Saturday 22 February 2020

MEASUREMENTS: Archimago's Colorful Speaker Cables, KnuKonceptz, AmazonBasics, and "freebie" speaker cable. (And changes at Audiophile Style.)


As hinted at with the article a couple weeks back of my simple DIY speaker cables, let's spend some time considering the basic speaker cable measurements, how these affect the hi-fi system, and get a sense of the magnitude of effects to be aware of.

If you read the audiophile publications, it's not uncommon to see articles paying lip service to the importance of science in cables and measurements, yet at its core, they're often written to perpetuate fears, uncertainties, and doubts (FUD). IMO, this recent Positive Feedback article is a good example in that it acknowledges resistance, inductance, and capacitance but goes out of its way to describe complexities without actually giving you any contextual information about the values of these quantities. Notice the title. The "snake oilish" nature of something isn't nullified just because acknowledgement of the term is used. And certainly, denial is to be expected from supporters of "snake oil".

Despite the example above (and countless others over the years), independent writers have discussed speaker cables and have provide quantification of their potential contributions. I remember this article from Audio Science Review showing significant variation in resistance between 12AWG cables (important to keep in mind!). A number of years before, there was a series such as this on Audioholics looking at the 3 main parameters for AC signals - resistance/impedance (R), inductance (L), and capacitance (C). Secrets of Home Theater and High Fidelity occasionally had cable reviews with measurements. Elliot Sound Products has a fantastic article circa 2003. Then there's this one from Roger Russell of Acoustic Research and McIntosh - essential reading for the general audiophile.

Even further back in 1991 with The Audio Critic issue 16 (PDF p.39-44), the late Peter Aczel wrote an excellent article showing some frequency graphs (see "The Wire and Cable Scene: Facts, Fictions, and Frauds Part II" p.51) modeled using difference cables connected to a Boulder amp and Carver "Anazing Loudspeaker" - great history back to the 80's in that article! He even compared with a tube amp with low damping factor and refers to audiophile magazine articles and writers of his day (Martin Colloms is brought up whom we still hear about these days).

That's quite a bit of information already and I think reading the articles above (even more links at the end) will give the audiophile more than enough historical and technical perspectives along with useful advice on what is or isn't important.

I suspect every audiophile regardless of "objective" or "subjective" leaning would agree that "The best speaker cable is no speaker cable!". That is, perfect transparency would be a direct connection between amplifier to speakers with no opportunity for a cable to introduce any potential power loss, change frequency response or time-domain characteristics of the signal. Alas, nothing's perfect, we do need cables in the real world, and as a result, we do have to consider the electrical properties. As I'm sure you've seen many times before, the basic schematic for a cable can be laid out like this - Rs = series resistance, Ls = series inductance, Cp = parallel shunt capacitance:


Remember that this the "lumped-element" circuit and if we want to be even more accurate, we could view the cable as countless little series resistors and inductors plus parallel capacitors. Depending on the overall strength of those 3 parameters, the cable can act to impede the voltage to our speakers (signal loss) or act as a type of very weak filter that can subtly change frequency and phase response of the AC signal.

Having said this, using the example of the DIY cable, let's measure it to provide a practical example. What's needed is an "LCR meter" with adequate precision since we will be looking at milliohms, microhenries, and picofarads! At home, I have a reasonably priced REED Instruments R5001 Passive LCR Meter (~US$250). I compared this to an excellent B&K Precision 880 I have access to at work and the results are very close down to +/-5 pF, +/-0.1μH, +/-1 mΩ with various low value capacitors, inductors (a few 10μH and 22pF parts I have around here), and a 20' length of 12AWG zip cord for AC resistance. Note that I'm certainly not claiming lab-calibrated absolute accuracy for such small values here at home although I think the results should be close; I'm simply aiming for the opportunity to allow us to compare the different cables' electrical parameters and appreciate the basics in context.

REED R5001 package with various basic test leads as well as tweezer SMT tester in plastic bag...
It's rather convenient measuring speaker cables terminated with banana plugs as one could just plug the cables directly in to get the reading. Just make sure to calibrate first with the shorting procedure.

Like many LCR meters these days, this one can give me results at various AC frequencies - 100/120/1k/10k/100kHz. For audio work, it would be nice to have had a 20kHz reading but this is just fine and the reading at 10kHz should be very close to what we would see at 20kHz. There is also a DC resistance measurement which I find is less precise with this device so I'll focus on AC resistance at the various frequencies for this report.

Here's a picture of the DIY cable from a couple weeks ago connected to the meter for parallel capacitance (Cp) measurements (for the "official" results, I measured them on the table rather than on the floor rolled up like this):


For series resistance (Rs) and inductance (Ls), I just secured the bi-wire + and - leads together tightly to maintain good surface contact and plugged in one of the banana plugs. Also for zip-cords, I shorted the parallel run when measuring.

Here are the results for my 9.75' length of DIY cable. While the LCR meter can measure both 100Hz and 120Hz, there is minimal difference so I'll just report the 100Hz value:


We can then calculate the amount of each parameter per foot as a convenient standard for comparisons:


Within the audible frequencies (at least measured up to 10kHz, but we can interpolate to 20kHz), the AC resistance is nice and low at <1mΩ/ft owing to the large combined 9 gauge copper conductor. Remember, this is all just with the GLS Audio locking banana plugs with the small set screws - welding/soldering could improve resistance even further.

Suppose we're use a 10' Archimago's Colorful Speaker Cable (total of 20' round-trip), with resistance as measured above, we can calculate approximately the amount of signal loss. To be conservative, let's use a 4Ω speaker load and we'll use the 10kHz measured resistance value as representative for audio frequencies:

Cable resistance Rcable = 0.82mΩ/ft x 10 ft x 2 (round trip) = 16.4mΩ

We can then convert this to a measure of Insertion Loss = 20 log (Rload/[Rload+Rcable]). Assuming an ideal voltage source amplifier with output impedance at 0, dB loss into a 4Ω load would be 20 log (4/[4+0.0164]) = -0.036dB.

Insertion Loss (10' Archimago's cable to  load) = -0.036dB

While with the Colorful Speaker Cable the loss is essentially nothing, we can imagine that with more resistive cables, the loss could be important if the load impedance is low. I have not seen any specific recommendations for what would be a "threshold" drop, above which we should avoid for hi-fi purposes. I think up to -0.25dB loss (connected to 4Ω load) would likely almost never be noticeable; for context, often our digital pre-amp volume controls jump in 0.5dB steps. -0.25dB for a 1W signal across 4Ω (2V/0.5A) would be about 2.5% loss. Working backwards, this means that the total resistance of the wire loop (back & forth) should be no more than 120mΩ to a 4Ω loadSince most speakers are of higher overall impedance than 4Ω, the signal loss will be even less and we can consider this 120mΩ total cable resistance as a conservative value to keep in mind.

Beyond resistance, speaker cables have "filtering" effects on account of their series inductance and parallel capacitance. Again, an ideal speaker cable should be absolutely transparent with no change to frequency or phase.

So, how much of a filter is Archimago's Colorful Speaker Cable? Let's have a look using the inductance and capacitance measurements, plugging in the numbers into an RLC filter calculator to obtain frequency analysis graphs.

Suppose again we use 10 feet connected to a hypothetical speaker load of flat 4Ω resistance, using a solid state amp with very low output resistance (high damping factor). Since we're interested in audio frequencies, let's purposely overestimate by using the average of the 10kHz + 100kHz results for inductance and capacitance. This works out to 1.07μH inductance and capacitance of 414pF for 10'. Here's what the frequency and phase graphs would look like up to 60kHz as the signal is presented to one's speakers. I've put a pointer at the 20kHz audible limit for easy reference:


As you can see, at 20kHz, the frequency roll-off for this cable is much less than -0.01dB even (it's around -0.0046dB to be more exact). The phase response is <2° at 20kHz, again totally benign. If we consider the effect of crossovers and speaker frequency responses, this cable "filter" effect is simply trivial.

Remember that these graphs are constructed with a hypothetical speaker load of only 4Ω. As discussed previously, speakers are reactive devices and the impedance (Z) varies quite substantially depending on the frequency of the signal. For most box speakers with cone tweeters, impedance by 20kHz would be more like 6Ω+ due to voice coil inductance which would make the "filtering" effect of this cable even more benign (with 6Ω load, at 20kHz, we're looking at -0.002dB, and less than -1.4°). It's typically with electrostatic speakers that we see such low 4Ω impedance into high frequencies.

Let's run some measurements to compare with other cables I have here. Note that for each cable, I will be using the same set of GLS Audio locking banana plugs to keep contact resistance constant, measuring at least a 15ft length to get a per foot average:

KnuKonceptz Kord Kable 12AWG


This is the zip-cord cable that I'm using for the Colorful Speaker Cable above (as per construction post). Basically it's two parallel runs of 12AWG oxygen-free copper (supposedly 4N = 99.99%). It's more flexible than most and the translucent bluish color is quite nice. It's not expensive - US$30 for 50'. How does this cable measure per foot?


Insertion Loss (10' KnuKonceptz Kable to  load) = -0.13dB

That's good. As we expect, a 12AWG cable like this would have very low signal loss at a length of 10'.

And if we plug in the inductance and capacitance values, here are the frequency response and phase graphs for the 10' length connected to a 4Ω load (again, let's use the average of 10kHz and 100kHz with Ls =  1.63μH and Cp = 186.5pF to slightly overestimate what 20kHz roll-off would be like):


A minuscule -0.011dB, and -3° phase shift at 20kHz. Nothing here worth getting concerned about for such a short but typical length of cable for audiophile systems.

AmazonBasics 14-Gauge Audio Speaker Wire - 99.9% OFC


Easy to get and inexpensive at US$30/100', this 14AWG cable from Amazon is comparatively stiffer than the KnuKonceptz. This stiffness to a large extent is due to the thicker individual copper strands, (64 strands total), red and black twisted pair. There's a fire resistant white "CL2 rated" cable jacket.


Insertion Loss (10' AmazonBasics 14AWG OFC cable to  load) = -0.18dB

A -0.18dB attenuation from cable resistance is still small in the grand scheme of things.

And the high frequency roll-off remains very tame (for 10' using Ls = 1.39μH, Cp = 287pF):


At 20kHz, we're looking at only around -0.008dB with -2.4° phase shift. No surprise, 10' of 14AWG OFC is good.

AmazonBasics 14-Gauge Audio Speaker Wire - Copper Clad Aluminum (CCA)



Probably one of the most inexpensive, but reasonably thick zip-cord cables you can get currently. At US$11 for 50', this cable from Amazon has clear outer insulation, is reasonably easy to work with and looks like copper. However, this is not a pure copper conductor:


Because copper is more expensive, this wire only has a thin outer layer of copper with an internal core made of aluminum (30-50% lower price for CCA wire). Resistance would be intermediate between aluminum (~60% of copper) and pure copper. Once you rub off the copper surface as happened with a few of the strands in the picture above, you'll see that it's silvery inside.


Insertion Loss (10' AmazonBasics 14AWG CCA cable to  load) = -0.31dB

We can see the effect of using CCA quite easily. Resistance is increased over that of the AmazonBasics OFC cable despite being the same gauge by 50-80% which was more than I had actually expected. Otherwise, inductance looks about the same as the OFC and lower capacitance presumably related to the dielectric used, the zip-cord geometry instead of twisted pair, and lack of the outer jacket.

This cable's insertion loss is over the -0.25dB arbitrary "hi-fi threshold" I discussed above when connected to a 4Ω load. I don't think it's unreasonable to suggest that even for a 10' cable run, for "hi-fi" purposes where we expect minimal compromises in sound quality, it's recommended that audiophiles simply use speaker cables better than 14AWG copper clad aluminum even if we're unlikely to hear a difference.

Graphs for conservative estimate of 10' cable using Ls = 1.74μH, 174.5pF showing the cable's "filtering" property:


This graph is not much different from the AmazonBasics 14AWG OFC cable as the inductance is similar. We're seeing attenuation around -0.013dB and concomitant phase shift of around -3° by 20kHz.

Finally, let's end off with the "worst" speaker cable you might run into at some point; a generic low-quality zip cord that comes with various mass-market amps and speaker.

Generic 20AWG Speaker Wire - Copper / Aluminum


What I have here is a 33' length of some unnamed 20AWG cable that came with an old home-theater-in-a-box from about 15 years ago. I think this wire came with a 5.1 Denon surround system my brother-in-law bought.

As you can see on the right, the conductors are made of different materials. One of the runs appears to be copper, the other aluminum (no copper cladding). For the series resistance, I averaged the measured results of the copper and aluminum runs (overall the aluminum run was about 20% higher than the copper). Inductance was not much different between the copper and aluminum wires.


Insertion Loss (10' Generic 20AWG Cu/Al cable to  load) = -0.55dB

With around -0.6dB signal loss through a 10' cable of this type, one should notice that this sounds different (probably "worse" as per discussion last week) compared to thicker 12AWG, pure copper cables switching back and forth in a direct A/B test.

Taking the average again between 10kHz and 100kHz to get an estimate of "filter" parameters, a 10' length will have approximately Ls = 3.34μH and Cp = 129pF to a 4Ω load:


At 20kHz, we're still looking at only a small -0.045dB attenuation compared to lower frequencies and about -6° phase shift due to the "filtering" effect of such a cable even though the resistance is significant enough to create almost -0.6dB drop into 4Ω load.

While the magnitude of these changes is far from terrible, for a "hi-fi" system, the insertion loss should be better.

Subjective

I don't think I need to spend much time on this :-). After I put together my Hypex NC252MP amplifier in December, I first used the AmazonBasics 14AWG OFC cable for about 3 weeks, then "upgraded" to the KnuKonceptz Kord 12AWG OFC for 2 weeks, then finally put together my Colorful Speaker Cable and have been using it for a little more than month at this point.

Did I hear a difference? No, as much as I would love it if my Colorful Cable sounded better because of better objective parameters and looks nicer. They all conveyed a nice sound stage, no "veils were lifted", no change in audible details or subjective PRaT. Remember, we're talking about 10 feet of cabling to each speaker with milliohms of resistance, an additional microhenries of inductance, and picofarads! I don't live under a radio tower to pick up audible interference using unshielded speaker cables so didn't experience any issue there either.

To be honest, I didn't try listening to that 33' of copper/aluminum 20AWG cable - we are talking hi-fi here and I just don't see the need when 12AWG OFC is cheap.

Conclusions...

I think it's clear from these measurements using various "non-exotic" speaker cables, quantified for LCR, that there's nothing all that magical between the measured values and the physical attributes of the cables. As you can see, zip-cords are rather boring lengths of wire and maintain low total impedance and capacitance at normal lengths - very poor "tone controls". As one who seeks transparency and fidelity, this is as it should be! ;-)

Of the three parameters, the most important are RESISTANCE (R) which when high leads to voltage/signal loss and INDUCTANCE (L) which when high will result in a dampening of high frequencies. Of the two, when it comes to actual potentially audible "decibel" change to the signal, RESISTANCE is the most important - thin cables, poor contacts, low metal conductivity - will result in your amp needing to work harder for the same overall amplitude due to signal loss. As for INDUCTANCE, notice that none of these cables, even the low quality 20AWG Cu/Al one, resulted in anything more than -0.05dB attenuation at 20kHz using a 10' length into 4Ω load. As a passive "filter", these cables I've shown here are extremely weak which is "good" of course because for the sake of transparency, we do not want our cables to significantly "color" the sound.

CAPACITANCE doesn't affect the audio frequencies like resistance and inductance. However for cable designs, there is typically an inverse correlation between capacitance and inductance. There are some cables where inductance is minimized to flatten out the frequency response even more despite already minuscule roll-off at 20kHz with zip-cords. To do this, the trade-off typically results in a cable with very high capacitance. We see this kind of trade-off done with heavily braided and Litz cables such as the Polk Cobra (7.5mΩ, 0.026μH, 500pF per foot), or less extreme, the Kimber 8TC (2.19mΩ, 0.037μH, 100pF per foot). Remember that very high capacitance cables can result in ultrasonic oscillations with some older amplifiers (Phase Linear amps from the late 70's + Polk Cobra cables have been said to be a poor match). These days, it's rather unusual to see such issues with modern amps and some cables may include a Zobel network (eg. the Goertz reported on here).

Ultimately, I think it's reasonable not to sweat the small stuff with speaker cables. Nothing wrong with adding "bling" and having fat cables but you're unlikely to hear a difference unless the designer purposely played with the LCR values beyond that of these zip-cords. Note that this is quite possible with some exotic cables. For example, if you look at the MIT Cables patents here, you'll see things like "Audio signal transmission line with low-pass filter (series or parallel inductance)" listing "20 microhenries to 1 millihenry" of inductance added to the cable. Wow! If we added just the lower amount of 20μH to 10' of KnuKonceptz Kord 12AWG, we'd create a low-passed-filtered cable with -0.5dB roll-off at 10kHz and about -1.5dB at 20kHz. An example of a "tone control". Do we seriously want to do this with "high fidelity" audio cables? Is this the best solution for "bright" speakers/amps?

A 12AWG copper zip-cord speaker cable like the KnuKonceptz Kord or likely the AmazonBasics 12AWG OFC (not measured) should have -0.25dB or less insertion loss at lengths up to a 20' loop into a low 4Ω load. This should satisfy the majority of audiophiles I suspect using a conservative threshold. Typical zip-cord inductance would not result in significant high-frequency attenuation at this length. For example, again using the KnuKonceptz measured values, a 20' length would result in around -0.013dB at 10kHz and less than -0.05dB at 20kHz. One's speakers' impedance variation, the speaker driver's frequency response, and room acoustics would totally swamp whatever contribution from the cable!

As a general rule of thumb for these LCR values and in order of importance:

1. Resistance (R)
   Low: <12AWG copper / <2mΩ/ft
   Medium: 12 - 16AWG copper / 2 - 5mΩ/ft
   High: >16AWG copper / >5mΩ/ft

2. Inductance (L)
   Low: <0.1μH/ft
   Medium: 0.1 - 0.5μH/ft
   High: >0.5μH/ft

3. Capacitance (C)
   Low: <25pF/ft
   Medium: 25 - 75pF/ft
   High: >75pF/ft

Based on these values, the DIY "Archimago's Colorful Speaker Cable" (<1mΩ, <0.15μH, 45pF/ft at 10kHz) would be classified as low resistance, low-medium inductance, with medium capacitance. A well-balanced combination I think; plus there should be enough "bling" in the appearance to make the cable look "audiophile". I think this last bit has always been the most important as far as anyone can see or hear :-).

Needless to say, I believe all audiophile speaker cable reviews should include measurements of these basic LCR value (here's looking at you The Absolute Sound and Hi-Fi+). Especially for $$$$ cables, wouldn't it be good for readers to anticipate losses based on resistance and the length they need, consider if the cable acts as a low-pass "tone control", and consider if capacitance might be too high for their amplifier? In fact, this might also be good for manufacturers so they can explicitly put warnings against certain cables that might not be compatible with their design - save on repair hassles for everyone!

If we look around the Net, we'll see articles and comments suggesting complexities beyond the basic measurements. Manufacturers will sometimes make a big deal about "skin effect" and "proximity effect" and why they prefer Litz wire. We can see that the skin effect results in higher resistance at 100kHz compared to 10kHz and below in my measurements - no need to get all excited since audio frequencies only go up to 20kHz. Furthermore, last time Mikhail linked to this EDN article discussing cables attached to crossover/speaker loads vs. resistive loads, with the long lengths used in studios, and picking up EMI. We can also find academic articles suggesting variation in low-level distortions with different speaker cables (like this). Perhaps experiments repeating some of this would be worth trying in the future. Again, remember to pay attention to the magnitude of differences reported and not get too anxious with small stuff.

[As an aside, notice that the "Archimago's Colorful Speaker Cable" 2x12AWG wires/run AC resistance (Rs) measurements show a big jump from 10kHz to 100kHz. Compare this to the thin 20AWG cable with only a small gain from 10kHz to 100kHz. This demonstrates that the Colorful Cable with its thick 12AWG wires have higher skin effect than the thinner 20AWG at ultrasonic frequencies. Ultrasonic frequencies will travel out at the surface ("skin") of the wire making the effective gauge appear smaller, thus higher resistance. While this is not "ideal", such frequencies are inaudible and since I am running the Class D Hypex NC252MP amp with 420kHz switching noise, skin effect lowering transmission of this ultrasonic stuff plus a bit of inductance is just fine...]

I hope that after showing the effects of standard cables, the Positive Feedback article I linked to at the beginning no longer seems all that thorough nor suggests that we need worry about speaker cables at normal lengths (the usual 6-20 feet for most of us, I suspect). Some of the discussions in the article like "velocity of propagation" down copper cables can be complicated to calculate given the variables (search for Telegrapher's equation and propagation speed if you want to see some formulae), but we're looking at fractions of the speed of light; something like 0.5-0.8c or around 500,000,000ft/s give or take a few million units. What's the potentially audible problem even if the dielectric slows things down somewhat?! Heck, even if right/left cables are grossly asymmetrical (but still realistic) like 20' one side and 30' the other, it'd be ridiculous to claim that we can hear the temporal difference in signal transmission!

To quote that article for contrast (bold added for emphasis):
"As with so many other things about cables, when capacitive discharge effects or the need to balance fields to optimize sonic performance were first announced, many greeted them as just that much more "snake oil" or manufacturer's hyperbole. The truth, though, is that they are just as much facts of physics as the easily-measurable differences in velocity of propagation brought about by changes in cable insulation, and are easy to hear by anyone with open ears, a clear mind, and a decent system
Don't just take my word for it; go listen to a cable!"
I don't know about the magnitude of Mr. Skoff's claimed "capacitive discharge effects" or "need to balance fields" for audio cables since the article provides no details nor can I find anything useful on his RSX Technologies site (at most we have vague wordy "white papers" that say essentially nothing about the products). At least I do agree with spending some quality time listening. Even better yet, try some unsighted listening instead of looking at the "bling" because that's how one truly "hears" with "open ears, a clear mind" - "easy" to hear a difference he says? ;-) I suspect if one does this and re-reads that PF article, one will find that "snake oil" and "manufacturer's hyperbole" are likely being peddled.

For your continued reading "pleasure" on speaker cables...

Roger Russell's excellent "Speaker Wire: A History" review.

Audioholics' big speaker cable "faceoff" with measurements from 2004... Notice AudioQuest, Transparent Cable, Zu Cable declined to have their cables submitted for measurements and comparison despite the cost they demand.

Audioholics' take on the myth of Cable Resonance. Dielectric Absorption in Cables Debunked.

Nice review - The Real Science of Speaker Cables (2016).

Audio Society of Minnesota Listening Test published in Stereophile.
Notice they used 16AWG cable with low impedance electrostatic speakers - why? 12AWG OFC cable here would have been inexpensive and realistic for audiophiles. Notice the $8000 cable did not do that well.

Great summary by Tonestack - also talks about effect of wire resistance on speaker impedance curve.

"Demystifying the effects of loudspeaker cables" from L-Acoustics, 2015 conference paper - focusing on cable loss at long lengths. Also some other details like skin effect.

An interesting DIY Braided Speaker Cable project. I'm suspicious of the capacitance measurement, but otherwise nice to see projects that include measurement results compared to Kimber 8TC (remember that years ago, I measured 8TC cables from my speaker output). Note that the SPICE filter simulation is with an 8Ω load whereas I use 4Ω above.

Great technical material on speaker cables from Elliott Sound Products with loudspeaker impedance simulation circuitry and cable performance all the way to 100MHz!

Electro-voice Cable Loss Estimator - a fun online calculator to try out!

Iconoclast Cable (Blue Jeans / Belden) - effectively 10AWG per run (24 strands, 24AWG) cable measured at 1.2mΩ, 0.08μH, 45pF per ft. Very low inductance while keeping capacitance in check is excellent. Check out the science in their Speaker Cable Design Brief. A pair of these, 10 feet, oxygen-free (electrolytic) copper (4N, 99.99% pure), terminated with banana plugs currently will set you back US$2200! Good value? That's relative...

I still get a kick out of these guys' marketing at NB Cables. Fancy looking, no BS cables with unbundled snake oil and burn-in. Reasonable priced products should be like that... (Unsolicited plug.) :-)

With this as a starting point, we might be able to do some measurements of more "exotic" speaker cables in the days ahead. In fact, I have one here which perhaps we can look at next time. Also, measuring total impedance of speaker+cable to look at the effect would not be difficult either (hinting towards what Peter Aczel did years ago which modeled the amp --> wire --> speaker system).

--------------------

This week, there was a "disturbance in the force" at Audiophile Style. No point speculating as to the whys of such a change; I'm sure it's multifactorial.

Clearly the changes (and implications) have not gone over well with a number of longtime participants.

Regardless, even if not the same going forward, the discussions at AS over the years have been important in changing the audiophile landscape. I trust these discussions will not be forgotten. Although some "subjectivists" might see this as a "win" of sorts, I don't think it means suddenly madness rules and all forms of illogical, irrational, fictional accounts of audio will suddenly spring up... These kinds of thoughts have always been present in audiophile forums including AS regardless of the rebuttals of objective-leaning participants. I remain hopeful that on the whole, audiophiles will appreciate a "more objective" philosophy over time and in greater numbers though education and desire to move forward as a respectable, technologically-based hobby.

As expressed by a number of objective-leaning members, by creating a specific "Objective-Fi" subforum and the uncertain prospect of confrontational comments potentially moved into that "ring-fenced" space, the impression is that the site no longer values objective perspectives. I certainly cannot fault those who feel this way since I think there should also be a "Subjective-Fi" subforum to truly show balance - this would be the place where the most "subjective-only" threads go; a place where anyone can say anything and never be challenged by any form of external reality-testing!

While over the years, I've often appreciated Chris' perspective, there is something in his "State of the Union" post I believe is highly incorrect:
"Because audio is an inherently subjective pursuit, it makes no sense to create a subjective-Fi sub-forum equivalent to the Objective-Fi sub-forum either."
Music, the emotions it elicits, and preferences we have about it are certainly inherently subjective. But audio, as in audio hardware that audiophiles are passionate about, is not "inherently" subjective and IMO very much the opposite! The mechanisms by which audio hardware operate are based on physics independent of the mind and its perceptions, invented by scientists, executed by engineers even if products are infused with artistic elements and beautified with industrial design. I hope that for audiophiles, this fact is also inherently understood and when highly questionable "subjective" claims are made, there will always be push back because that is simply honest and even necessary. If there is no push back during discussions veering into fiction/delusion/snake oil, then I think this would be a sign of failure for a public forum even if hosted privately, and I trust Chris will recognize this as a descent into a likely useless "echo-chamber".

Let's face it, the audio Industry will always bias towards subjectivity because it prefers to deal in "opinions" and conclusions can always be steered towards "go listen for yourself" - which these days typically means putting money down, dealing with shipping, and the hassles of a return possibly. The only discipline asked of its consumers from the Industry is a habitual desire to upgrade. Likewise the Industry is threatened by systematic evidence because they would rather provide the information straight from their Gurus and through highly questionable whitepapers with no corrective feedback mechanism.

Yes, a subjective perspective is easier to deal with, it's less emotionally taxing. Nobody wants to be told one is wrong using factual evidence these days, right? Not just audiophilia, but it seems to me that in many areas of public discourse, having "fun", accepting claims of "lived experience" no matter how ridiculous, is more important than education, understanding, and appreciating what is truth.

Remember the outcomes of intolerant subjectivism as we've seen with Michael Lavorgna's first iteration of AudioStream. Look what happened to the once-vibrant InnerFidelity without Tyll's objective analysis.

Will participant numbers on AS actually grow? Will web traffic increase or decrease? Will discussions actually be useful and meaningful without as many points of view even if this dares to at times create conflict in a marketplace of ideas?  What is a forum without its debates and disagreements after all? We shall see how this plays out...

--------------------

I recently watched and very much enjoyed Jojo Rabbit - funny and yet I think important film for our time. Definitely one of the best movies last year. The eclectic Jojo Rabbit soundtrack is quite nice as well with a selection of pop, rock, jazz, instrumental score, classical - great to see that dynamic range wasn't squashed with an album average of DR12.

Again, remember folks to get me your blind test results before the end of April!

Have a great weekend and enjoy the music as we enter March, everyone! March will be a busy time at work and Spring Break... Might not be around as much.


Addendum: February 24, 2010
Recalibrated my meter and repeated the measurements of the AmazonBasics 14AWG cables tonight (OFC and CCA variants) with >30' lengths to get a better 1' average reading. Found that the CCA had comparatively higher resistance compared to the OFC than I first recorded using a shorter length of wire.

I think it's useful to have these AmazonBasics cable results as an example of the level of "performance" one can get very cheaply these days for comparisons.

11 comments:

  1. To what extent do you think your measurements and conclusions would extend to headphone cables? Thanks.

    ReplyDelete
    Replies
    1. Hi George,
      Yeah, to various extents this is possible, but remember that headphones are typically higher impedance devices that speakers and current transmission demands much lower than amp-speaker combos. Plus lengths of cables lower as well...

      Perhaps one day I'll get a change to look into this more :-).

      Haven't looked at the data much at all but I know InnerFidelity did have articles back in the days like this:
      https://www.innerfidelity.com/content/headphone-cable-measurements-wrap

      Delete
  2. I'm a little surprised that your resistance and inductance measurements aren't monotonic functions of frequency and that they vary as much as they do at low frequency.

    A very simplistic model suggest that, for 12AWG copper, you should get

    100 Hz: 3.09 mΩ/ft
    1 kHz: 3.10 mΩ/ft
    10 kHz: 3.45 mΩ/ft
    100 kHz: 8.53 mΩ/ft

    and, for 14AWG, you should get

    100 Hz: 4.91 mΩ/ft
    1 kHz: 4.92 mΩ/ft
    10 kHz: 5.15 mΩ/ft
    100 kHz: 11.05 mΩ/ft

    as a result of the skin effect.

    In particular, there should be hardly any frequency-dependence below a few kHz, just because the skin-depth is larger than the radius of the wire. But you found a much bigger change in resistance between 100 Hz and 1kHz than between 1 kHz and 10 KHz.

    ReplyDelete
    Replies
    1. Just to clarify where I'm coming from, I view your measurements of 12AWG and 14AWG copper zip cord as "calibration". If they measure as expected, that should boost our confidence in the measurements of your more "exotic" cables.

      My "model" is that the zip cord consists of two straight cylindrical copper conductors (of radius 1.026mm and 0.814mm respectively), separated by a sufficient distance so that we can neglect their mutual inductance (relative to their self-inductance) and (when the ends are shorted together) the capacitance of the cable.

      The formula for the complex impedance of such a conductor can be found on Wikipedia:

      https://en.wikipedia.org/wiki/Skin_effect#Impedance_of_round_wire

      In addition to the resistance per unit length quoted above, this model predicts the inductance/unit length.

      12AWG
      -----

      100 Hz: 0.191 μH/ft
      1 kHz: 0.191 μH/ft
      10 kHz: 0.180 μH/ft
      100 kHz: 0.076 μH/ft

      14AWG
      -----

      100 Hz: 0.192 μH/ft
      1 kHz: 0.191 μH/ft
      10 kHz: 0.187 μH/ft
      100 kHz: 0.096 μH/ft

      The model is (evidently) not as good at very high frequencies where (because of the skin effect) the self-inductance goes down, while the mutual inductance goes up.

      Delete
    2. Hi distler,
      Yeah, I appreciate the feedback. As I mentioned at the start of the article, there are certainly limitations to the LCR device itself and I wanted to basically give readers a taste of the kinds of values we're looking at given the very small magnitude of the numbers themselves.

      This limitation is also magnified by the lengths I'm using for the measurements - only ~10' for the DIY cable and 15' for the Amazon zip cords. For the Amazon cables, I could re-do the measurement with 30+' since I have some around but since I don't need any more DIY stuff, so long as I can verify that the AC resistance is generally as expected and inductance within the "ballpark", it's good enough.

      Anyhow, thanks again and I am curious so will see if I can at give the Amazon cables another shot with longer lengths for the machine to measure after recalibrating.

      As for the "exotic" cable, let's just say that the results are in fact rather expected!

      Delete
  3. The only time I ever heard a difference, and this is more of simply functioning cable vs. non-functional, was with a pair of Elac UB5's I was demoing.

    These are known to be "hard to drive" and I had a length of Amazon Basics copper-clad-aluminum. During the guitar strums of Flametop Green by Daniel Lanois, the speakers would crackle pretty bad. I thought the speakers were busted, until I switched to some OFC, and they played totally fine. I switch back to CCA, they crackled... back to OFC, they were fine. Since then I swore off CCA although I have no idea why they were behaving that way (yes, I even used different runs of CCA to make sure it wasn't a specific length that I was using).

    Any ideas what would cause this?

    ReplyDelete
    Replies
    1. Interesting experience Ishm!

      The only measurement of the ELAC UB5 I can find is this one:
      https://www.soundandvision.com/content/elac-uni-fi-ub5-speaker-system-review-test-bench

      They say the UB5 impedance reaches only a minimum of 4.05Ω at 132Hz which isn't terrible (with -23° phase angle). They are not efficient speakers however with sensitivity at 85dB/W/m.

      What amplifiers were you using? How long were the cables? Possibility that you were playing loudly and somehow reaching the clipping point of your amp exacerbated with the CCA cables? (Will need to have a listen to this Lanois track!)

      Delete
    2. Unfortunately this was a time I had several things going in and out, and I'm not sure what amp I was using at the time. May have been a NAD 7240pe, which I use now, or might have been any number of others. It was certainly fine in the power category and I have a small room and don't listen at crazy levels.

      The cables were probably not more than 10ft. It surprised me that this happened with higher frequency notes rather than low frequency.

      Delete
  4. How do we explain when we experience different sounds in different cables of decent or better quality?

    If there's only the material, the resistance, the inductance and the capacitance to use in trying to explain it, what do we do to avoid being seen as maniacs or snake-oil victims?

    I mean, if someone can hear a difference in the soundstage with 2 different speaker cables, and the known parameters simply isn't enough to explain it, then what to do?

    I feel there has to be more to it, as in science in this area is not complete.

    ReplyDelete
    Replies
    1. There are so many contributing factors to how we perceive what we hear. I haven't changed my setup in months but I can walk in and listen to the exact same song a day apart and it sound different. Some days I feel my system lacks punch and clarity and all of a sudden I want new speakers. The next day I am more than satisfied with my system. Without getting too deep, how someone feels at the time can influence what we perceive we hear. When I am in a good mood, my system just seems to sound 'brighter' and clearer, when I am in a sad or frustrated mood, I feel my system mellows out.

      It's also how we perceive cables. When told you're using a more expensive, exotic cable you generally hear it because you 'want' to and you start looking for things that aren't there and start justifying it in your own mind.

      The 'science' of cables hasn't changed for decades and I don't expect it ever will. Cable companies are just coming up with new ways and buzzwords to sell their product. Good ole reliable thick gauge copper isn't 'sexy' anymore and no one can justify the price tag if the seller doesn't make bold, left field and exciting claims.

      Another issue is the 'stuck up' (for lack of a better term) attitude the audio industry seems to have. If someone buys an expensive cable (for whatever reason) and someone else can't hear the difference they supposedly can, the person will simply claim their ears are more trained or the other listener is less experienced. It has created this elitist mentality that dealers seem to feed off. Go to a 'high end' audio store and watch them push the more expensive cables, tell them you can't tell the difference and they'll tell you they can as they have more experience and 'listen to systems all day every day'.

      What's also interesting is when people claim a cable 'increases the bottom end' or 'makes the system brighter'. Isn't that the opposite of what one would want in a cable? A simple EQ adjustment can do that on even the most basic systems these days. Wouldn't we want a completely neutral cable that replicates what the system (and artist) is trying to produce then adjust using EQ to satisfy personal needs.

      I could give examples and throw around theories for hours haha. Personally, I find the psychology behind cables far more interesting than the science. The science is solid and unchanged, the psychology will always vary.

      Delete
  5. How about doing measurements of the frequency response of your speakers with the various cables, or perhaps recording a song with your microphone and then change the cables and record the song again, and then show us the "difference" between the cables, just to get it over and done with and show that there is no difference, even when implemented in a real system (so not just in theory)?

    ReplyDelete