Originally Posted by chesebert
geez...all these talk and we end up with nothing useful!
What were you expecting?!
I have a few (useful?) thoughts to add.hciman77
has provided a link
with measurements with respect to the skin effect. According to these, an average speaker cable of 10 feet produces a drop-off of almost 3% of loudness intensity (-0.25 dB) at 20 kHz due to the skin effect. That's actually not much and well within the bandwidth of modern electronics' variation. It's certainly nothing one could deduce a dull or underrepresented treble of. The main question is: «Is this audible at all?» My experience tells me that it most likely is. On the other hand, I don't have measuring instruments which would allow to make conclusions about the correlation between certain measuring values and heard sonic characteristics with electronics and cables.
Nevertheless, I have made a lot of cable experiments with the goal of minimizing the skin effect -- by means of a drastic increase of cable surface relative to cross-section surface. I achieved this by using magnet wires with (finally) extremely low diameter (down to 0.04 mm). With the clear trend: the more extreme the ratio, the more treble sparkle. Not exactly higher treble intensity, rather higher quality, clarity and detail. At the same time this design approach could also lead to overemphasized brilliance and smoothness or even sleekness, especially with bigger cable lengths, so that a conventional braided cable with its drier characteristic could offer the impression of higher accuracy and definition, despite the duller upper end.
Apart from geometry, the materials also played a certain role. Silver and silver plating had a characteristic impact. As many other audiophiles have discovered, silver tends to sound bright and detailed, up to edginess in some cases.
As stated in an earlier post, I have occupied myself extensively and intensively with the construction of loudspeakers and thereby experienced unexpected phenomena: Even barely measurable frequency-response or (accompanying) phase-response changes can have audible consequences. So I switched a variable capacitor of (max.) 470 pF parallel to C2 -- the second component of a 4th-order high-pass filter for a tweeter with crossover at 1730 Hz -- with a total capacitance of about 14.3 microfarad.(See black box with knob at the right side)
You can imagine the influence of the difference between 14.3 and 14.30047 or e.g. 14.30022 microfarad: it's barely measurable in terms of frequency or phase response, although theoretically it leads to a microscopic change of both. However, to my ears different positions on the scale revealed a multitude of «sweet spots» -- meaning adjustments which seemed like making «click», allowing the music to appear as a homogenous, coherent entity -- and an overwhelming area where this sensation was absent, although there was a constant minimal change of characteristic nonetheless. (I should add that the crossover-network components as well as the involved speaker chassis were pedantically measured beforehand and brought into exact accordance with established filter formulae.)
Later I detected that I could get similar effects by pushing the tweeters (on the top of the cabinets) forward and backward by fractions of millimeters or even changing the position of the whole speaker by similar amounts -- also implicating a slight shift of phase between tweeter and woofer relative to my listening position. I have to admit that I certainly wouldn't be able to get such results with any unfamiliar equipment. They were just possible after many hours of intensive occupation with the characteristics of this pair of speakers.
Back to the 3% drop-off at 20 kHz. As said, this is not much, and most people won't be able to detect a difference of volume level in that range between two samples; add to this that 20 kHz may not be audible to adults anyway, so if we take a drop-off in the range of 0.15% at say 12 kHz into account, this seems even less likely. But in fact we don't deal with two samples of different loudness intensity, but we're talking of sonic-balance variation! It may not be possible to associate a 3% drop-off with a lack of treble, though, but as modern (headphone) amps show, they can sound clearly different despite minimal measuring variations. BTW, the main candidates (or should I say: my
main candidates?) for responsibility for sonic differences are harmonic distortions. This despite the fact that THD in most cases resides below 0.01%.
Depending on the slope characteristic, a HF drop-off may be perceived differently: once as dull and lackluster, once as accentuation of the lower treble and even as harshness or graininess. Keep in mind that a frequency response in the form of a straight line is perceived as uncoloured, even if it's slightly tilted to one or the other end. The same applies to a slightly convex or concave curve, as long as it's smooth (and of course not too extreme). But once the curve is inhomogeneous, it's perceived as colored.
I can't pretend if the perceived coloration in the case of minimal deviations such as 3%/0.25 dB isn't in fact a consequence of inevitable phase distortion, but since FR and phase distortion always appear parallel, it doesn't matter that much. Just one more example. I once got my Metaxas Solitaire (a power amp with extraordinary HF bandwidth) modified: The only sonically relevant modification was the removing of a small inductor coil (meant to prevent HF oscillation) right before each channel's speaker terminal with a value of 0.01 mH -- corresponding to a low-pass corner frequency of 127 kHz. After the modification, the amp sounded significantly smoother (although you would rather expect a smooth low-pass filter to create smoothness) and showed finer «grain».
The conclusion of this long post: My theory is that minimal FR- and phase-distortion patterns in cables are responsible for the perceived sonic characteristics..