No need to repeat my name in such a condescending manner!
I’m not debating the authority of the article. Rather, the fact that this has nothing whatsoever to do with LAN isolators or the erroneous association you have made. LAN isolators provide no benefit to the listening experience. Indeed noise in the digital domain really isn’t an issue until conversion to analogue, which is handled by the DAC. Buy a good quality DAC and these problems aren’t an issue.
Really funny. My local dealer sells snake oil for a living. Should I also trust his advcie, because he does it for a living? Really poor argument.
Thank you Martin for fighting for the scientific truth.
Happy to have found a kindred spirit in this never-ending argument!
I keep googling this “high impedance noise” thing as it applies to Ethernet expecting to come up with some IEEE reference or, at the very least, a manufacturer of actual networking hardware (the real stuff, not the audiophile stuff) and all I get is a bunch of jargon on audio forums or links to companies selling audiophile isolation devices.
Stray currents are a problem when POE is involved, but there are solutions to those issues through sound engineering. There are also some specific use cases (typically medical) where stray current is the difference between life and death but those are all centered around the potential for accidental contact between line voltage and Ethernet. Granted, the 802.3 spec should take care of those cases, but one can never be too careful. Regardless, in audio we aren’t dealing with life and death (no, really we aren’t) so I’m inclined to look to the the army of engineers employed by the $NNN billion dollar networking industry, which is the foundation for our modern lives, than seek advice from jargon-fueled forum posts.
On the other hand, if the audio networking guys are right then I have serious grounds for disputing every charge on my credit card statement. After all, if my bank isn’t using proper isolation and regeneration then how can I be assured that their processing is bit perfect???
Go ahead and be flippant. Don’t confuse data integrity and noise. The data can get the to the DAC bit perfect yet have a lot of noise come along with them and effect the what comes out of the analog output stage of the DAC. If what you suggest with your comment were true, we would not care that Ethernet had isolation transformers and DACs would not need to do any isolation of their inputs whatsoever.
Right, WE do not care
Well, you should!
Funny that you eulogize one guy for his electronics knowledge then ignore one who works for DCS one of the top audio brands in existence.
I think you’re getting confused about who’s confused!
Let’s spell things out:
The so-called tweaks are nothing but a go-faster stripe*. They may make someone happy, they may even deceive the those easily persuaded, but in the end they offer no measurable improvement.
*go-faster stripe in British. noun. informal. a decorative line, intended to be suggestive of high speed, on the bodywork of a car.
Thanks for the translation. Before, you finally cleared up what the meaning of “trout” was. I looked up it in the Urban Dictionary, but there was no conceivably germane entry. Go figure. Need a Brit dictionary.
The choice to use transformers was to mitigate the potentially significant “ground offset” between Ethernet nodes. Not for safety. Once transformers were decided upon, it was simple to build in to the spec that those transformers needed to block x amount of voltage.
Of course. But, according to you, that noise does not matter. You say it “3” noise in the digital domain noise “does not matter” yet you call out noise rejection as a positive for Ethernet which is in the digital domain. So which is it? Either it matters or it doesn’t.
Noise in the digital domain is an issue because it effects the analog output of a DAC. No one said that the bits are not “faithfully transmitted and received”.
That is a nice story but “deal with” is certainly variable and almost never complete. Typically, the less noise that is passed to the digital input of a DAC, the better the analog output of a DAC.
Can you point to any part of the ethernet signal that is not noise?
All of the digital domain signal is ‘noise’. All of the data is ‘noise’
From the analog perspective - every single part of the ‘good’ ethernet signal is noise if any gets transferred to the analog sections
‘High End Audio’ people seem to believe that the ethernet signal is stripped away from the analog leaving noise and the analog signal behind - at least that would be the only way most of the justifications for magic cabling / devices seem to work.
‘Buffer’ seems to have been an uninvented concept as well… (even ignoring that each switch port is doing this with every packet as well - unless you have cut-through)
In my post I summarised points previously made to clear up some confusion or misunderstanding.
Ethernet (the type we’re taking about) uses twisted pairs aka differential pairs (or balanced) vis-a-vis single-ended. The principle reason magnetics are used is isolation, that is to protect hardware from faults in mains wiring and devices.
Common-mode rejection means that noise does not affect the differential signal. Similarly, noise which is present within the digital domain has no impact on delivering data bit perfect.
Ultimately, Ethernet works because it is well engineered. There is no need for “audiophile” cables and isolators. @AE67 is correct, we don’t need to be concerned with the internals of Ethernet to enjoy music.
Well engineered DACs are designed to operate in the environment we use them. They filter RF noise from the digital signal. Some do this better than others, but the answer to that problem–if indeed you can perceive it–is to invest in a better DAC not snake-oil solutions based on pseudo-science and FUD.
Thankfully the recording studios where this source is created are fully wired with cabling made from the tears of angels and never ever compressed, filtered or buggered about with…
No. The isolation transformers block a certain amount amount of common-mode noise and probably a smaller amount of normal-mode noise. The transformers don’t block all common-mode noise and certainly don’t block all noise. Again look the link I provided with John Swenson showing the noise graphs of what gets pass through Ethernet.
Again, for the 57th time, no one is stating this noise is causing bits to get flipped. The noise IS getting passed to the receiving PHY. The less noise the receiving PHY gets, the cleaner its output.
Otherwise known as the “get a life” law 
Sorry, too much noise is interfering with my enjoyment of music.
For the 58th time, that is completely wrong! Please stop this misinformation campaign!
No, it is not wrong…but hey, you believe what you want.
In my case a tried a pair of TPLink converters and the results were not satisfactory, losing articulation in the bass and resulting in a music scene devoid of air, so I decided not to use them.