Sound Quality - Rule 5: Use Ethernet between Core and Output. Really?

It occurred to me that in this earlier post, I was only speaking of a traditional setup where we use “the Internet” to get the “files” or stream to a transport that will ultimately pass the data to a DAC.

I don’t have much experience in the situation where the transport is the DAC and is also the output device (wifi soundbar, Sonos, Alexa, etc… While I certainly have devices like these, I wouldn’t consider them high-end enough to care about them doing anything than generating music for background and not for critical listening.

I understand there are a whole new class of devices, like the Devialet speakers, that do this and are intended to be high end, critical listening devices. I’m not that old, but this seems pretty “new-fangled” to me, so I can’t speak about it, but I would hope that a $2-3k speaker like the Devialet was built to be used as a direct consumer of Internet-like data, and that it would have the buffering and reclocking necessary to remove wifi, Ethernet, etc. from the equation as a factor in sound quality.

My system sounds the same over ethernet and WiFi. I don’t buy it. WiFi is frowned upon for one reason they want to remove the most troublesome link for support queries as WiFi is a fragile thing and diagnosing someones problem if there on WiFi I don’t believe it affects sq just stability and adds more latency which can cause other issues.

1 Like

Slim Devices (Later bought by Logitech) had a high end device like that in already in 1996, the Transporter
Stereophile’s review used the wording “Is a high-end music server the audio equivalent of polishing a turd?”
Read more at https://www.stereophile.com/mediaservers/207slim/index.html#ePbFVXAOlqpYSzii.99

All the discussion around wired, wifi, WAV sounding better than FLAC, you name it, was already going on back then. :slight_smile:

Auralic reccomend a WiFi connection for their Aries in order to avoid a galvanic connection. I would reccomend an Ethernet connection for stability, but I don’t understand that it is inherently better SQ.

Interesting, but wouldn’t a hop through an ethernet switch essentially achieve galvanic isolation? I ask this because I don’t know the answer and galvanic isolation has always been an interesting topic to me.

Asking about sound quality in these forums (fora?) presents a bit of a minefield (probably as with anywhere else). First of all, there seems to be a distinction frequently made between what audio nerds would call “sound quality” and something else that encompasses more catastrophic problems. The audio nerd’s “sound quality” would include notions like dynamic agility, soundstage size/dimension, tonality, detail, imaging, etc. The other way the term “sound quality” gets used here, especially by the Roon staff, encompasses more obvious problems/issues, such as drop-outs, stuttering, etc. My guess is that the Rule 5 statement about wi-fi is probably aimed at that second use of the phrase “sound quality.” You might experience drop-outs.

It seems that your questions are more about the first sense of the term. I don’t think Roon staff will take on your questions, if that is your meaning. They don’t seem eager to have those discussions, and who can blame them? Isolating one thing, such as wi-fi or RAAT or USB or whatever, is always an all-else-being-equal endeavor. Problem is, nothing is ever equal when it comes to sound quality (in the audio nerd sense).

If I was to answer your question, I would first say that the quality of the associated equipment might be dispositive of your questions. A pair of $500 speakers, and the other equipment normally associated with such speakers, is unlikely to reveal sound quality differences between wi-fi and ethernet. How much do you have to spend on a stereo to hear that difference? I have no clue. I could not hear differences in speaker cables on a pair of $5500 speakers (purchased in 2000) until I changed to a pair of $8k speakers in 2013 (not having changed anything else in the system). Really, my first answer should be to ask about how your speakers and your listening position are placed in your listening room (unless you’re rocking headphones).

I think the best way to scratch your itch (unless your itch is solely to understand why there would be a difference) is to try it both ways in your system. If you don’t hear a difference, be thankful and enjoy, or be poorer buying equipment that can reveal the difference. If you hear a difference, you should still ask yourself which one you like best. If you hear a difference, and yet further differences going from one wireless router to another or from one ethernet switch to another, then I will congratulate you on having a sufficiently resolving system, and I will wish you luck avoiding the madness that chasing the sound quality dragon can bring down on you.

It is so much easier to have this discussion if the only thing that “sound quality” means is whether or not the music drops out or stutters or there is some other basic failure. Yeah, if the hardware is working within specs, wi-fi sound quality is “good.”

I guess there is another issue lurking here: if you can’t hear a difference, is there hardware out there that can reveal it? In just about any context, usually the answer is yes, unfortunately. That can be a source of frustration or an opportunity for Buddhist wisdom. Go to an audio show and listen to the Sygergistic Research demonstration of their tweaks. Their sound is made blatantly obvious. The stereo that makes it obvious (in a noisy hotel room) probably costs $500k or more. At least with the wi-fi/ethernet question, it is easy to determine a difference in your system without having to spend hundreds/thousands of dollars. Bulk ethernet cable, although also blamed for sound quality differences, is not crazy expensive.

I hope this was somewhat helpful to you.

6 Likes

Tonight to also put the topic of ethernet cable changing the sound I decided to capture digitally the headphone output of my Naim Uniti Atom using a pcm recorder so to compare the outputs of local USB stick playback, ethernet and Wifi.

  • Took one file and copied it to usb stick an stuck it in the Atom (Sufjan Stevens Chicago from Illinoise as its a very dynamic track with a lot going on.)

  • Connected up a professional solid state digital recorder that records 44.1 wav to the headphone out of the Atom and adjusted volume to 50% and set the record levels on the recorder ensuring there was no distortion or clipping. Listened to output all good

  • Recorded file playing back by USB x2 times as a control.

  • Played file again via USB this time without the lan cable connected and it moved well away from the Atom. Switch is at the other side of the room so I have a 5m cable

  • Played same file from NAS via uPNP via Wifi again no lan cable connected (did not use Roon for the test as its output does come out at a slightly different volume for some reason)

  • Played same file from NAS via uPNP on Lan (my normal setup)

  • All files copied to PC and edited to be all matched up to the millisecond using Adobe Premiere

  • Exported them back out as uncompressed wav so all the same duration

  • Imported in Audacity and compared waveforms at normal and highly zoomed in levels

  • A/B across tracks randomly in realtime in Audacity

So what did I find and hear when I could accurately switch between them all and compare them?

No difference that I could see whatsoever.No increased noise floor in the silent areas, no difference in peak levels, no difference in waveforms and switching between was seamless with no jarring or noticable difference it felt as one continuous track from one source. So I think that wraps it up for me. It shows at least in my system no added noise is happening in the D/A stages from ethernet and wireless is the same as ethernet and Lan. Yes my system might be adding noise all over but it still shows there is no difference between sources and with the Lan cable removed and put some distance from the audio kit. You might all say using the headphone output is not the same as using the speakers. Perhaps but you cant control the environmental noise if recording from speakers and it still uses the same DAC regardless and it seems to be the theory thats this is where the noise causes the problems. Why not try it for yourselves.

4 Likes

@Antonio_Bendezu: Yes, I don’t really have any questions surrounding ethernet cable as applying to the more ephemeral sound quality issue; it will either play, not play, or piss you off noticeably.

What’s interesting is after about 15 years solidly in this hobby, I have found enough nirvanna for myself in my big rig, that comprised of some nice stuff (see my profile), but the things that have remained constant are the Vandersteen 3a Signatures, 2x 2Wq subwoofers, and V’steen’s balanced crossover to drive the subs in the unique way they do.

The only moving parts on that system has been DACs, going from an Empirical Audio OR1 reclocker to I2S to a Emp. Aud. modded Benchmark DAC-1, to an OR5 to an Audio-Gd M7 vis HDMI I2S, ultimately to a Chord DAVE.

I remember in the recent years exclaiming to myself how good that system sounded, even before the addition of the DAVE. At no point have I had upgrade-itis on that system, however the DAC hand-me-downs were the result of upgrade-itis on my fullsized headphone rig.

Finally, I hit nirvanna in my headphone rig after switching from dynamics (every Audeze flagship since the LCD-2) when I went with a pair of Stax SR-009s and the most expensive energiser that Mjolnir Audio made.

For me, nirvanna involved in being in the multiple 10s of thousands, but fortunately, I am happy enough that the itis has left me.

For that, I’m thankful. :slight_smile:

This is what I was trying to say, albeit very clunkily. (I’ll blame it on recent surgery :slight_smile: )

I’d say that if tweaks and cables and software had any effect, it would be on the wire between the transport and the DAC (and not TCP/IP, wifi, or ethernet wire). I say that with a caveat of “personal opinion” because I got dinged pretty badly on Hydrogen Audio forums 14 years ago for violating their ToS and saying something sounded or was better without empirically qualifying it, unlike your excellent analysis post that I am replying to.

The important thing is getting the wired and wireless segments in your network stable and reliable. The downside with wifi is that if you live in a densely populated area, there of potential for interference. The downside of wired is the potential for noise and interference from damaged / improperly functioning or poorly engineered network hardware, power supplies or wiring.

The best thing I’ve done to improve the stability and reliability of my home network is replacing all the questionable wiring, replacing consumer-grade hardware (router, WAPs, etc.) with gear from Ubiquiti Networks and putting all the critical components on UPS power. Did it have an effect on sound quality? Certainly not to my ear on either wired or wireless endpoints, but what it did do was keep things from getting flakey over time, requiring devices to be power cycled, with the added benefit of robust monitoring capabilities (e.g. up-time, packet loss, load, etc.) just in case I start getting paranoid :ghost:

That being said, I have zero interest or belief (from a technical perspective or otherwise) in ‘audiophile’ Ethernet cables, but I’m not one to try and tell someone what they may or may not hear in their own system. If you can afford it and you perceive an improvement that justifies the expense in your mind, then go nuts!

2 Likes

Hello Robert

You can try this for galvanic isolation : Network isolator, MED MI 1005

MED_MI1005_02

Nice! Cheaper and “prettier” than the device featured recently at audiostream: https://www.audiostream.com/content/barn-gigafoilv4-–-inline.

I wonder how they compare, though I suspect my system wouldn’t reveal it.

I second the gear from Ubiquiti. I use 2 of their APs at home, controller-less, and have setup a couple smaller business <10 people, networks using their gear. It’s enterprise-grade comparable to the big names but without the price tag, the downside is you have to be fairly tech savvy to setup and support is via forums but once it’s done you’re good to go.

All the tech support people at my work use ubiquiti gear at home and at work for WiFi aps. I am looking to get some at some point money permitting and replacing my router with one of theirs.

Another +1 for Ubiquiti - replacing my APs with a couple of carefully configured, standalone AC Pros has transformed my WiFi stability.

1 Like

Ubiquiti is nice, but I’ve only touched their switches and maybe a router. I worked one place that the congestion was so bad, and they were trying to run 4000 sq/ft. on a couple of home grade Best Buy router jobs, that I finally convinced them to put in 5 or 6 of the OpenMesh APs, which has a “cloud management GUI.”

It sufficiently did away with our wifi issues, especially when I got us also onto a fiber connection: https://www.openmesh.com/

1 Like

I’m using all Ubiquity/Unifi gear - my core is located in another building over 100m away with fibre connecting two switches together and most of my endpoints run via AP Lites. It works well for the Apple TV (under the AP) and Sonos Play:1 (inbetween a few walls and the AP). but the poor little Raspberry Pi:3 drops out often. It has absolutely lousy built in Wifi. It’s metres from the AP, with nothing in the way. Some times it works great, others its hopeless. Often restarting the AP will fix it up.

This is what’s being refereed too when ethernet is preferred - it will never drop out. Great wifi does have the potential to drop out. Roon incorrectly reports this as a data issue - which is confusing.

There is a new rasbperry Pi releases now with dual band wifi so it’s sure to be an improvement.

Have you tried a WiFi dongle with the RPi? I use an Edimax Nano on a couple of them and have no dropouts at all (previously was unusable).

Personally while I think it’s interesting to hear the views and scientific dive into Ethernet vs WiFi etc I would challenge anyone to tell the difference between both in a true blind test. If I ever get to the stage where I am even trying to find a difference between WiFi and Ethernet transmission of digital music then I have come to the point if life where I have become more interested in “hifi” than I have music.

5 Likes

I enjoy both music and hifi, in much the same way I used to enjoy building RC airplanes and flying them, and riding motorcycles of all kinds and maintaining them. Two facets of the same hobby, trick is to figure out the balance that works for you and enjoy.

6 Likes