Scientific Measurements on Cables

[George 47]

A recent paper has looked not just at frequency domain differences (minimal) but time domain differences and measured differences. At least someone is going beyond the standard measurements to observe differences. https://www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web&cd=&ved=2ahUKEwig5ILFmtv5AhWMhFwKHTUpDKUQFnoECAYQAQ&url=http://boson.physics.sc.edu/~kunchur/papers/Audibility-of-cable-pathways--Kunchur.pdf&usg=AOvVaw0DPhd-nUb2LmIZl-dADk8U

You may get a security notice, ignore it, the site is genuine.

 

[George 47]

Forlorn hope: can we keep it civil and try to criticise the paper and the science, not the messenger?

 

[Devil]

I’ve got to congratulate the author for finding a cable with such poor LCR values as cable B, I’ve never seen a cable measure so badly.. (not even freebies that measure pretty much the same as cable A).

So as always, if you design a cable with high enough capacitance/inductance to act as an audible tone control, those are probably the results you’ll get.

I’ll add that I’ve not really read the article properly (tl:dr!) but that’s the first thing I checked and seemed unusually high. However if anyone can glean anything else then please feel free to enlighten me..

 

[Devil]

Edit scrub that, that they’re comparing xlr to rca connections, that’s dependent on what they’re plugged into, more than what they’re made of. If that’s the case then it’s not really a good comparison imo..

 

[rabski]

As above. Cable A 112nH inductance. Cable B 1265nH inductance. Cable B more than three times the capacitance of cable A.

I'd be far more surprised if they sounded the same really. Single ended or balanced regardless.

 

[Devil]

But even so the circuits will be different as xlr runs at higher voltages afaik.
So you’re just not comparing like with like lol..

 

[rabski]

But even so the circuits will be different as xlr runs at higher voltages afaik.
So you’re just not comparing like with like lol..

It states that they level matched, which would negate any effect on the listening tests. I agree the electrical effects will differ, but capacitance combined with inductance will act as a filter that will not vary by voltage. If it did, speaker crossovers would not work.

 

[Paul55]

As above. Cable A 112nH inductance. Cable B 1265nH inductance. Cable B more than three times the capacitance of cable A.

I'd be far more surprised if they sounded the same really. Single ended or balanced regardless.

Cable B is four times as long as A.

 

[Paul55]

A long, long time ago I happened to have two 10m lengths of thin coax and suitable adaptors to run some tests on an M-Audio MobilePre audio interface.

I don't know where the slight upward tilt came from, or the general tendency to get brighter, the Mobile Pre wasn't an expensive interface, it had what I needed at the time, microphone inputs with phantom power.

Anyway. 1m, 10m and 20m of interconnect from out to in and measure FR. An audible difference due to cable capacitance interacting with output impedance.

Something that bothers me in the article linked in the OP is their measurements of cable noise with one end open circuit. Not sure that's a reasonable thing to do. One thing that would be interesting to see tested is an expensive cable against an equivalent plain old appropriate copper job. A pairing where a cynic would be surprised to find an audible difference, rather than where it's not that surprising.

 

[George 47]

I liked the discussion in the paper on the invalidity of quick A/B tests and why they can go wrong with A influencing B and the reverse. I agree that the measurements of Cable B (Monster cable) are unusual for audio cables but note the measured frequency response and the comment: Fig. 3 shows the measured frequency responses of the two interconnects. Although interconnect A’s response is more nearly constant, both cables have deviations that are within ±0.005 dB over the 16 Hz–22 kHz range, which are expected to be subliminal

He states he deliberately chose differing cables to maximise the potential differences in cables but the frequency response measurements showed that there should not be an audible difference, but in blind listening tests, with different people and training, differences were heard that were statistically significant.

I like the reference to psychoacoustics as well and how it influences the outcomes from people listening rather than the measurement says this, therefore that is all that matters.

I need to wade through the rest.

 

[tuga]

Does it look like a test designed to prove a point of view?
I‘ve seen several cables tests with measurements other than just frequency response and they all showed differences. Audible differences?

An LCR comparison between a Blue Jeans LC1 and a Kimber will also show measured differences by the way.

If we don’t know what measurements mean and how they correlate with listening/audibility then they’re best left alone.

 

[George 47]

Does it look like a test designed to prove a point of view?
I‘ve seen several cables tests with measurements other than just frequency response and they all showed differences. Audible differences?

An LCR comparison between a Blue Jeans LC1 and a Kimber will also show measured differences by the way.

If we don’t know what measurements mean and how they correlate with listening/audibility then they’re best left alone.

Aren't all tests? The point of view should come from the data rather than agendas.

The blind listening tests showed a statistically significant difference once the usual A/B test was changed to take account of psychoacoustics.

Isn't that the aim of the paper to try and introduce time-based measurements rather than frequency-based measurements to get a better correlation?

I do have reservations.

 

[oldius]

I was always taught that only one variable should change to make the results valid in a scientific test: in this case, only the cable should change but its length should be the same.

The question remains, I think, do cables that have matched resistance and capacitance sound different to each other? I would expect differences if either of those two variables are hugely variable.

 

[Paul55]

I liked the discussion in the paper on the invalidity of quick A/B tests and why they can go wrong with A influencing B and the reverse. I agree that the measurements of Cable B (Monster cable) are unusual for audio cables but note the measured frequency response and the comment: Fig. 3 shows the measured frequency responses of the two interconnects. Although interconnect A’s response is more nearly constant, both cables have deviations that are within ±0.005 dB over the 16 Hz–22 kHz range, which are expected to be subliminal

I don't understand their frequency response measurements. One cable causes a bass droop, the other a lift? The cables have an effect near DC? And how did they measure the parts of the balanced cable? Clearly this couldn't have been measured as installed. That all seems a bit odd. Along with the very high noise pickup reported. Anyway if the measured capacitances are correct, and the 100R output impedance assumed, there is no audible impact from cable capacitance.

I'm also uneasy about the switching arrangement at the power amp between single ended and balanced. There will be a balanced/single ended (or vice versa) conversion going on which may or may not be transparent. It certainly needs to be proven so.

It would have been interesting if they tried a range of cables, even if not via exhaustive testing. A basic XLR-XLR for example.

As soon as you start look hard at this kind of thing it becomes extremely difficult to avoid confounding factors. Better just to put a record on.

 

[tuga]

Amateurs, I say. Nothing to see