Hi All,
I ran into a Spanish forum some time ago, and found an explanation why cables can not alter sound...
The reasoning behind it is that an electrical signal travels with (let`s say, on average), 80% of the speed of light through a conductor. That means that, unless we're talking about really long cables, a waveform in the audible bandwidth, can not do a "full stroke" within the length of a normal sized interlink or a loudspeaker cable, and therefore can not be altered.
Maybe a rather far-fetched explanation, but it somehow seems feasible to me... Any thoughts?
Thanks,
Peter
Well, I would hardly call that proof that cables can not alter sound. In general cables do not make an audible difference, but there are other considerations besides the speed of electrical transmission.
@ All,
Here is the original article:
http://www.matrixhifi.com/ENG_marco.htm (http://www.matrixhifi.com/ENG_marco.htm)
Well, the explanation is plainly wrong.
Although they usually don't (audibly) alter the sound, cables can alter the waveform: Just make it long enough and at least the amplitude will be reduced audibly.
A short cable of insufficient cross-sectional area will introduce appreciable resistance, which, combined with the varying inductance with frequency of a speaker will cause a change in the frequency response of the system. If this is considerable, it can be audible.
The length of the cable vs the frequencies of the signal prevents stand waves and similar effects, but that's not the whole story.
Of course quite inexpensive cables have low enough resistance (and inductance at audio frequencies) for them to have no audible effect on the sound.
Cheers,
David.
All real cables have non-zero impedance and can and do alter sound. It is a question of degree. Are the changes audible in the context of other imperfections in the system? For cables of reasonable length and appropriate construction, the answer is generally no.
All real cables have non-zero impedance and can and do alter sound. It is a question of degree. Are the changes audible in the context of other imperfections in the system? For cables of reasonable length and appropriate construction, the answer is generally no.
<nit-picking and elaboration>
Cables in themselves do not make sound. They carry electric signals that can be transformed into sound by a loudspeaker. Talking about the sound-altering abilities of a cable can be misleading, as some believe that sound is a perceptual measure just like color is (according to wikipedia, color is, sound it is not). It is fair to say that the behaviour of typical audio-cables connected to typical audio equipment used for typical audio purposes seems to be well understood by "science"*), and that theoretical models fits measurements very well.
When comparing those models/measurements to what we have learned about the human hearing in listening tests, it is hard to understand the fuzz about audio cables. When doing double-blind listening tests, we have been unable to prove that theory and measurements are inadequate. I believe that economical, sociological and psychological models are better able to explain the phenomenon of audiophile cables, but I am not trained in either.
There is still the remote possibility that some small percentage of listeners have super-abilities and that none of them have ever participated in such listening tests. Or that cables somehow affect the subconsciousness in some way that blind listening tests cannot record. Or that existing science was struck by "statistical bad luck" or bad experiment setup. The scientist should always consider the remote possibility of such strange explanations. Only a fool will clutch such straws as evidence that his faith-based conclusion is as good as anything. The practitioner would be well adviced to spend efforts else-where.
-k
*)Who gets to decide? The majority of those having a relevant Phd? The one scientist with the most citations within the field? No matter what sensible metric, I am pretty confident of the outcome.
Darn! It sounded so GOOD
What amazes me is that all the companies that make hi-end loudspeaker-cable, interlinks and power cables actually accomplish this without any design specifications that include the specific audio properties that can be measured (or at least verified in some way) afterwards, let's say in Quality Control.
Darn! It sounded so GOOD
What amazes me is that all the companies that make hi-end loudspeaker-cable, interlinks and power cables actually accomplish this without any design specifications that include the specific audio properties that can be measured (or at least verified in some way) afterwards, let's say in Quality Control.
Do you see the minute differing design specifications published for brands of gasoline? For toasted oat cereal? 87 octane is 87 octane is 16 gauge copper is 16 gauge copper. Just because something can be measured does not mean it is relevant.
Darn! It sounded so GOOD
What amazes me is that all the companies that make hi-end loudspeaker-cable, interlinks and power cables actually accomplish this without any design specifications that include the specific audio properties that can be measured (or at least verified in some way) afterwards, let's say in Quality Control.
Just because something can be measured does not mean it is relevant.
...But if the design goal is making a cable that, as close as possible, is capable of transporting an audio signal unmodified (a.k.a. a "better" cable), and there are no known measurements that can actually confirm this behavior, how can a company begin a development process for this?
In other words "Because something can not be measured, it does not mean it is irrelevant" either.
In other words "Because something can not be measured, it does not mean it is irrelevant" either.
Anything that a cable can do to an electrical signal can be measured. And, the measurement equipment is far-far-far better than human hearing.
I believe that the article only applies the length argument to cable inductance.
Capacitance and resistance are given separate coverage.
In other words "Because something can not be measured, it does not mean it is irrelevant" either.
Anything that a cable can do to an electrical signal can be measured. And, the measurement equipment is far-far-far better than human hearing.
Which is exactly my point. Yet, manufacturers of cables claim to be able to develop cables that have properties they can not verify by known testing methods. I find that odd.
In other words "Because something can not be measured, it does not mean it is irrelevant" either.
Anything that a cable can do to an electrical signal can be measured. And, the measurement equipment is far-far-far better than human hearing.
Which is exactly my point. Yet, manufacturers of cables claim to be able to develop cables that have properties they can not verify by known testing methods. I find that odd.
You find it odd that marketers of a product claim properties which can not be disproved?
How delightfully uncynical.
Now I don't know the validity of this measurement and if the findings where reproducible, but a e-shop here in Sweden Fynda (http://fynda.se) has performed a measurement with the same cable in 0.75mm and 2.5mm and as you can see the measurements are not linear with frequency, so the behavior seams to match that of 2Bdecided's explanation.
Jamo D450:
(http://fynda.se/bilder/kabelmatningjamo.jpg)
Dynavoice Magic F-6:
(http://fynda.se/bilder/kabelmatningmagic.jpg)
Translation please?
wow, terrible cables if true.
.
I'm guessing that what we are looking at is basically the frequency response curve of the speakers and how that differs with the speaker cable.
Yes, I always suspected the Dynavoice to have slightly more sub-bass than the D450 (but I never dared to mention it here, TOS#8 you know...)
Good thing you only said "suspected," since TOS #8 explicitly prohibits graphs.
wow, terrible cables if true.
You do realise that this is a graph of amp->speakers and not the cables right
Translation please?
It basically only sais that the scale is dB, that a difference in 3dB is the same as doubling of volume and that both cables are terminated with gold plated banana pins.
So then it really provides us with no useful information about the sound of cables?
Why not plot the difference between the two cables?
...I think I know the answer.
Considering that one cable has a 0.75mm^2 cross-sectional conductor area (~19 AWG), and the other has 2.5mm (~13 AWG), all this test does is prove the laws of physics correct: less copper = higher resistance/foot, which, in this case, would be easily measurable given the substantial differences in the wire sizes.
A much more interesting/amusing test would be one between two cables from different manufacturers (or maybe better yet, two different "grades" of cables from the same manufacturer) with the same cross-sectional conductor area.
Why not plot the difference between the two cables?
...I think I know the answer.
If you're suspecting that it would be flat, a close look should convince you that is not exactly the case. Difference appears to vary by 0.5 dB or so across the measured frequency range. This is caused by complex impedance of the speaker. This is potentially an audible difference even after level match.
Considering that one cable has a 0.75mm^2 cross-sectional conductor area (~19 AWG), and the other has 2.5mm (~13 AWG), all this test does is prove the laws of physics correct
And it hints that 19 AWG is not quite big enough for wiring these speakers.
If you're suspecting that it would be flat, a close look should convince you that is not exactly the case.
Well within 2 dB, Notat, with the biggest delta being at the low end trending down as it goes up in frequency. It's damn near flat, no question about it; any ripple would be dwarfed by what happens by simply moving your head.
I'd also like to see additional plots, hopefully giving some king of indication as to how precise the measurements were. Do you care to comment on what that delta might be?
And it hints that 19 AWG is not quite big enough for wiring these speakers.
Suggesting that someone needs a thicker gauge than 19 AWG is hardly unreasonable, nor necessarily expensive (though I suppose "expensive" is subjective). Suggesting from these plots that clear audible degradation is a slam dunk would be unreasonable, however.
So then it really provides us with no useful information about the sound of cables?
Sorry! But then again that wasn't my goal, I just wanted to point out that even a change in only the resistance and inductance can introduce non linear differences. This is not about cable A is better/worse than cable B.
Yes a graph of the differences would be more interesting since that is basically what I wanted to display, but since I don't have access to the raw data I couldn't do that. But I think that it is quite clear that there is close to 1dB differences on some frequencies and 0.5dB in others indicating that it is not a linear difference. Since we don't know what measuring equipment was used we cannot know if these changes are due to precision, so this is far from a proof, it was just that I found these images and thought them interesting.
so this is far from a proof, it was just that I found these images and thought them interesting.
Yes, I agree. It's just another example of what we already know: if the cable is too thin/long, it's resistance will be too high and hence significant wrt the varying impedance of the speakers. Easily measurable (and, if taken to extremes, audible) frequency response variations will occur.
Respected 2.5mm speaker cable is less than £2 a metre. By respected, I mean genuine copper, with plastic coating that doesn't react with copper or air or human fingers, and has been available for years.
There's plenty of stuff on eBay that works out less then 50p a metre. I have some, but I'm not qualified to say whether the metal and plastic are genuine and will last ten years. Ask me in ten years . At least it has the polarity marked, which some brands of "respected" budget speaker cable do not.
I previously used 0.75mm mains cable (flex), sometimes over a long and unequal run. There must have been measurable loss and frequency response variation on at least one channel, but due to the room and speaker placement that cable was used in, that set-up was actually the
best that my own stereo has ever sounded. So much for speaker cables being important! (TOS 8? How exactly do you propose I ABX different rooms in different houses? )
Cheers,
David.
(TOS 8? How exactly do you propose I ABX different rooms in different houses? )
A Portal gun (http://firsthour.net/screenshots/portal/portal-gun-blue-orange-portals.jpg), obviously.
(TOS 8? How exactly do you propose I ABX different rooms in different houses? )
A Portal gun (http://firsthour.net/screenshots/portal/portal-gun-blue-orange-portals.jpg), obviously.
Or much simpler: just find a theatre with a revolving stage...
wow, terrible cables if true.
You do realise that this is a graph of amp->speakers and not the cables right
busted ;>
Why not plot the difference between the two cables?
...I think I know the answer.
If you're suspecting that it would be flat, a close look should convince you that is not exactly the case. Difference appears to vary by 0.5 dB or so across the measured frequency range. This is caused by complex impedance of the speaker. This is potentially an audible difference even after level match.
Well steep and/or rugged graphs might make it difficult to spot variability of the difference with the eye, so it is certainly a point taking the difference.
(I have to admit though, that the first answer that struck me when Greynol asked «Why not plot the difference», was «because it would be moderated» :-o)
Perhaps it was open to interpretation, but that wasn't the correct answer.