whether cables have a sound or not. (Unless someone just doesn't want to know and no answer is going to be definitive enough for them.)

 

[mikeaj]

So is the reservation about which measures and tests are being used (e.g. FR, nonlinear distortion) or the established levels of audibility (because of experimenter incompetence, sample size, playback system, or other aspect of previous tests)?
 
Not only are you talking about levels far below suggested levels of audibility, but levels that are far below such effects as positioning of your head in the room (speakers) or headphones on the head, L/R imbalances of transducer sets, expectation bias, what part of the sound somebody is focusing on, etc.
 
If it were just about a few A/B studies of 50/50 results, that would be one thing certainly, but nothing else stacks up either... Actually, even if it can be reliably demonstrated to make a difference, it would certainly still be very low in priority on the list of things to get right in an audio playback system.

 

[bigshot]

even if it can be reliably demonstrated to make a difference, it would certainly still be very low in priority on the list of things to get right in an audio playback system.

That's the one bit of logic that escapes a lot of people. Minute differences just aren't important. If you have to jump through hoops to just register the difference at all, perhaps there are other things that are more important.

I remember the old saw that being an audiophile is worrying about the last 1%. But the problem is, they usually worry about that long before they even address the 99%.

 

[iim7V7IM7]

Oh boy, an "X-factor" defense.
The measurements I am talking about are the output of the equipment while playing music or test signals. Either null tests or comparison of transducer output as percieved by calibrated microphones. Not electrical characteristics of equipment.
Soundstage is a minefield of psychoacoustic entanglement. A listener can perceive, modify or ignore soundstage. He can also create it from a mono presentation. Not meaning quantum entanglement, the only thing more baffling than audiophiles discussing soundstage.

I must be wrong and you must be right.

 

[Clarkmc2]

Quote:

I must be wrong and you must be right.

Not at all. We were talking about different things. You're a smart guy!

 

[Sandman65]

There have been some excellent and patient posts delivered in this thread!
 
I do believe people can hear differences but I believe these differences (aside from poor termination, inline circuitry, etc...) are more likey created by stimulus completely seperate from some newly discovered "phenomena" that redefines established electrophysical theories and laws.
What I believe is it is more likely that changes in hearing from one listening session to another (and recollection of such when trying to ascribe differences) that creates the "difference"
The cable people tell you its the cables and others say its the nature of human hearing...I could mention placebo but I won't.
 
The funny thing about recollection is it is the argument often forwarded by those against blind testing, the argument being the individuals perception of difference varies to such a degree that, even after relatively short periods of time, it becomes impossible for human hearing to remain "constant" while switching back and forth. Therefore if human hearing is so variable that blind tests are inconclusive the arguement over efficacy of cables seems to me a bait-and-switch where you are being told it's the cables whereas it's more likely human hearing is going to tell you theres a difference regardless if you are listening to the same cable or trying to compare between various flavors?
 
When you do a industry survey of cable proponents you'll often find they have direct or indirect financial interest in your purchasing decisions.
It's a sales/marketing/advertisers job to seperate you from your money and unfortunately there are many unscrupulous types out there that will play on your passion for music..
I tend to place more validity in arguments posed by people who have no finiancial stake in the outcome.
 
It's like the buddy who talks you out of taking the nasty she-beast home at closing time...you're mad at him at first but once you sober up (and remain STD free) you're viewpoint will change.

 

[iim7V7IM7]

How about just measuring the minute differences and then comparing the figures to established thresholds of audibility? We know pretty well how much frequency response has to deviate to be audible, and we know how much distortion can be heard. It seems to me, even without listening tests, it should be pretty obvious whether cables have a sound or not. (Unless someone just doesn't want to know and no answer is going to be definitive enough for them.)

You could try to conduct a design of experiments, a full factorial or partial factorial design (if you felt you already knew which inputs influence human listening outputs). You would need to block the source material and listening sequence to evenly spread listening fatigue and or bias. You could try to analyze the data with an analysis of variance (or ANOVA) to see which measurable electronic measures most influenced the specific human listening output response. I think the challenge would be in sample size and proper categorization of the human hearing output given the variability. You would need a very large sample size. You would lastly need to agree on what a "significant difference" was and how confident you want to be. As differences getter smaller and confidence levels get higher, sample size (n), gets larger.

 

[iim7V7IM7]

Not at all. We were talking about different things. You're a smart guy!

Sorry, I misunderstood you. thanks for your kind words.

 

[Chris J]

Quote:

its been pointed out that the 1st post had some basic rhetoric errors - a false equality of comparisons between links in the audio reproduction chain re "audibility"
 
headphone all vary by more than DBT/ABX thresholds - we know they "sound different" - measurement of output sound frequency response alone is sufficient to establish that differences are audible
 
wire/cable electrical interconnect electrical characteristics can be calculated, measured - for most situations in headphone audio reproduction their interaction with normal equipment does not give changes in frequency response that would be expected to be audible  - often by margins of 100x
 
some don't seem to get the basic idea - you have to have numbers, sensitivities, knowledge of thresholds - just because every physical assembly has (possibly microscopic) different properties and some differences in some properties can lead to changes big enough to be heard doesn't mean any difference whatsoever can be heard
 
 
tubes made at the same factory can differ in critical electrical characteristics by >10% - vendors measure, charge extra for matching tube's gm to even 10%, even dual triode tubes - both sections in the same envelope can have Vt difference up to 2 V
 
in many common tube circuits, "normal" new production parameters range for the same model tube, and service life drift, wearout differences can have effects on amplifier response that are clearly above DBT/ABX audibility thresholds

 
Transistors also have wide variances in their parameters: on the order of approx.+/- 50 %, depending on the parameter. Do all Op Amps sound different? Do various samples of discrete amplifiers sound different?
A properly designed SS amp sould have enough negative feedback to swamp out component variations.
 
Depending on the type of capacitor, capacitors can also have wide variations in parameters and still be in spec.
 
So what are tube rollers hearing?
Differences between brands? i.e. Sylvania 12AX7 sounds like this, Mullard 12AX7 sounds like that, GE 12AX7 sounds all warm and smoochy?
Or, more likely, sample to sample variation? In spec tubes with out of spec tubes?
Which will be more pronounced in low, or even zero loop feedback amps, and less pronounced, to non exsistent in high feedback designs.
Is this what capacitor rollers are hearing?
 
As I mentioned earlier:
A fair comparison would be comparing in spec tubes to in spec tubes.
Not comparing old worn out, out of spec tubes with unused, in spec tubes.

 

[mikeaj]

Quote:

You could try to conduct a design of experiments, a full factorial or partial factorial design (if you felt you already knew which inputs influence human listening outputs). You would need to block the source material and listening sequence to evenly spread listening fatigue and or bias. You could try to analyze the data with an analysis of variance (or ANOVA) to see which measurable electronic measures most influenced the specific human listening output response. I think the challenge would be in sample size and proper categorization of the human hearing output given the variability. You would need a very large sample size. You would lastly need to agree on what a "significant difference" was and how confident you want to be. As differences getter smaller and confidence levels get higher, sample size (n), gets larger.

 
Factorial design on what factors?
 
If you're going by things like resistance, inductance, capacitance, etc., there will still be complaints about the length of the cable used, some other factors you don't test for, the mechanism by which you're varying the factors (do you need to build a different cable with each of the desired properties?), and other problems.  Also, the effect of cable parameters depends on the amplifier and load.  Test different combinations of those too?  I mean, do you really want to map out an entire design just so, for example, you can eventually get up to some combination where there's a long-enough cable with enough capacitance and an unstable enough amplifier to cause problems?
 
It's a pretty big mess.  Now, if you can run trials for a few million participants...
 
Or would you try to test directly for frequency response variations by level, nonlinear distortion (but what type? how much of each harmonic?) at different levels, etc.?

 

[iim7V7IM7]

Factorial design on what factors?

If you're going by things like resistance, inductance, capacitance, etc., there will still be complaints about the length of the cable used, some other factors you don't test for, the mechanism by which you're varying the factors (do you need to build a different cable with each of the desired properties?), and other problems.  Also, the effect of cable parameters depends on the amplifier and load.  Test different combinations of those too?  I mean, do you really want to map out an entire design just so, for example, you can eventually get up to some combination where there's a long-enough cable with enough capacitance and an unstable enough amplifier to cause problems?

It's a pretty big mess.  Now, if you can run trials for a few million participants...

Or would you try to test directly for frequency response variations by level, nonlinear distortion (but what type? how much of each harmonic?) at different levels, etc.?

To your point, you would need to standardize these. Or on the other hand, perhaps cable length in itself is a factor to be studied. In any case, a large confounded mess, with messy endpoints. I was more responding in theory. I tend to agree that there is a low probability of success.

 

[Chris J]

I've been following your discussions WRT cables with interest, the question is still out there, with respect:
 

Why Pick On cables?​

 ​

And no, I am not trolling.