Ethernet switch

Sorry for delay, I missed this bit.

Who on earth is asserting that “analog(ue) principles smoothly transit to the digital domain”. Definiitely not me. I assert that a superfically digital device can have an impact in the analogue domain; there is no domain transference, the impact of a switch is all-analogue. I’ve worked in digital/IT for decades, perhaps I should mention that.

I really thought we were done.

You have Nigel! That is precisely what you are asserting. You claim that an audiophile switch(X) used properly (Y, right spot in the chain) can produce an audible SQ improvement(Z) at the DAC output over that same chain without the switch. That’s your claim; tell me how I am wrong. It’s as if I said “this Riedel stemware actually allows the wine to open up in a discernibly better way.” Both might be true or false or a combination. To both I say, “let’s check!”

This is a red herring, and a deflection. I am asking for no such thing. I’ve been a lawyer for over forty years with a lot of courtroom experience. The respective burdens of proof in criminal and civil law exist to test whether a case has sufficient merit to be heard. Juries (and judges) in each have different standards of proof for assessing the evidence they have heard. “Beyond a reasonable doubt” (criminal) is not remotely what I am asking for, and neither is “preponderance of the evidence” (civil). I have repeatedly said “give me something to go on that suggests something is really happening here other than your cognitive bias,” or in the alternative, some blind testing that shows you can determine the difference more accurately than a coin flip. Show me one audiophile switch maker that can show any difference at the DAC output that might be the explanation for what you are hearing. Show me one review with a rational explanation or hypothesis for what might be happening, and I’ll be all eyes and ears. But you haven’t.

Forget about whether I would find it convincing; you haven’t offered anything.

I’ve recounted my crude blind testing, and I have offered the measurements of others. You’ve really responded to neither except for this:

Well, there is another explanation staring you in the face: a properly designed DAC (or Upsampler) will reject that noise even if it’s on the line just before the Ethernet port and it will never make it to the DAC output.

But that’s it. However unreasonable you see my “evidentiary demand,” you’ve offered nothing other than subjective impressions. Please correct me if I am wrong.

This was a harken to the point made by @Anupc that audiophiles want to bring their experience with legitimate differences produced by cables and such in the analog domain over to the digital domain. That’s all. Poor choice of wording on my part, no need to exaggerate what I said.

You seem like a really nice guy, and no, that’s neither patronizing nor condescending. We could probably have a lot of fun with this over beers or tea or Pinot. But I think we’ve taken it as far as ir can productively go—and probably a bit farther—in this forum.

1 Like

We can be, whenever you want to be. I did suggest earlier we were destined never to agree.

I said it can, not that it will. You’re extrapolating inaccurately and unhelpfully. You said it doesn’t make any audible difference in your system, I said it did in mine. Quits?

Yes, it was a deflection, fair cop.
I can offer you a rational explanation but already know that you will reject it as it does not confirm to your requirements for evidence (=measurement) nor does it align with your own experience. So what’s the point.

That final sentence is unfair. You’ve offered your personal experience and I’ve offered mine. My hearing a difference is as valid as your not hearing one.

I have dCS DAC. The difference is clearly audible using this DAC. I believe it to be properly designed.
I don’t see you offering anything other than subjective impressions either; if I’ve missed your measurements, post me a link to remind me. I can offer you theories but you’ll reject them. Life’s too short.

I didn’t exaggerate, I interpreted.

Agreed x2. I am a nice guy and I’m sure you are too; you write well and patiently but we simply have different experiences, expectations and requirements. Hey, we’re both dCS fanboys, right?

Actually, my original response to your post was specifically to this point that you made about Media Converters;

That suggested to me a lack of actual knowledge of how Media Converters and SFPs work, hence my reply.

In any case, if your only point was actually;

Then, yes. No arguments from me :+1:t2: :grin:

I think we both agree that whether it’s an “Audiophile” Switch or not, having it close to the streamer to minimise the length of an unshielded Ethernet Cable is generally a good idea.

And if that’s not possible for whatever placement reasons, then having a long unshielded Ethernet cable is easily mitigated with a simple Media Converter/SFP and good PSU close to the streamer. Which has the upside benefit of complete galvanic isolation from the rest of the upstream networking components and no downside. And this is exactly what many have done.


Thanks @anup!

Yes, I should not have said “most”. My own experience with such converters was some time ago and with different DAC etc, augmented by similar experiences reported by others. I know how they work but wasn’t fully abreast with the details of the standards to which you referred originally, so will investigate.

In fact, I bought 10m of fibre and a pair of budget converters to experiment with a couple of months ago but have yet to get around to doing so. The first thing I’ll do is confirm they meet that standard. Then I’ll listen. And then I definitely definitely won’t measure! :smiling_face:

Best wishes,


Interesting and thankfully civilized thread which highlights the crux of any discussion about sound quality.

  • Each person hears what he/she hears. The experience is what it is.
  • Whether the experience is caused by (expectation) bias or by an actual difference in the signal that is being played is difficult to say when differences are small.

I am an electrical engineer and would define myself as an objectivist. I am open to the thesis that the ear can hear differences that cannot be measured by the standard measuring suites used in the industry which are primarily based on steady state sinusoidal signal or combinations of these. I have significant doubts that we fully understand the limits of hearing and what measurements would reflect these limits.

I use a digital front end comprised of

  • Trendnet TCF-1000 Ethernet to SFP converter
  • Melco S100 switch connected to the Trendnet via multimode optical fibre. The S100 is powered by an LPS.
  • Melco N1A streamer feeding a Neukomm CDA126S USB DAC. The player port of the N1A is connected to the Rossini’s Ethernet Port

To my ears introducing each of these elements has brought an improvement. Otherwise I would not have bought the gear after listening to it. All comparisons were done sighted, opening the door to all sorts of bias such as wanting to justify the investment done after the decision was made. Some blind comparisons were done with friends, resulting in the same ranking as my own listening.

My thesis what is causing the differences some people hear when introducing some Ethernet switches is a difference in EMI in the analogue domain. There cannot be a difference in the digital domain. There is no digital effect that would cause a difference in soundstage. Snap crackle and pop yes, but not differences is tonality.

Why no equipment manufacturers has done the obvious eludes me. Putting an optical SC type socket on a DAC would do away with this whole conundrum of Ethernet effects. An SFP socket would not solve the issues, as it again introduces a foreign electrical component into the receiver. The optical / electrical conversion needs to be part of the DAC, so its quality can be controlled by the manufacturer of the device. I would expect a significant increase in consistency of system quality under different configuration scenarios using this approach.

Having said all of this, I hugely enjoy the music my APEX upgraded Rossini produces!


They have! I posted a few examples in the post to Nigel above.

If you want to take a look at one specific flagship DAC with an SFP cage built directly within the DAC: Linn Organik DAC.

As I mentioned to Nigel, SFPs operate in very precise electrical manner. The “issues” you suggest is not as you might expect.


On that, we all three agree! :beers:

Now what do we do?

1 Like

I should have thought that was obvious! Music and drink. :wink: :musical_score::beers:


It’s ok, Greg @PaleRider, I’ve found it. Phew, I thought we had finished for a moment, and was wondering what to do with my empty life…

If we can cross out “to minimise the length of an unshielded Ethernet Cable”, we definitely, universally, 100%, wholly, agree.

My point is that, if a switch is to do anything at all to minimise RFI/EMI then it should be at the very last point in the playback chain before the streamer. That’s not about having the shortest possible unshielded cable; it’s about giving the switch the opportunity to make its biggest possible difference.
In your view and in your system, that might be a big fat zero; in others, it might not. My point is that a switch can never make a bigger difference at the other (router) end of things than it can here. Never. It is there to minimise accumulated RFI/EMI whatever the sources of that might be. In an optical system, you could argue that it makes no difference at all so why have one; you could argue it makes just as much of a difference if installed at the router end (and you’d mean zero); I’m suggesting that, for the avoidance of doubt, if you’re going to use one then do so just before the streamer.