Slashdot Mirror
$340 Audiophile Ethernet Cable Tested
An anonymous reader writes: Ars Technica has posted a series of articles attempting to verify whether there's any difference between a $340 "audiophile" Ethernet cable and a $2.50 generic one. In addition to doing a quick teardown, they took the cables to Las Vegas and asked a bunch of test subjects to evaluate the cables in a blind test. Surprise, surprise: the expensive cables weren't any better. The subjects weren't even asked to say which one was better, just whether they could tell there was a difference. But for the sake of completeness, Ars also passed the cables through a battery of electrical tests. The expensive cable met specs — barely, in some cases — while the cheap one didn't. The cheap one passed data, but with a ton of noise. "And listeners still failed to hear any difference."
Okay, digital data is supposed to be easy 1 and 0 communication. But when you get down to the physical media, said binary digits are represented by physical phenomenon. So +3.3V = 1, 0V = 0 type stuff.
Voltage, resistance, EM waves, magnetics, etc... You're actually back in the world of Analogue, and here you have to worry about noise.
When you're moving data as fast as you can, or storing it as densely as you can, interference becomes more likely. For example, you'd think that +3V =1 and 0V = 0 would be easy, but when you're flipping the signal as fast as you can, you end up with the cable possibly acting like a transformer or capacitor. So the voltage might run a bit higher, a bit lower, a bit faster, a bit slower, etc...
Radio transmissions, Solar noise, close by electrical cables, other data cables with parallel runs, etc... The world is 'noisy' even if you're using wires.
That's why you have error correction in digital communications. So the 'occasional' bit can become flipped and the system transparently recovers it, and you get your transmitted data, identical from the other side.
I don't read AC A human right
that audiophile either means "wealthy" or "sucker" or maybe both...
I get that you want the best possible sound... and in some cases the placebo effect may actually help you enjoy your music more... but are there really enough of these people to base a business on?
I suppose, if you don't have to do anything except throw some gold plating on a connector and you are already in the cable business.. why wouldn't you?
This reminds me of another product, I think I saw on /, a while back... it was just a little plastic riser that kept your cables elevated off the floor and separated from each other... the cost was something like $100 per "device"...
Well... whatever floats your boat...
My eyes reflect the stars and a smile lights up my face.
They did the listening test first, before the cable quality check. And yeah, they were really, really skeptical of the cable, but they'd had enough of the 'if you've never listened to it, how can you know?' kook crowd - they wanted a to do a real test just because people don't test these types of claims. They either write them off as ridiculous (as they nearly always are) or they buy into them completely.
So here: Objective, blind test. No difference as far as anyone they tested could reliably tell. (They did have one person in the test correctly guess which cable was which. Out of two who tried - the rest didn't even try.)
The reason why it's not there is because ethernet is supposed to be just a dead simple layer 2 multi-access data protocol. Adding that kind of shit just contributes to latency (any kind of error correction involves additional parity bits and more processing.) And yes, I'll grant you that the added data and processing for error correction is tiny, but multiply that by a billion in large scale networks and you can see where there's a problem
If you need error correction, use UDP and handle it at the application layer, that way you aren't negatively impacting every other application that doesn't need error correction.
Ethernet at the MAC layer has digital data but the wires of course are all analog. If it wasn't analog there wouldn't be a bit error rate. The point was that even with a lot of noise the digital signal was still extracted correctly. The noise on the cheap cable was very bad, falling well outside the specs, and yet it still transmitted the data. The "noise" either causes the packet to be dropped or the packet is successful. The specification is about the limit of what is allowable so that it works even in a worst case scenario for what it is designed for, and they weren't doing anything close to having difficult operating environment for ethernet.
Actually, in pro audio ethernet is used with proprietary protocols, handled by black box ASIC chips with special switches. I deal with this crap in the studio. Where I am they use it mostly for the personal mixers providing monitor outputs, but some places use it for inputs too.
Also, warning for those about to embarrass themselves making fun of "directional" cables, that means it is grounded at one end, and you put all the grounded ends in to the same device to avoid ground loops. If you don't know what it is, it must be brain-numbingly stupid... right? Ignorance is bliss.
We use normal cables, sure. But we do buy expensive ones with nice plug shielding, because musicians may or may not even be sober at work. Expensive cables isn't just for fancy looks, that build quality can make a real difference. Plus, it might be a lot easier to get the bean counters to agree to buy premium cables, than to get them to agree to replace equipment. They might tell you, "use the spares until they fail" in which case you'll regret not having spent their money on the "over"-priced ones.
Same with instrument cables. No, a brand new expensive guitar cable does not sound better than a cheap one. But after 300 shows, the cheap one craps out during a show or session, and the expensive "hifi" one didn't because it has premium long life rubber and better plug strength. So it does actually sound better once you factor in the way it sounds when equipment fails and you can't even hear the instrument. Most of that benefit is in the middle price range, of course. A $1000 guitar cord probably has metal mesh "shielding" that substantially increases cable strength, but the $300 one already has plastic mesh that will provide more than enough abrasion protection.
Also... some commenters don't know this, apparently, but a "placebo effect" is a real effect. It doesn't mean it is a scam, it means the people were successfully tricked into getting healthier faster, or in this case, to have more fun. If you "trick" them into thinking the artist is more artsy, they might enjoy it more too. Pretty snooty to claim they're not really enjoying the subjective aspect of their choices as much as they claim to... especially if you're also claiming that due to the placebo effect they really are enjoying it more!
Gigabit (only BaseT, not fiber) ethernet has error correction inherent in the trellis coding used at the phy layer.
10GBaseT Ethernet (802.3an) uses a [2048,1732]2 LDPC forward error correction scheme. Fiber base GigE typically does not, except alternate encapsulated carriers like SDH VCAT or OTN (G.709).
G.975 for 10GBaseT over submarine cables uses a Reed-Solomon code with a 7% overhead. 10G WAN Phy has no FEC, 10G on G.709 OTN has FEC inherited from G.709.
In general 40G and 100G systems have Reed Solomon error correction on 25GBaud carriers, and not on 10GBaud carriers. Of course on G.709 there is FEC in the transport.
So really it's a mixed bag, but to say ethernet has no error correction is a false-hood, especially if you are talking about the most common layman GigE. It's likely in future that all higher baudrate fundamental carriers will have RS-FEC builtin, as they are running into limitations of the physical optical channel that can be mitigated with FEC.
Except that the $300 cable isn't grounded on either end, and shows a high level of crosstalk. So the arrows on this cable are just to make people think it's worth the $300.
This cable is better than a $2 cable: It's well built, and meets the specs - barely. But you can get $10-$20 cables that are as well built, and meet the specs with less margin for error (these literally tested as 'within the specs' by less than the margin of error on the testing device) easily.
'Sensible' is a curse word.