Cabling, testing, and sound quality

[mojidooji]

I posted a question in another thread asking whether or not this testing procedure was possible and got no answers, so I'm posting this to get some feedback on this idea before looking into it further.

There is a part of the argument of why all “good” cables sound the same that has bothered me; because inductance, capacitance, and resistance are the only properties that effect sound. I say this because it is true if, and only if, all good cables sound the same. People supporting this argument often attempt to prove that cables indeed sound the same by using the least reliable instrument to measure sound quality, besides the ones which cannot detect sound – the brain. Besides the problems of bias and memory, we also must contend with changing metrics, attention span, and the uncertainty of subjective measurement. I am not certain that most people are even capable of recognizing anything but the most blatant differences between two sounds. My point is that the psychological issues probably far outweigh any other single factor, most of the time, so I question the validity of every test method because it cannot control for the most important variable. Without this, no test can be conclusive.

The only way to fix this is to remove the subjective component of the testing process which, I think, means we should redefine “sound quality” before coming up with a more reliable test procedure. Since “sound quality” is purely qualitative, by definition, it is not to be confused with accuracy, which is a measurable quantity consisting of frequency and amplitude over time. I’ve long thought the anti-cable-difference camp was trying to go the wrong way down a one way street, so I think we should dispense with the assumption that resistance, inductance, and capacitance (in a “good” cable) were the only factors we should care about. I don’t think any reasonable person could debate the issue of differences when they are measured in the only relevant units; frequency and amplitude over time.

We can let philosophers debate whether one should purchase more good than can be heard. So, is this possible?

 

[mtrycrafts]

I posted a question in another thread asking whether or not this testing procedure was possible and got no answers, so I'm posting this to get some feedback on this idea before looking into it further.

There is a part of the argument of why all “good” cables sound the same that has bothered me; because inductance, capacitance, and resistance are the only properties that effect sound. I say this because it is true if, and only if, all good cables sound the same. People supporting this argument often attempt to prove that cables indeed sound the same by using the least reliable instrument to measure sound quality, besides the ones which cannot detect sound – the brain. Besides the problems of bias and memory, we also must contend with changing metrics, attention span, and the uncertainty of subjective measurement. I am not certain that most people are even capable of recognizing anything but the most blatant differences between two sounds. My point is that the psychological issues probably far outweigh any other single factor, most of the time, so I question the validity of every test method because it cannot control for the most important variable. Without this, no test can be conclusive.

The only way to fix this is to remove the subjective component of the testing process which, I think, means we should redefine “sound quality” before coming up with a more reliable test procedure. Since “sound quality” is purely qualitative, by definition, it is not to be confused with accuracy, which is a measurable quantity consisting of frequency and amplitude over time. I’ve long thought the anti-cable-difference camp was trying to go the wrong way down a one way street, so I think we should dispense with the assumption that resistance, inductance, and capacitance (in a “good” cable) were the only factors we should care about. I don’t think any reasonable person could debate the issue of differences when they are measured in the only relevant units; frequency and amplitude over time.

We can let philosophers debate whether one should purchase more good than can be heard. So, is this possible?

Well, perhaps you may need to do some more research into how psychoacousticians conduct their research, how sound quality is judged by research places like Harman International or the National Research Center in Canada that has researched in the filed for many decades, etc.

One cannot remove a subjective component that is in deed vital in this hobby but, one can control bias, shortness of memory for small differences, etc. You just cannot ignore the decades of replicable research that has gone before us in acoustics.

And, cable parameters, those 3 that you mention is it. Large enough variations in the right parameter for the particular cable use is in fact audibly different. Small differences are just not perceptable by the human brain, sorry.

 

[mojidooji]

mtrycrafts,

Perhaps I wasn't clear, or perhaps you have read too much into my post. I realize that this is an emotionally charged issue for many people and I am attempting to approach the topic in a non-threatening way to get the best possible feedback.

My interest is not psychoacoustics; it is how to measure the accuracy of an audio system. I'm not sure what the study of how the brain processes sound has anything to do with the accurate measurement and comparison of audio signals, but research on the topic does not indicate that the brain makes objective measurements. People like what they like for all sorts of reasons and, while it is an interesting field in psychology, I don't think it addresses quantitative analysis.

While I appreciate your input, I think you have wrongly judged my motivations on this topic, aside from dodging the question. It should have been obvious from my post that I have an academic interest in specific sorts of measurements and analysis of those measurements to test a specific theory. If you are open to sharing your expertise, please reconsider my original post and answer accordingly.

 

[avnetguy]

Maybe it would help if you could elaborate on the specific theory you're testing for?

 

[mojidooji]

avnetguy,

I'm not sure if that would help, since I am looking for methods to compare the frequency and amplitude over time of multiple audio signals, which seems to be a quite specific query.

But, to answer your question, I would like to test the theory that properly working cables do not have audible differences, in real-world conditions.

 

[markw]

A well done double blin test seems to meet your needs but, for some reason, many people object to this.

 

[Nestor]

Define "frequency and amplitude over time".

 

[Rickster71]

I realize that this is an emotionally charged issue...

It should have been obvious from my post that I have an academic interest in specific sorts of measurements and analysis of those measurements to test a specific theory.

While this is one of the more objective on-line forums, and with that being said; I don't know that with it's 51,000 members (mostly laymen) it's may not be the best place for purely academic pursuits.

Unless you're looking to study the human condition; I don't know that a public forum is the best place for 'non-emotional' academic interests.

 

[avnetguy]

But, to answer your question, I would like to test the theory that properly working cables do not have audible differences, in real-world conditions.

If you're looking for an "audible difference" then the human component would be included in the testing. It's one thing to electronically measure a difference and something else to say we can actually hear the difference.

 

[TLS Guy]

mtrycrafts,

Perhaps I wasn't clear, or perhaps you have read too much into my post. I realize that this is an emotionally charged issue for many people and I am attempting to approach the topic in a non-threatening way to get the best possible feedback.

My interest is not psychoacoustics; it is how to measure the accuracy of an audio system. I'm not sure what the study of how the brain processes sound has anything to do with the accurate measurement and comparison of audio signals, but research on the topic does not indicate that the brain makes objective measurements. People like what they like for all sorts of reasons and, while it is an interesting field in psychology, I don't think it addresses quantitative analysis.

While I appreciate your input, I think you have wrongly judged my motivations on this topic, aside from dodging the question. It should have been obvious from my post that I have an academic interest in specific sorts of measurements and analysis of those measurements to test a specific theory. If you are open to sharing your expertise, please reconsider my original post and answer accordingly.

Don't get fancy, its simple. Wire is wire and the small differences in cables inductance and reactance are insignificant, compared to other errors in the system.

As far as speaker wires are concerned resistance is the only factor of importance. For turntables capacitance is an issue, and easily corrected for. For FM and digital signals, the cable impedance is important, but for any reputable manufacturer, will easily meet the specs.

Sounds to me as if you have an agenda to persuade people to throw cash on snake oil by the back door.

All you need for you need to satisfy yourself, are sources (generators), a scope, impedance and capacitance bridges. If you can measure distortion with any available distortion meter, then good luck.

If you want a time waste project, be my guest.

 

[mojidooji]

markw,

Double blind tests do not meet the requirement of measuring the accuracy of a signal and comparing it to a reference, which is my interest.

Nestor,

Those three (along with phase) is an audio signal.

Rickster,

I'm seeing that you're probably correct. Though my interest is a method for such measurement and comparison, I've not yet encountered an answer to my question. Perhaps providing background was a mistake since readers seen to be focusing on it rather than my question. I was hoping that some of the more knowledgable members (or the people running the site) would have an answer. From my reading, it seems that what I may want is software that can create a time-based fast fourier transformation such as WavePad FFT Sound Analyzer, along with software to compare multiple TFFT graphs. If such a thing exists, I would like to know what it is.

This is not my specialty, so I've had some difficulty figuring out what I need to perform the tests.

 

[mojidooji]

If you're looking for an "audible difference" then the human component would be included in the testing. It's one thing to electronically measure a difference and something else to say we can actually hear the difference.

Assuming some difference was found, using known thresholds as determined via psychoacoustic research.

 

[TLS Guy]

markw,

Double blind tests do not meet the requirement of measuring the accuracy of a signal and comparing it to a reference, which is my interest.

Nestor,

Those three (along with phase) is an audio signal.

Rickster,

I'm seeing that you're probably correct. Though my interest is a method for such measurement and comparison, I've not yet encountered an answer to my question. Perhaps providing background was a mistake since readers seen to be focusing on it rather than my question. I was hoping that some of the more knowledgable members (or the people running the site) would have an answer. From my reading, it seems that what I may want is software that can create a time-based fast fourier transformation such as WavePad FFT Sound Analyzer, along with software to compare multiple TFFT graphs. If such a thing exists, I would like to know what it is.

This is not my specialty, so I've had some difficulty figuring out what I need to perform the tests.

Why on Earth do you want a complex waveform? All wave forms can be broken down to pure sign waves. Your plan will be meaningless utter chaos. Your plan does not have an ounce of merit.

All you need are sign waves.

You need to know resistance per foot of the cable. The bandwidth of the cable, and the transmission impedance per foot.

Then you can measure any signal loss and distortion. For digital signals data loss, but if the impedance is correct you won't improve on it.

I think you are a time waster.

 

[mojidooji]

Sounds to me as if you have an agenda to persuade people to throw cash on snake oil by the back door.

I'm not sure why you would think this, or why we have already reached the point of personal attacks. If I seek to do precisely what I claim, it would be impossible for the results to be used as snake oil. If differences in the measured signals are below the threshold of perception, the test will validate the theory that no differences can be heard. If the differences are above the threshold, the test would falsify the theory and, by definition, no longer be "snake oil". Since all of the data and methods would be public, anyone could repeat the study and verify the results.

If you are aware of research that has already tested this theory, I would be interested in reading it. Keep in mind that in order to address the various claims made by supporters and detractors of the theory, my tests will measure audio reproduction systems rather than just cables.

 

[TLS Guy]

I'm not sure why you would think this, or why we have already reached the point of personal attacks. If I seek to do precisely what I claim, it would be impossible for the results to be used as snake oil. If differences in the measured signals are below the threshold of perception, the test will validate the theory that no differences can be heard. If the differences are above the threshold, the test would falsify the theory and, by definition, no longer be "snake oil". Since all of the data and methods would be public, anyone could repeat the study and verify the results.

If you are aware of research that has already tested this theory, I would be interested in reading it. Keep in mind that in order to address the various claims made by supporters and detractors of the theory, my tests will measure audio reproduction systems rather than just cables.

Here is one on this site for speaker cables on this site for a start.

There are numerous others, if you research your topic, which is a prerequisite and required for any serious research project.

The problem is the snake oil people won't, can't, understand or believe the results.

 

[Swerd]

There is a 20 year-old paper published by the Journal of the Audio Engineering Society. You should be able to find it a university library.

The author made lots of measurements of 12 different examples of speaker wire (including several expensive or “exotic” speaker wires, inexpensive zip-cords 12 to 18 gauge, and 7 gauge automobile battery jumper cables. Measurements were resistance, capacitance and inductance and figures showed how impedance varied with frequency. Further tests were also done with the various cables when differing loudspeaker loads and amplifiers were included.

Davis, Fred E.
Effects of Cable, Loudspeaker, and Amplifier Interactions
(1991) J. Audio Eng. Soc.,Vol.39, No.6, pp 461-468

Abstract
Loudspeaker cables are among the least understood yet mandatory components of an audio system. How cables work and interact with loudspeaker and amplifier is often based more on presumption and speculation than on fact. The literature on loudspeaker cable behavior and effects is minimal. Measurements were made with 12 cables covering a variety of geometries, gauges, and types. The measured data indicate distinct differences among the cables as frequency-dependent impedance, subtle response variations with loudspeakers, and reactance interactions between amplifier, cable, and loudspeaker. In some cases the effects of the amplifier overwhelm the cable's effects. Mathematical models that provide insight into the interaction mechanisms were constructed and compared to the measured data.

Conclusions
If loudspeakers were only simple resistance, then large, low-resistance cables would not be a bad idea. However, loudspeaker systems exhibit frequency-dependent complex impedance that can interact with the reactive components of amplifier and cable. The best response was obtained with low-inductance cables and an amplifier with low-inductance output and a high, frequency-independent damping factor.

These tests have shown that the best way to achieve adequately low resistance and inductance in a cable is by using many independently insulated wires per conductor rather than one large wire. Efforts to reduce the skin effect (such as Litz construction) will help, but due more to the reduction of inductance than the reduction of the skin effect. Inductive reactance is more significant in large cables than the skin effect. If an amplifier does not disagree, larger capacitance in a cable is not significant since this component is comparatively small and reduces amplifier and cable inductive reactance effects.

The best performance was measured with the multi-conductor cables Spectra-Strip 138-064, Kimber 16LPC, and AudioQuest Litz. Smaller multi-conductor cables such as Kimber 8LPC, Kimber 4PR, and Spectra-Strip 191-036 also performed well.

Of the two-wire cables, 12 AWG provided the best Interface, with reactive loads, while both smaller and larger gauges (3-7 AWG and 18 AWG) showed greater high-frequency drop and interaction with capacitive reactance in a load. 12 AWG seems more than adequate, even for demanding systems, high power levels, and reasonable lengths.

The effects of 3.1-m cables are subtle, so many situations may not warrant the use of special cables. Low-inductance cables will provide the best performance when driving reactive loads, especially with amplifiers having low damping factor, and when flat response is critical, when long cable lengths are required, or when perfection is sought. Though not as linear as flat cables, 12 AWG wire works well and exceeds the high-frequency performance of other two-conductor cables tested. By the way, keep the auto jumper cables in the garage!​

All this is well and good. But it avoids the larger question. Do any of these measurably different electrical parameters lead to differences that listeners can actually hear? To address that you need reliably done measurements of human perception – blind listening tests. I can easily understand why no one would want to go to all the trouble of testing this scientifically, but to my knowledge no one has ever published a properly controlled scientific study of this. You could be the first .

 

[mojidooji]

Here is one on this site for speaker cables on this site for a start.

There are numerous others, if you research your topic, which is a prerequisite and required for any serious research project.

The problem is the snake oil people won't, can't, understand or believe the results.

TLS Guy,

Thanks for your response.

As I said, this isn't my specialty and since I am at the beginning of research, I am seeking input on where I might find specific information. Perhaps it is a waste of time, but I would like to know why before I shelve the project. Having read the articles on the audioholics web site are what brought me to ask questions here, so I am aware of the claims behind opinions on this site.

Regarding "snake oil people", certainly some have a vested interest in misleading claims. However, not all of them do; and building a straw man (assuming this is what's happening) to knock down may obfuscate valid claims. What interested me in this topic is that the primary method for determing whether or not differences in various components is audible is extremely unreliable and time-consuming.

 

[markw]

markw,

Double blind tests do not meet the requirement of measuring the accuracy of a signal and comparing it to a reference, which is my interest.

As you specifically mentioned here:

I would like to test the theory that properly working cables do not have audible differences, in real-world conditions.

You clearly state you're testing the audiability of "properly working cables in real world conditions".

As one of your learned pretentions must be aware, this cannot be measured and, as another has stated, the human hearing system must be involved for this. And, to avoid any biases , blind testing is a must.

But, let's be real here, shall we? You're just trolling. You simply want others to play with since you're obviously tired of playing with yourself.

 

[highfigh]

markw,

Double blind tests do not meet the requirement of measuring the accuracy of a signal and comparing it to a reference, which is my interest.

Nestor,

Those three (along with phase) is an audio signal.

Rickster,

I'm seeing that you're probably correct. Though my interest is a method for such measurement and comparison, I've not yet encountered an answer to my question. Perhaps providing background was a mistake since readers seen to be focusing on it rather than my question. I was hoping that some of the more knowledgable members (or the people running the site) would have an answer. From my reading, it seems that what I may want is software that can create a time-based fast fourier transformation such as WavePad FFT Sound Analyzer, along with software to compare multiple TFFT graphs. If such a thing exists, I would like to know what it is.

This is not my specialty, so I've had some difficulty figuring out what I need to perform the tests.

Look into AES and IEEE documents. Also, I would think ASTM have done work in this area. Cables have been tested and they have been blessed many times over the decades. Look for the tests that have been done by the credible agencies I mentioned.

 

[mojidooji]

http://www.apiguide.net/04actu/04musik/AES-cableInteractions.pdf

An interesting read.

markw,

As I said earlier, I would prefer to measure output and compare differences to known tested thresholds rather than measure nothing directly and ask people to tell me what they think they heard. Does anyone here disagree that human perception is of dubious accuracy? Specifically with a multi-varient metric?

Is it reasonable to assume that any difference above the threshold is potentially audible? I think it is. So, the "human" factor is considered by using hearing thresholds to determine if a difference is potential audible.

highfigh,

I am now, obviously, aware of AES and will have much reading to do. Interestingly, Floyd Toole's (of NRC) research is along the same lines I am proposing, though he has concentrated on loudspeakers and has developed a method to train listeners to determine loadspeaker quality in a way that closely matches the results of his lab measurements.

Understanding the differences between double-blind results and the limits of audibility would be a nice bonus and go a long towards reducing uncertainty in listening tests.