Head-Fi.org › Forums › Equipment Forums › Sound Science › Help with a philosophical problem with blind testing cables
New Posts  All Forums:Forum Nav:

Help with a philosophical problem with blind testing cables

post #1 of 51
Thread Starter 
Some people in this forum have been trying to come up with a protocol for blind testing cables in a relaxed setting. Rather than get cranky about it, I've been thinking about it. Unfortunately I've hit a bit of a logical dead end. Please follow my logic to see if you can reconcile this problem for me.

First, a cable is designed to be a passive device. It is, by design, not supposed to change the sound. Yet, people are hypothesizing that is not simply a passive device, but it is somehow changing the signal it is carrying.

So, in short, this is stating a passive device is actually an active device. But if a cable can be an active device, why not other passive devices like headphone jacks, headphone plugs, connectors, wire in amps, dacs, sources, the cases of equipment, solder, etc?

You might think this is ridiculous, but people have claimed these things make a difference. Certainly people buy expensive headphone jacks, headphone plugs, special RCA jacks, special solder, in the hope of not degrading their electrical signal flowing through their headphones.

So here is where I'm stuck. If you claim a negative result from a cable blind test, how can you attribute it to cables since there are so many other passive devices that can be changing the sound? How can you say it's not the RCA plugs on the amp that are contaminating your results? How can you ever claim a negative result is valid if you cannot demonstrate all passive devices in the test are indeed passive?

So what about a positive result? Again, you've got the possibility of contamination from other passive devices. How do you know it wasn't just the cables, but some interaction between the cables and some other passive device? Or maybe it was the passive device itself and not the cables. The solution to this is to increase the number of passive devices the cables are tested with and show that the positive result is consistent across multiple passive equipment. I've been thinking about how to address this statistically, but run into other philosophical blocks.

My point is this, if you state that a cable (a passive device) can have a sonic difference from another cable, then you must (or risk being labeled a hypocrite) also accept there is the possibility that other passive devices can lead to sonic changes. But, if you follow this rationale, since any audio system consists of multiple passive devices, it becomes very difficult to point out that one passive device (e.g. a cable) is responsible for a certain sonic difference.

My only solution (at this point) is to approach this problem in a completely different way. Active devices are supposed to change the electrical signal, so blind test active devices.

If you have a negative result with active results, then you may (you don't have to) conclude that since you have a negative result from active devices, you will also have a negative result from passive devices since by definition passive devices are supposed to not change the signal and active devices do change the signal. Thus, if you cannot hear the difference between devices that do change a signal, it is not likely you will hear the differences between devices that are not supposed to change the signal.

If you have a positive result, then you can say "I've found a positive result at this (active devices) level so let me now go to the next (passive devices) level."
post #2 of 51
Quote:
Originally Posted by odigg View Post
My point is this, if you state that a cable (a passive device) can have a sonic difference from another cable, then you must (or risk being labeled a hypocrite) also accept there is the possibility that other passive devices can lead to sonic changes. But, if you follow this rationale, since any audio system consists of multiple passive devices, it becomes very difficult to point out that one passive device (e.g. a cable) is responsible for a certain sonic difference.
You lost me at this point. If the systems sounds a certain way, and you only change one of your "passive" devices (let's say it's the cable), and the system sounds different (for the sake of argument, let's say it sounds more sibilant), why is that sound change not attributable to the cable change?

Quote:
Originally Posted by odigg View Post

If you have a negative result with active results, then you may (you don't have to) conclude that since you have a negative result from active devices, you will also have a negative result from passive devices since by definition passive devices are supposed to not change the signal and active devices do change the signal. Thus, if you cannot hear the difference between devices that do change a signal, it is not likely you will hear the differences between devices that are not supposed to change the signal.
It seems to me this is a non sequitur. The fact that two CD players sound the same doesn't mean two cables sound the same. There's two many other variables at play. (I do agree that differences between CD players are generally more pronounced than cable differences, but I'm just focusing on the logic of your proposition.)
post #3 of 51
Lets assume three different types of passive devices, each with a rolloff at a certain frequecy as follows:
device A: -X dB
B: -Y dB
C: -Z dB

However, these devices have a lenth, so the proper units for that rollof is actually (dB rollof)/(meter of device).

Now, in an audio system, we have maybe 1 total inch of device A (2.54 cm = .0254 meters), and lets assume 2 total meters of B, and .5 meters of C.

So rollof is added to the system for each device as:
A: -X(dB/m)(.0254 m)
B: -Y(Db/M)(2 m)
C: -Z(dB/m)(.5m)

As you can see, so long as Z is not absurd (say 10^6 dB/m), it will have very little effect. We can see that perhaps if we put wood in place of device A, knowing wood has a rolloff of about 9.99% at all frequencies, all we need is to try to patch one tiny connection with all wood and we have problems.

Liken this to water pipes. You have a system with three pipes in series, all the same diameter, one of material A, another of B, and a third of C. You have 10 meters of pipe A, 1000 meters of pipe B, and .01 meters of C. We assume that the current of water dissolves a little of the surface material of each pipe, and lets assume it does this to each material at the same rate. In the end, you would expect much more of material B in your water, and a neglibible amount of C, right?

Note that the above explanation is one I made up, and it assumes that you believe in the whole cables thing. I do not (am willing to be proven wrong), and thus just put forth what sounded like a reasonable and somewhat (definately not totally) scientific argument, based on my education as a physicist and chemist.

-Nkk
post #4 of 51
Thread Starter 
Quote:
Originally Posted by PhilS View Post
You lost me at this point. If the systems sounds a certain way, and you only change one of your "passive" devices (let's say it's the cable), and the system sounds different (for the sake of argument, let's say it sounds more sibilant), why is that sound change not attributable to the cable change?
If we are talking about a cable (we usually talk about cables), a cable consists of multiple passive components. You have the plugs (minimum 2) at the ends of the cable, the solder, and even the sheath if that offers any shielding properties.

So how can you say a change (or no change) is strictly from the wire part of the cable? It seems to me that the connection between the plugs on the cable and the plugs on the active device (amp, dac, etc) is a likely area for some interaction considering you have mated two separate devices together.

So, to demonstrate the wire part of a cable is responsible for the change, you would have to try that wire with different connectors, solder, etc. Only if you could demonstrate the change held with difference passive parts, could you say the wire is responsible for the difference.

To me this is important because parts like jacks (RCA, stereo 3.5mm) are probably made from different metals and compositions based on supplier cost, availability, etc. There is probably also some batch to batch variation in composition. So we have no way to state that the connector you got with the cable you are testing (let's say X brand Cable) has the same connector as the X brand cable somebody purchased a year later even the connectors look the same.

So it's important to demonstrate it's the wire itself that is responsible for the sonic change, not the other parts of the cable.

Also, this assumes a fixed interaction between two passive components. What if this interaction itself (such as between female and male connectors) was complex and active (e.g. wire reacts differently to different connectors)? So you couldn't just switch cables with on one amp and generalize a positive or negative result across different equipment. You'd have to demonstrate this with different equipment to show the interaction across multiple components is consistent

Quote:
It seems to me this is a non sequitur. The fact that two CD players sound the same doesn't mean two cables sound the same. There's two many other variables at play. (I do agree that differences between CD players are generally more pronounced than cable differences, but I'm just focusing on the logic of your proposition.)
I'll elaborate my point a little more.

It's easy to show a *measurable* differences between active equipment. Using RMAA, you can do it even on a basic desktop PC with a stereo in.

From what I've seen from cable measurements (nick_charles did some recently), finding reliable differences across cables, or even across trials, is much more difficult.

So my logic is, if you (not everybody, just the person who is doing the test) cannot hear a difference between two active devices that reliably measure differently, how will you hear a difference between two devices that nobody has (to my knowledge) demonstrated measure differently (at least in the aubible range).

Maybe we can agree to disagree?

nkk - I like your water analogy. I think the problem is that since differences in passive devices have not been measured reliably, it becomes problematic to point at one (when multiple devices are in the chain together) as the culprit in a sonic change. With your water analogy you can measure the content of each pipe in the water to calculate the rate of dissolve. With passive devices nobody (as far as I know) has done this.

Another problem is, your water analogy assumes a static rate of dissolve. What if the pipes dissolved at vastly different rates based on the water that went through it? It would then be harder to pinpoint how much each pipes dissolves in water.

Hence my problem with the mental block when thinking about how to go about a blind testing cables. It seems very complex.
post #5 of 51
Quote:
Originally Posted by PhilS View Post
You lost me at this point. If the systems sounds a certain way, and you only change one of your "passive" devices (let's say it's the cable), and the system sounds different (for the sake of argument, let's say it sounds more sibilant), why is that sound change not attributable to the cable change?
Well, could you measure a difference in the frequency response? Was it accompanied by excessive high frequency ringing? You might have a really badly designed amplifier that has skimped on feedback and is prone to oscillation

The difference then would be pretty easy to hear and measure between cables - but instead of changing cables, you should change the amplifier and ask for money back ...
post #6 of 51
Thread Starter 
Quote:
Originally Posted by sanderx View Post
Well, could you measure a difference in the frequency response? Was it accompanied by excessive high frequency ringing? You might have a really badly designed amplifier that has skimped on feedback and is prone to oscillation

The difference then would be pretty easy to hear and measure between cables - but instead of changing cables, you should change the amplifier and ask for money back ...
I think it's largely been established that nobody has shown (at least as far as I know) audible measured differences between cables, at least when talking about measuring with test equipment.

But the claim here is that audio cables can make a difference, and they are making a difference in a way that is not being measured by standard (or known) equipment measurements.

Hence the need for a blind test that addresses some of the issues with typical ABX protocols (short trials, out of comfort zone, etc).

But, as you've seen with what I've stated already, the exact rational for this blind test has me running around in circles. I'm wondering if there is a way out of this conundrum.
post #7 of 51
Not only do the "passive" parts of a connection make a difference, sometimes they are the most important part of the difference. A cable can actually have its impediance, capacitance, RFI/EMI shielding, etc. measured. And there are certainly ways to maximize or minimize these parameters based on certain geometries and such.

But RCA jacks and plugs and other "passive" devices generally go unmeasured. The metals they are made from (brass, copper, rhodium, silver, nickle, etc.) all effect the sound in different ways. Even more important, if a jack is copper and the plug is rhodium, or some other different combination, that will be different than if they are of the same metal.

Then there is break-in. Some will say that is a fallacy, but I have heard it enough to say it is a fact. Not only do cables break in, but if you change one out it will also take time to "settle", so quick A-B comparisons by plugging and unplugging components have fatal flaw in my opinion. Not as drastic as between a new vs a broken in cable, but still there somewhat.

Then you also need a system that is resolving enough to let you hear the difference. I think this is where the naysayers screw up the most. Since they don't believe in differences they don't believe in better (i.e. usually more expensive) systems to run their test on. In which case they are probably right, it won't make much if any difference.

You also need music that you are intimately familiar with. using someone elses test track will take repeated listenings for you to get all the nuances. An essential pre-knowledge component of the test. So group tests are usually flawed in that way.

The best test is to try something for yourself and come to your own conclusions in your own rig, over time and within your own budgetary constraints. Nothing like experimentation and gathering real self-explored knowledge. Quick A-B comparisons only let you focus on one or two aspects of the sound and "wow" factors tend to take over. Listening over time evens this out and allows you to explore different aspects with different music in a relaxed and known environment without pressure.

It's like a golfer with 10 swing thoughts. Way too much to process and get right. It's better to focue on one at a time, or maybe two. Otherwise the overload will cause a misfire and the results to be thrown out, and lead to fatigue. Same with quick A-B testing.

Enjoy,
Bob
post #8 of 51
Two logical problems with your theory:

1. Assumptions on what passive and active devices are. (Neither defined clearly nor proven what they are.)
2. Assumption that only active devices should/can affect sound.

If you're saying that only devices that have power applied (IC, opamp, tubes, etc.) are active, therefore only active devices change how something sounds would be a blown theory when it comes to something like a speaker or a crossover. Speakers are, by your loose definition, passive and shouldn't affect the sound produced.

It is already proven scientifically that different metals and alloys have varying degrees of electrical resistance and have different conductive characteristics, and have limits to their high-end frequency response (although normally in MHz, not 20KHz). This also defeats the idea of passive devices not having an effect on the sound.


Starting with a bad set of assumptions can't lead to anything usable. Keep going with your pursuits, but make the appropriate corrections as you go.
post #9 of 51
Yup, it is a philosophical problem.

Underdetermination is a potent weapon in the skeptic's arsenal.
post #10 of 51
Stong Undertermination = tells us that there is no way to distinguish between theories with the same observable consequences – called empirical equivalence – and points to the existence of an infinity of possible theories consistent with any finite data set. It relies on an implicit separation of theory and observation: when we say that the evidence underdetermines the theory choice, we run up against theory-ladenness. Since we cannot distinguish between theory and observation in a straightforward fashion, we cannot appeal to or rely on theory-neutral observations and say that these disallow the possibility of making a choice.

Weak Undertermination = it is always possible to construct alternative theories which are empirically equivalent and also share many of the characteristics we desire in scientific theories.

To summarise, underdetermination is almost an acceptance that we are creative in our explanations and can typically find more than one for a given puzzle. It speaks against a naïve form of empiricism and is only a problem for those who suppose that there is nothing more to science and scientific theories than an appeal to data.


Nice link. I posted what I believe are the relative explainations above. Maybe it will open some eyes to people's over aggressive opinionating. It's amazing how polarized this argument about cables and testing can become, when all it really takes is some first hand experimentation. Then simply decide for yourself.

Bob
post #11 of 51
Quote:
Originally Posted by ServinginEcuador View Post
It is already proven scientifically that different metals and alloys have varying degrees of electrical resistance and have different conductive characteristics, and have limits to their high-end frequency response (although normally in MHz, not 20KHz). This also defeats the idea of passive devices not having an effect on the sound.
In absolute terms you can find measurable differences between cables, I have done this myself, however when the differences are at such low magnitudes i.e 100ths or 1000ths of a db at any given frequency , which is typically what I found[1], then you are asking a great deal of human perception to detect these differences.

Any way you cut it unless you deliberately set out to mangle the sound using capacitors, resistors and so forth two cables of the same length and approximate same gauge at any price will be extremely similar in their frequency response and that response will be flat as a pancake at audible frequencies and why would you want anything other than a flat frequency response ?


1 - Solid copper, stranded copper, silver plated copper, stranded
silver
post #12 of 51
Thread Starter 
Quote:
Originally Posted by ServinginEcuador View Post
Two logical problems with your theory:

Starting with a bad set of assumptions can't lead to anything usable. Keep going with your pursuits, but make the appropriate corrections as you go.
I think that's exactly the problem. I'm trying to eliminate assumptions but I'm going around in circles. So far all I am able to do is define active as "supposed to change the signal" and passive as "not supposed to change the signal."

Note that I say "supposed" not "does."

An amp is supposed to change the signal as it takes a small signal and makes it bigger (more voltage, current, take your pick). A cable is not supposed to change the signal. A crossover is designed to change the signal, therefore it is active by this definition.

Quote:
Originally Posted by BobMcN View Post
when all it really takes is some first hand experimentation. Then simply decide for yourself.
See, that's my problem. Based on experimentation, I've become skeptical of sighted listening tests. So I'm trying a blind listening test and trying to experimentally control my variables. The goal is not a quick A/B test, but a long term test in more typical listener conditions.

But there seem to be so many variables and I'm going around in circles.
post #13 of 51
Maybe you're trying to quantify it too much. Perhaps you just need to sit back and listen and see which you enjoy more.

Now I'm not saying that you shouldn't also quantify some parts of the test, just that there's probably an unquantifiable aspect of all of this also.

As I see it, and per my experience, there are 2 ways to listen: critically and non-critically. When I make a mod to a component, or try out something new, be it tweak or cable or whatever I go into critical listening mode. I hear the twinkle, the boom, the souondstage, the details, the imaging, the naturalness, the timbre, etc. This is not a very enjoyable way to listen.

Then I sit back the next day and relax. Read a book with the music playing. Does it draw me in? How do I feel about it now. Is it making me happy or pulling me into a critical mode again. All of this contributes to your overall enjoyment so it is an important, if not quantifiable aspect of the sound.

Yes there are a lot of variables, so only try and change one thing at a time. And remember that perfection is not achievable, only enjoyment.

So Enjoy,
Bob
post #14 of 51
Odigg, I am going to share my opinions towards what you wrote here.

If I am not wrong the main problem you are having with cable's test is that you think that not only the wire inside the cable can be responsible for a SQ change, but the whole of what a cable is (from the wire to the connectors). I believe you are entering in a world of relativism, a place where absolutely everything might be the reason for some people listening to some differences, hence doubting DBTs.

I think the solution here is to get back to what we as humans can hear, to the limits we have, avoiding getting into relativism, but taking in account that some people have better hearing capabilities than other people.

Assuming that everyone can hear from 20Hz to 20KHz, it should not matter what materials are used in a passive device if the measured frequency response stays unchanged between those audible frequencies.

If by doing a blind test, and trying to control more variables than in a sighted test, without matching volume output, and with something costing more money than other thing, the differences disappear, then we have concluded that either there are no differences, or our method is incomplete. Both might be true, and the problem here is the latter one, to what degree? Well, as we are human beings, we have a limited capability that has been surpassed by machines, we are better of relying on data measured by some machines than our own opinions.

Relativism is also making you wonder if a passive device can exist, or if they are all active devices to a certain degree. It might be the latter, but on something that goes above our human thresholds of listening. If they are above that, then there is no problem with that, because getting back to the relativism, those devices will be passive for us.

I expect you to tell me if I got any of your ideas wrong, and for you to clarify if what i have answered is getting away from your conundrum or not.
I also feel I have left some sentences unfinished, but I am writing this wilst listening to music, so it becomes a bit more difficult than usual to concentrate on what I am saying.
post #15 of 51
Quote:
Originally Posted by BobMcN View Post
Not only do the "passive" parts of a connection make a difference, sometimes they are the most important part of the difference. A cable can actually have its impediance, capacitance, RFI/EMI shielding, etc. measured. And there are certainly ways to maximize or minimize these parameters based on certain geometries and such.
While this is true on the face of it, if you actually do the measurements, you will find that the differences are far below what is likely to be audible by humans ... and worse, considerably below the noise of available music sources and recordings. Claiming to hear differences in noise that is 20-30db below the theoretical noise floor of what dvd-audio can deliver requires extraordinary proof.

Quote:
Originally Posted by BobMcN View Post
But RCA jacks and plugs and other "passive" devices generally go unmeasured. The metals they are made from (brass, copper, rhodium, silver, nickle, etc.) all effect the sound in different ways. Even more important, if a jack is copper and the plug is rhodium, or some other different combination, that will be different than if they are of the same metal.
It is wrong to state that these will affect the sound. There is no basis for such claims.

Quote:
Originally Posted by BobMcN View Post
Then there is break-in. Some will say that is a fallacy, but I have heard it enough to say it is a fact. Not only do cables break in, but if you change one out it will also take time to "settle", so quick A-B comparisons by plugging and unplugging components have fatal flaw in my opinion. Not as drastic as between a new vs a broken in cable, but still there somewhat.
Again, there is no proof. Worse, there is a large body of counter-evidence.

Quote:
Originally Posted by BobMcN View Post
Then you also need a system that is resolving enough to let you hear the difference. I think this is where the naysayers screw up the most. Since they don't believe in differences they don't believe in better (i.e. usually more expensive) systems to run their test on. In which case they are probably right, it won't make much if any difference.
Cable non-believers seem to be the ones in possession of most of the highly precice audio gear. Or do you have some other kind of revealing in mind?
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Sound Science
Head-Fi.org › Forums › Equipment Forums › Sound Science › Help with a philosophical problem with blind testing cables