Darko on Ethernet vs. WiFi

John Darko discusses Ethernet and WiFi from a noise perspective. He basically says it depends on the design of the equipment, what it is sensitive to.

A good discussion. Not perhaps news, but it goes against the conventional wisdom which often takes the form of dogma. As I have often argued…

Added text version to the video:

I cannot watch the video right now. Are there any analog noise floor measurements to determine to which, if either topology the equipment design is more network noise sensitive? Or is the discussion limited to anecdotal observations of “blacker backgrounds,” etc.?


Darko doesn’t do measurement.

It’s anecdotal, but at least he says he can’t hear a difference between the two.

Yeah, but come on… they’re both packets over the network with no guaranteed delivery timing. Should be the same. But the poor guy has to sell advertising, so he has to keep saying something to keep drawing views.


But one is immune from mains RF and one may not be as immune as many think/thought… Not that’s there’s a free lunch with WiFi of course - it generates it’s own RF obviously and depending on the end user’s WiFi system, reliability can be an issue.

There’s no panacea with any single solution - it’s all dependant on implementation, especially at the receiver end of course. In the end we still have to listen and decide for ourselves if we’re happy with the overall implementation.

I’ve never owned any Auralic gear but their Chief Design Guru (one man band?) seems to do many things rather technically well.

1 Like

Delivery timing?
His discussion is about noise, not jitter (which does not exist on networking).

He’s right of course - the end result is very much dependant on implementation… that’s been the case for a long time, even before he existed :wink:

There’s no panacea with any single solution. Until the measurement gear becomes as sensitive as the ear/brain system, we still need to listen in the end.

But I do like learning about the potential technical mechanisms at play.

None of us here really have any expertise but if anyone has a chance to speak to the designer/s of their own gear, they may find that while they know a heck of alot (to be able to make great sounding gear that we all love, using existing measurement gear) there’s still a lot they themselves don’t know or can’t explain… yet…

Yes, but that is itself a significant statement.
Conflicts with a lot of firm convictions in these pages: network is better than USB, wired is better than WiFi, server out of the room is better than inte room, regardless of the implementation.

I would even include in that list driving a MicroRendu with the hypercapacitor LPS-1 is better than with a switching power supply. Seems plausible, a quieter PSU certainly can’t do any harm. But how much does it improve the sound? That depends on the DAC it feeds, and it may not be significant at all. I own that chain, and I have found that with the mR/LPS-1 feeding an 818v3 which feeds DSP8000SE which contain the DACs, I have not found a difference from network direct into the 818v3, or USB direct into the 818v3. And yet, these statements are anathema to many.

1 Like

Not if you consider the last part of my post for added context:

" if anyone has a chance to speak to the designer/s of their own gear, they may find that while they know a heck of alot (to be able to make great sounding gear that we all love, using existing measurement gear) there’s still a lot they themselves don’t know or can’t explain… yet…"

But if you haven’t heard a difference does that mean the difference doesn’t exist? Someone else will not hear it? I’m not making claims, I’m only asking questions, respectfully.

Before the Darko video existed I posted this:

So you claim to hear no differences upstream of your DSP speakers but I claim I did - who is correct? I haven’t got a clue to be honest. We can only report what we observe and shake hands and head to the pub to share a beer hopefully :grin:

If/when the measurement gear becomes as sensitive as the ear/brain system one day, there may be a lot less arguing right?.. who am I kidding :grin:

But as I said I do like the discussion. On that…

Here you mention a ‘quieter PSU’ but are you only talking about Vdc output ripple?

What about leakage current measurements between the PSU’s and RF filtering differences between the PSU’s? Do these have an effect?

Are there truely RF immune DACs / analogue components out there? If so, which ones? And I’m not talking about the marketing material.

It may go deeper than just ‘quieter’…

How much someone cares about this is another issue. I know you like to discuss these things (since you posted the topic) and I do too but not everyone does and that’s ok.

I say all that as I’m enjoying the music (in case anyone asks) :grin:

1 Like

Uh, no. You are overstating human hearing sensitivity and underestimating measurement capability. Frequency domain analysis oft extends below -150 dBFS. At anything approaching normal playback levels, you cannot hear levels that low.


1 Like

Is that the only analysis that the designers of your own gear consider? That’s all that matters?

It would be interesting to hear if the designers of your own gear agreed with you (or not).

1 Like

I use wired because it’s less problematic from a networking point of view, but would agree that WiFi has better theoretical isolation. Ethernet magnetics can have different performance characteristics for common mode EMI rejection. That is not to say, of course, that any difference would be audible.

I know wired is no different to wireless at least in my setup as I have tested tirelessly and recorded the output of both digitatly and compared the audio waveforms and listened to the outputs I also compare then to local usb stick playback too as that’s supposed to be Rolls Royce of Digi playback. Exactly the same. I have changed routers, cables nothing changed. The only disadvantage of wireless is general stability.

1 Like

Anecdotal but I had endless problems feeding my CXN via wireless, I had eventually was forced to run a cable , not a blip since

Case Proven , I measured it by doewntime and level of cursing :mask:

1 Like

Ha, the swear jar got full in record time Mike?

We’ve all been there at some point with networking woes.

I’m close to a new Router with the contents !


I am going to disagree and here is why: It’s clear that we aren’t measuring everything that is important. If we were, then the equipment with the best specs would ALWAYS sound the best. It doesn’t so there aspects of equipment that we are not measuring or we are not understand the measurements we are getting.

As William Bruce Cameron said:

“Not everything that counts can be counted, and not everything that can be counted counts.”

1 Like

Darko added a written version to the original video version:

Tell you a secret: when I wrote Down-to-earth recommendations to newcomers I was being polite. I really thought we could all use some down-to-earth talk.

Consider a garden-variety idiot, like myself. Although my main system, with a Meridian box, had both network and USB inputs, I had another DAC/headphone amp with USB. And although I had the Core in the same room and could connect it with USB, I “knew” that network is better so I bought a MicroRendu. And then I read about the clever engineering (supercaps!) going into the LPS-1, and I “knew” how important power supply decrapification is, so I bought it. I never tested whether I had a problem with sound quality — no need, I “knew” these were important improvements. So I spent $1000 and now have three extra boxes (switching PSU, LPS-1, mR) feeding the DAC.

Somebody should save me from myself. And perhaps others too.

You ask, if I can’t hear a difference, can others? Does it exist? What do I care? Before I spend $1000 I should investigate if it is needed or useful in my system, not be taken in by shiny boxes and experts.

I have always done a lot of evaluation of speakers, and cars, before I buy. But not for electronics. And especially not for decrapifiers. Why? Because the allure of something for (nearly) nothing is irresistible?