'Audiophile' Ethernet Cables.

[Paul R]

In my system I was very diappointed in a wireless bridge. Then I found the existence of MoCa, where you transport network over a coax cable. You need two MoCa adapters and some extra coax cables. Assuming you have your tv set near your hifi, it's a great way to have reliable wired network, without putting high frequencies on your mains. It's strange that you don't read more about MoCa on audio fora.

 

MoCa or optical can work great.

 

With MoCa, one has to be sure than any internet signal from your cable company is not being transmitted on the segments of cable you are using in your house, and that there are no signal filters between you either. With optical of course, you have to pull optical fibre cable.

[plissken]

So, you are saying WLAN is superior to Powerlan. Convenience wise for sure, but what about the quality of the signal? Any explanation why WLAN > Powerlan?

I'm not trying to be irritating, just wanna learn as much as possible to make an informed decision.

 

Thanks,

Pumpe19

WLAN is technically superior to wired LAN in the instances of Ground isolation and radiated EMI. It gets you most of the benefits of optical without the hassle also.

 

I have multiple WAP's (wireless access points). One of these is dedicated to my 2.0 system. It has it's own SSD and on it's own channel.

 

Dlink, Asus, Netgear for the affordable options or go enterprise wireless. Ubiquiti makes a really great line of affordable enterprise class AP's.

[esldude]

WLAN is technically superior to wired LAN in the instances of Ground isolation and radiated EMI. It gets you most of the benefits of optical without the hassle also.

 

I have multiple WAP's (wireless access points). One of these is dedicated to my 2.0 system. It has it's own SSD and on it's own channel.

 

Dlink, Asus, Netgear for the affordable options or go enterprise wireless. Ubiquiti makes a really great line of affordable enterprise class AP's.

 

The Ubiquiti gear is highly recommended. Not quite as user friendly to setup though not too terrible. Very good gear however while costing little more than regular consumer gear.

[plissken]

Says I. I'm using my ears just as you wanted me to do and still you keep on talking about measurements. You're not reading what I'm posting. That's fine.

We're off topic, so please move your discussion elsewhere.

 

Sent from my HTC One_M8 using Computer Audiophile mobile app

 

Here's the thing: If you know what is in the chain you didn't do just ears evaluation.

 

So here is a primary question that I have: Do you think that during playback that when the cable is unplugged you could raise a hand and lower it again when the Ethernet cable is plugged back in? This is all with the understanding that playback isn't going to be interrupted.

[Samuel T Cogley]

@plissken,

 

I'm quite intrigued by the idea that you can insert and remove an Ethernet cable and not cause an audible interruption in the audio stream. I imagine this might be something like LACP, Spanning Tree, or some combination.

 

Can you give some technical detail on the test setup?

 

Thanks in advance!

[plissken]

@plissken,

 

I'm quite intrigued by the idea that you can insert and remove an Ethernet cable and not cause an audible interruption in the audio stream. I imagine this might be something like LACP, Spanning Tree, or some combination.

 

Can you give some technical detail on the test setup?

 

Thanks in advance!

 

LACP is a protocol to bridge LAGs automatically. LAGs are either dynamic or static. LACP, in layman terms is the dynamic portion.

 

Spanning Tree is a split horizon algorithm @ layer 2 of the OSI model (for Switches) that prevent broadcast storms by establishing a single switch as the root bridge (think of it as the master switch) and then designating a single root port on any other given switch that will flow traffic and shutting down other interfaces to prevent looping.

 

Those other ports will still be in a learn/listen state and if the designated root port goes down then switches will bring up the redundant links.

[plissken]

The test setup is an SG-200 Cisco Layer 3 Managed switch. Computer acting as a server, a client machine with two Intel server class NIC's in an Active / Active team.

 

Playback software is JRiver and would be bit streamed. Buffers can be dialed in in JRiver including reserving 1GB of system RAM. So you could copy over an entire album on GBe Ethernet in ~8 seconds, pull the plug, and listen for the next 50 minutes.

 

My point being is that after that transfer, even if the cable is left plugged in, the OS is going to power down the NIC:

 

 

All these reviewers of Audiophile Ethernet have been evaluating them and they have most likely been shut down by the OS 95% off the time. It's the very nature of Ethernet.

[plissken]

Without wishing to create additional turmoil, I do have one question for plissken:

 

Do you have any personal experience with Ethernet cables like the ones manufactured by AudioQuest or Wireworld?

 

 

AQ and Chord.

 

ArsTechnica did a tear down of the AQ Vodka RJE. The terminations are of poor quality and the high Cross Talk #'s support the photos.

 

Also there is no center X spline to keep the pair lay cross talk minimized. Everything else is your average run of the mill CAT6 STP cable.

 

On the subject of STP cabling: Shielded networks are end to end implementations. Introducing a cable with shield actually tied the the 8P8C could introduce noise and you wouldn't necessarily need any music playing to hear it.

 

So the AQ could actually degrade your setup IF they've attached a drain wire to 8P8C shield.

 

If the shield is left floating @ both ends you should be fine.

[Mercman]

AQ and Chord.

 

Thanks!

[Samuel T Cogley]

The test setup is an SG-200 Cisco Layer 3 Managed switch. Computer acting as a server, a client machine with two Intel server class NIC's in an Active / Active team.

 

Playback software is JRiver and would be bit streamed. Buffers can be dialed in in JRiver including reserving 1GB of system RAM. So you could copy over an entire album on GBe Ethernet in ~8 seconds, pull the plug, and listen for the next 50 minutes.

 

My point being is that after that transfer, even if the cable is left plugged in, the OS is going to power down the NIC:

 

[ATTACH=CONFIG]32450[/ATTACH]

 

All these reviewers of Audiophile Ethernet have been evaluating them and they have most likely been shut down by the OS 95% off the time. It's the very nature of Ethernet.

 

With these large buffer sizes, how can a tester (or the test administrator) be certain when the data stream has moved from one cable to the other and back?

 

EDIT: Sorry, I think I get it now. Are you demonstrating that the Ethernet cable is moot once the file is transferred to the client's buffer?

[Ralf11]

I thought everything would be Optonics by now. What happened?

 

Also, where is my flying car?

[plissken]

With these large buffer sizes, how can a tester (or the test administrator) be certain when the data stream has moved from one cable to the other and back?

 

You hit the nail on the head and it's the entire point.

 

You are playing back out of buffer. The audio you are listening to could have been delivered 10, 20, 30 seconds prior to the buffer and the NIC is actually in a power minimal state.

 

How can the reviewer have all this prose about a cable, that is in most likely hood, sitting on an interface that is made nominal by the OS? While they experience all these sound stage/depth/blossoming/vivid color improvements.

 

So here is another way to observe: Anyone can get a copy of JRiver and use the 1GB buffer option. Wait for playback to start and have someone plug/unplug the cable. You record the minute/second this event happened.

 

You should be able to go and get the cheapest cable from Amazon and hear the difference. The reason for this is once you have pulled the Ethernet cable you are no longer in a "Mixed Signal System" as far as the network is concerned. If the cable is indeed wrecking things you'll know.

 

Again I've laid out the conditions that I would be willing to concede I'm wrong. All I have seen is audiophile mantra of "I trust my ears" and they really don't.

[plissken]

I thought everything would be Optonics by now. What happened?

 

Also, where is my flying car?

 

On the car thing ask Elon.

[mansr]

On the car thing ask Elon.

All he's got to do is merge Tesla and SpaceX.

[plissken]

All he's got to do is merge Tesla and SpaceX.

 

They have a manned module, that still needs approval, and the X has wings. Put it together!

[Ralf11]

you can't just Put it together!

 

the system has to be carefully integrated so each component works with the others

 

and no testing allowed!

[Jud]

EDIT: Sorry, I think I get it now. Are you demonstrating that the Ethernet cable is moot once the file is transferred to the client's buffer?

 

 

That is the idea.

 

Also, Ethernet runs through transformers that ought to get rid of any noise.

 

On the other hand: I had a circuit in my system that acted like a "noise sniffer." These things are not subtle. To give you an example:

 

 

So we're talking about very plainly evident noise. And I heard it from my system with this circuit in there. There were several measures I took to try to eliminate the noise, including simplifying system layout and de-spaghetti-fying my cables (signal and power). Since I had an Ethernet cable in my system (Blue Jeans Cat6a), even though I thought it was unlikely noise could be propagating through it, I decided to try a ~$60 used Ethernet isolation transformer. (They're made for medical installations - the one I bought is sold by a company named Baaske.) *It very audibly reduced the noise* (though it didn't eliminate it).

 

I eventually completely eliminated the noise by other (inexpensive) means - a small $96 isolation transformer for the SMPS power supply of my laptop to plug into.

 

What this says to me is that it isn't the Ethernet signal that carries some sort of distortion from the source. Rather, it's just the electrical connection made by the cable that carries noise which can be propagated through the system from anywhere (such as, in my case, a "wall wart" laptop recharger). Though I was quite dubious that such noise could make its way through the isolation built into Ethernet, when amplified by the "noise sniffer" it was incontrovertibly there, and incontrovertibly reduced by the cheap little Ethernet isolation transformer. And it was eliminated entirely by a small isolation transformer on the power, not the signal, side of the system.

 

Now would "audiophile" Ethernet cables have stopped the noise? Don't know and don't care - I'm not spending my money on one. As so often in these endless debates, both sides are so busy shouting at each other they can't be bothered to do any research into what's actually happening and what (simply and cheaply) might be done to stop it.

[plissken]

Now would "audiophile" Ethernet cables have stopped the noise? Don't know and don't care - I'm not spending my money on one. As so often in these endless debates, both sides are so busy shouting at each other they can't be bothered to do any research into what's actually happening and what (simply and cheaply) might be done to stop it.

 

Cool box Jud. If there was noise in a system I would investigate. Nothing is sacred when trying to rid the systems of demons.

 

Most likely the noise was making it through the PSU of the switch/router where the CMNR nature of CAT5/6 won't reject it because it's not an induced field.

 

Some of my approaches are to determine just such items as the Ethernet leg being a path for noise. I'm not disagreeing with that. But I can tell you that noise is going to be there with another Ethernet cable.

 

It's why I'm a proponent of Wireless. Knowing you I'm sure you pulled the plug of both Ethernet and the laptop PSU during playback to hear the hash instantly disappear

[Samuel T Cogley]

You hit the nail on the head and it's the entire point.

 

You are playing back out of buffer. The audio you are listening to could have been delivered 10, 20, 30 seconds prior to the buffer and the NIC is actually in a power minimal state.

 

How can the reviewer have all this prose about a cable, that is in most likely hood, sitting on an interface that is made nominal by the OS? While they experience all these sound stage/depth/blossoming/vivid color improvements.

 

So here is another way to observe: Anyone can get a copy of JRiver and use the 1GB buffer option. Wait for playback to start and have someone plug/unplug the cable. You record the minute/second this event happened.

 

You should be able to go and get the cheapest cable from Amazon and hear the difference. The reason for this is once you have pulled the Ethernet cable you are no longer in a "Mixed Signal System" as far as the network is concerned. If the cable is indeed wrecking things you'll know.

 

Again I've laid out the conditions that I would be willing to concede I'm wrong. All I have seen is audiophile mantra of "I trust my ears" and they really don't.

 

My favorite is, "it's not possible to measure what I'm utterly certain I'm hearing", which is kind of the textbook definition of delusion.

[plissken]

My favorite is, "it's not possible to measure what I'm utterly certain I'm hearing", which is kind of the textbook definition of delusion.

 

The issue is the complete dismissal or not knowing of the explanations that Siemons and T.I. give in regards to unshielded Ethernet and the inherent noise rejection characteristics.

 

One issue is there are only so many ways to construct this type of cabling. The standard determines this.

 

Another issue is they trot out cable construction and it's ability to be better immune to outside noise. The question I have is: What outside noise?

 

It's bogey man construct. Because they need to actually *gasp* measure the environment. A cable can't protect from something that doesn't exist.

 

Siemons, in the paper I linked to in the other thread, considers a heavy industrial environment capable of emmiting ~10v/m EMI to the cable. Even at 18v/m Siemons was seeing noise rejection up to 30Mhz. If you take into consideration knock on harmonics of some fundamental above 30Mhz its going to be nothing by the time it gets to 20Khz.

 

Shielded cabling, if implemented correctly, can provide, again according to Siemons, 100-1000X the noise immunity. All with bog standard CAT6 STP. But in a home environment there is most likely nothing to protect against.

 

Now if you have noise making into the network switch via a single ended power supply, then that noise is directly on the 4 pairs of copper. The same 4 pairs that are going to be there in a $2400 Chord Ethernet cable. All the shielding and construction aren't going to do anything to resolve that type of noise.

 

This is the information I have, this is what I know. I'll welcome any counter point but 'I heard what I heard' doesn't cut it.

[r_w]

I appreciate that's a valid criticism in some cases but what about; air, soundstage width/depth/height... how can we measure that?

 

 

Surely not everyone who talks of that stuff is deluded?

 

... not having a dig BTW, I would very much like to know how to measure it.

 

 

(ought to mention that boutique Ethernet cables are not something I would ever recommend... go fibre)

 

 

My favorite is, "it's not possible to measure what I'm utterly certain I'm hearing", which is kind of the textbook definition of delusion.

[Samuel T Cogley]

I appreciate that's a valid criticism in some cases but what about; air, soundstage width/depth/height... how can we measure that?

 

 

Surely not everyone who talks of that stuff is deluded?

 

... not having a dig BTW, I would very much like to know how to measure it.

 

 

(ought to mention that boutique Ethernet cables are not something I would ever recommend... go fibre)

 

Euphonic colorations (some dislike the word "distortions", but they are equivalent) may exaggerate or soften some pleasing or irritating nuances of recorded music. Don't misunderstand, I sometimes find them quite pleasing (but also sometimes irritating) myself. It's part of the gear designer's "art" to subtly manipulate technical attributes to achieve these colorations. Depending on the gear and environment, some of these perceived nuances could be real or could be placebo or a combination of the two.

 

I see too often in audio gear discussions the unshakable belief that the existence of these subtle manipulations is an indisputable indication of "quality".

 

My point was only that an audio enthusiast (not saying you) who has no interest in addressing their own expectation biases because they believe they are immune is at best a fool, and at worst delusional.

[mansr]

Euphonic colorations (some dislike the word "distortions", but they are equivalent) may exaggerate or soften some pleasing or irritating nuances of recorded music. Don't misunderstand, I sometimes find them quite pleasing (but also sometimes irritating) myself. It's part of the gear designer's "art" to subtly manipulate technical attributes to achieve these colorations. Depending on the gear and environment, some of these perceived nuances could be real or could be placebo or a combination of the two.

 

I see too often in audio gear discussions the unshakable belief that the existence of these subtle manipulations is an indisputable indication of "quality".

 

I prefer the reproduction system as transparent as possible. Then I can apply whatever colourations I desire digitally.

[plissken]

I appreciate that's a valid criticism in some cases but what about; air, soundstage width/depth/height... how can we measure that?

 

You test to see if perception follows the reality of what cable is providing the data. But you might as well transfer a CD over with one cable to Directory A and use another cable to Directory B and then listen locally and figure out if A or B was one cable or another.

 

The terms you brought up have nothing to to with data transmission and everything to to with the final playback chain: DAC, Amp, Speakers. Source material being equal in all regards.

 

Surely not everyone who talks of that stuff is deluded?

 

... not having a dig BTW, I would very much like to know how to measure it.

 

 

(ought to mention that boutique Ethernet cables are not something I would ever recommend... go fibre)

 

The litmus test is simple. Start playback and have someone unplug the cable with out your knowledge. If you can tell when the cable was unplugged then fix your setup by either going optical, going wireless, or getting a competent DAC.

[plissken]

 

 

Surely not everyone who talks of that stuff is deluded? ... (ought to mention that boutique Ethernet cables are not something I would ever recommend... go fibre)

 

A person is deluded if they think that a true UTP CAT6 (one that passes spec) is going to sound differently then another UTP CAT6 (one that passes spec) are going to have these differences are delusional.

 

Ethernet is far and away the most highly engineered data transmission system available to us in the enthusiast audio community. Nothing is left to chance.

 

There was a thread here where someone had a shielded Supra CAT6 and it made the system noisy. This is totally doable if your entire system isn't designed to be shielded and I could fix that in the field by re-terminating the cable with out the shield tied or simply replace it with a UTP cable.