dcollins wrote on Fri, 07 December 2007 01:14 |
Shouldn't this be trivial to measure electrically?
|
Spot on.
Talk of microdiodes and other esoteric stuff is a classical attempt to explain a phenomenon (audibility of differences in conductor material) which has not even been proven to
exist yet. The reasoning goes that if one can offer some plausibloid sciency sounding explanation (sufficiently convoluted to buy into oneself) this excuses one from doing the groundwork. The #1 hallmark of pseudoscientific "explanations" is that proving the truth of the explanation is invariably harder than proving the truth of the original claim.
The person proffering such an explanation will then go on to say that they themselves don't have the equipment to prove this, or worse that such equipment doesn't yet exist, but that once the equipment and/or exorbitant funding become available their hypothesis will certainly be confirmed. We're lulled into believing that it's all but ready to be written in schoolbooks and that proof is nothing but an expensive formality which it is more economical to forgo.
You can scare these people witless by showing that their explanation makes predictions that are easily testable with the current state of technology. The microdiode hypothesis predicts that the resistance of a piece of copper wire is non-linear. Needless to say, no nonlinearity has been shown so far. That should not be a surprise. For starts I didn't know you could make a diode using intrinsic semiconductor material. Secondly, what current density do we need to get enough voltage across these "microdiodes" for them to start affecting conductivity? Enough to liquefy the copper or will we need do vaporise it? So far for "microdiodes" and their psychological resonance with "microdetail". Maybe the genius who knows something that only the inhabitants of Betelgeuse might have managed to prove might first see if his suggestions are at least consistent with the stuff we know already.
Still, the most important thing is that explanations are only called for inasmuch as the claimed phenomenon is first shown to exist. This is not a principle, it's a matter of economy. Sometimes it makes sense to skip a controlled trial if the explanation is easier to prove than the original observation (e.g. astrophysical phenomena that are observed only once). Even then, the explanation itself should be proven in full.
In this case the claimed phenomenon is:
"Conductor composition in loudspeaker cables makes an audible difference."
Next, it is attempted to prove the truth of this statement by this explanation:
"Copper has a non-linear resistance
AND
non-linearities of this order of magnitude are audible."
Whoa. We've just replaced ONE hard-to-prove claim by TWO hard-to-prove claims. Worse still, the second
is of the same nature as the original claim itself. Proving the existence of a non-linearity at -180dB still means nothing if the audibility of this isn't also demonstrated. I can't see how we're making headway like this.
Take by contrast the casual observation that a Cary 300B amplifier sounds different from a Halcro DM58. You can skip the listening test here. Just measure distortion of either. The Cary amp measures THD in whole percents, the Halcro amp does not distort measurably at all. The audibility of 2% of distortion has already been rigorously proven so here a measurement is more economical than a rigourous listening test.
Ah nothing starts the day like trashing some pseudoscience good