Here's a bit more analysis, for what it's worth. My question would be, what audible difference exists between the Kimber 8TC and the "standard 10 AWG" Bluejeans cable, and is that difference audible?
In an effort to answer that, or at least scale the results, I took advantage of an online circuit simulator, Circuit Lab
(
https://www.circuitlab.com)
Taking the data from both reviews, I constructed a rough simulation of the cable including measured R, L and C. I did not include the skin effect because it is not simple to simulate using the available simulation components, and I recognize that skin effect may add additional differences to the results.
The basic cable equivalent I used is this:
View attachment 12105
Values shown are for 20' of Kimber 8TC. The values for 20' of Bluejeans "standard" 10AWG were also simulated. I do recognize that the network is not an accurate representation of a long cable, which actually has all three properties distributed over the length. To simulate that accurately the network would have to be many sections of R, L and C to represent the cumulative effect. This would essentially create a multi-pole filter, so the extreme HF characteristics would be different than a single section network, however for the purposes of a 20' test piece, the difference in performance would be quite small.
I needed a load as close to a real speaker as possible. Rather than invent one, I borrowed a circuit from a Stereophile article from 1995, "Real-Life Measurements" by John Atkinson. Atkinson modified a circuit by Kantor by adding a Zobel network to more closely approximate the high frequency characteristics he'd measured on real speakers. The circuit performance seems reasonable, and can be seen on the link below. It may not simulate any specific speaker, but at least it's similar to speakers in general, and frankly, was handy.
View attachment 12106
Input was applied at the junction of R9, R11 and R12, with reference to ground. The Stereophile article discussing this network is below.
Real-Life Measurements | Stereophile.com
AC frequency response simulations were run on the cable simulation network with values entered for each cable. Then I had my turn to fight with Excel's graphing (dis)abilities. What a mess. Anyway, the frequency response at the simulated speaker input for each cable is shown. The response was run from 20Hz to 20KHz, and the vertical axis is scaled in dB, with the maximum being a fraction of 1dB.
View attachment 12102
Just for grins, here's the difference between the two 20' cables measured at the load:
View attachment 12103
Now, as to what you are looking at, this is essentially a voltage frequency response measurement at the load input. It does not represent an acoustic response in a room, as the measurement is electrical only. The difference plot is what's most interesting, though. Note, for example, the cable with "worse" measurements shows more HF response, which, if we held to the idea that L and C were "bad" if they were higher, doesn't seem to make sense. However, speaker loads are reactive. What we have is a resonant circuit with the cable and the load. The actual resonant frequency is far above the audio band, but we are seeing the beginning of its effect. It is evident that lower distributed R, L and C contribute to flatter overall response, even though the specific amount of change is miniscule.
To offer some conclusions, measurements of electrical parameters of short lengths of speaker cable may seem to exaggerate their real impact, or even the exact direction of change if not considered as part of the entire circuit. The net change between the cables is well below audibility, even at the extreme. The maximum difference between the two occurs at a frequency where small changes are difficult to hear. To support that claim, I give you a graphic from David Clark's 1982 AES paper, "High Resolution Subjective Testing using a Double-Blind Comparator" in which he determined the ability to hear level matching at different frequencies. While not exactly indicative of the ability to hear level changes as they relate to frequency and Q (bandwidth) of the response deviation, it does illustrate the relative ability to discern level variance vs frequency.
View attachment 12104
Each speaker will present a different load, so the FR curves and difference curves shown above will not represent the exact differences in specific cases. The load simulator is intended to be an average, more or less typical two-way bookshelf speaker. More complex speakers will have different results, but there's not reason to expect any of them to be extreme enough to become audible.
In short, I believe the differences between the Bluejeans cable and the Kimber 8TC to be completely inaudible.
