[castleofargh]

i made similar expierences for ethernet as other audiphiles, tho i believe that its not the data itself or the internet wouldnt work as it does

i havent spent "big bucks" on switches/ethernet cables but i expierenced how the inbuilt wlan of the raspberry pi makes things worse, disabling the bt/wlan module improves sound quality and i was using a cheap d-link wlan stick for a while, i switched to a fritz.box wlan stick recently and it improved sound quality further
(also interesting to note is that i hear improvements if i underclock the raspberry pi)
the reason i use wlan is, that i think this kind of galvanic isolation is hard to achive with ethernet cables (atleast at a comparable price point) tho maybe i change my mind once i tried a good ethernet setup since the wlan stick itself can introduce noise, and maybe the fact it runs on the ghz range makes things dramatical worse

since i dont dismiss all science etc (i just think some of it is wrong, like it hasnt be discovered in "blind tests" yet, so science people dont care about it) the most logical thing (for me) is actual noise which somehow influences other stuff, even if it cant be measured on the dac output
i come from a standpoint where i hear differences first and try to find a logical explanation to kinda try to make it a "whole picture" to go from there

but the best source i came across so far which kinda line up with my current "audiophile" believes (which might be interesting to some) is the website&youtube channel "Alpha Audio"
they "kinda" blind tested (tho you guys will critique this) ethernet switches and also measured them and there was kinda big differences, for example power supply noise distributed to the ethernet ports, port to port noise isolation etc (which was surprising for me since the common explanation includes that ethernet ports are completely isolated)

https://alpha-audio.net/review/a-very-deep-dive-into-network-switches-listening-and-measuring/2/
( HOW AND WHY we test network switches... )
( Live BLIND test switches - do switches matter for audio quality? )

the blind tests on youtube where you can listen yourself is a really interesting approach but micing (tho they do a good job i think compared to other youtube videos like this) and specially youtube makes it very hard to hear differences, tho i heared some differences myself but they seem very minor in comparision if you test this yourself

Alpha Audio is actually the only really good channel i found which try to combine objective and subjective matters (which is the approach we need imo)
kinda wondering why i never found measurements like this but glad they did it, and i think they are genuine and dont try to push audiophile stuff just because the sake of it
the tests also show that they dont think the most expensive solution is the best... and also measurements show this

I feel bad criticizing the few people who put in the efforts to get somewhere, but throughout the 2 videos, I cannot help but notice inconsistencies within the guy's own rational.
We start with the assumption that maybe noise and stuff in the switch bleed into the rest of the playback chain and impact the output signal. As a hypothesis it's one I surely would consider myself, I would look for other stuff too but it's one of the potential causes that might deserve checking out.
But Then why only measure the switches? How is that ever going to tell us anything about how some DAC or streamer after it is being affected?


The blind test is... well, as the other guy said, not scientific. And not that much of a blind test when the first dude reacts and even comments before the second one has finished describing his impression.
Also the guy on the left said that it's fun. I'd add him on some government watch list just for daring say something like that about blind testing. Informative, eye opening, mostly disappointing, often a torture, I would fully understand saying that about blind testing. But fun? Never cross that guy! He's the danger.

 

[Ghoostknight]

And not that much of a blind test when the first dude reacts and even comments before the second one has finished describing his impression.

yes i also didnt liked this as a first saw it, tho i think he doesnt do this everytime and both take notes which often kinda line up, they usually go over them in the end

overall yes its not a scientific blind test but i really like the effort and maybe they do it this way (not ABing) because they found its more effective (i usually test myself in a somewhat similar manner, tho i dont compare 8 devices at once )
(for me) they do a great service for the community and they are the only ones that even do it in this quality (even if it could be arguably improved in some things) and i dont think they just do it to push audiophile claims, i dont see the incentive unless they get paid big bucks by "audiophile" companys under the hand which i find hard to believe specially because they make some companys look kinda bad in some tests, specially the objective measurements!!

i guess saying that "its fun" wasnt really seriously meant, i understand it this way, ethernet switches are a heavly controversional topic (for a reason... it sounds ridicoulos) and they were curious themself (tho i guess they were already somewhat aware), so its "fun" to see the comparision and a somewhat conlusion, specially because cheap commercial switches were compared against audiophile ones

 

[VNandor]

Oh my god, I can't believe it, this guy plugged in his 2 prong power cord the wrong way around. This mistake completely invalidates any testing he has done!

 

[gregorio]

since i dont dismiss all science etc (i just think some of it is wrong like it hasnt be discovered in "blind tests" yet, so science people dont care about it).

Firstly, you’re making the exact same self-contradiction AGAIN! You “don’t dismiss all science” but you “just think some of it is wrong”, which IS dismissing it!

Secondly, double/blind tests are NOT used to discover if things exist, measurements, calculations, null tests and other objective means are used to discover if something exists and if it does, only then are DBTs are used, to determine if that something is audible.

Lastly, if it hasn’t been differentiated in DBTs, then the logical conclusion is that it either doesn’t exist or is inaudible.

the most logical thing (for me) is actual noise which somehow influences other stuff, even if it cant be measured on the dac output …

No, I’ve just told you what the logical thing is. If there’s no (additional) noise measured on the DAC output then how on earth is it “the most logical thing” to assume that noise (which isn’t there) is causing an audible issue? So that is NOT the most “logical thing”, it’s the most ILLOGICAL!

i come from a standpoint where i hear differences first and try to find a logical explanation to kinda try to make it a "whole picture" to go from there …

Again, that’s nonsense! It is NOT a logical explanation, it’s illogical!

Firstly, you keep falsely making claims of hearing differences but you have not ascertained that you can actually hear a difference!

Secondly, try to actually think of the basic logic here, rather than just falsely claiming logic and then doing the exact opposite! What about those who think they’ve seen/experienced ghosts, unicorns, a flat earth, mermaids, etc. Is science just wrong because it hasn’t figured out how to measure/detect any of these things yet and doesn’t have the “whole picture” or is science right and it can’t detect these things because they don’t exist outside of some peoples’ imagination? If it’s the former, then you must believe the earth is flat and that unicorns, Santa Claus, etc., are all true/real?

So although you seem to have no realisation of the fact, you “come from a standpoint” that’s utter, illogical nonsense! If you have even the slightest interest in logic (as you repeatedly claim) then you must first determine if there are actually any differences, if they are actually audible and then you try to find a logical explanation. You CANNOT start with a false assumption (that you actually hear a difference) and try to find a logical explanation for it because you do NOT have the basic facts, let alone the “whole picture” and therefore, your explanation will not be logical/rational!

but the best source i came across so far which kinda line up with my current "audiophile" believes (which might be interesting to some) is the website&youtube channel "Alpha Audio"

You mean the best source for supporting a nonsense audiophile claim. It certainly is NOT the best source (or even close to it) for the actual facts, for how Ethernet actually works and therefore why their claims are nonsense.

they "kinda" blind tested (tho you guys will critique this) ethernet switches

You mean the fact that it wasn’t a blind test. The guy on the right actually told the test subject he was changing switches and then stood up and did so. Are you claiming the test subject didn’t hear him say that or see him go and change switches? If not, then how is it any “kinda” blind test, don’t you know what a blind test is?

and also measured them and there was kinda big differences, for example power supply noise distributed to the ethernet ports, port to port noise isolation etc (which was surprising for me since the common explanation includes that ethernet ports are completely isolated)

No, the “common explanation” is that the data is isolated from noise after it’s been received, when it’s buffered. So you clearly don’t even know/understand what the “common explanation” is! And, they did NOT measure this, all they measured is noise (which they themselves injected!) into the transmitting ports.

This is ridiculous, they’ve got tens of thousands of dollars worth of measuring equipment but apparently don’t know where to measure or even what a blind test is! Are they really so ridiculously ignorant after buying all that high tech equipment or do they actually know what they’re doing but are deliberately scamming/misleading their audience?

the blind tests on youtube where you can listen yourself is a really interesting approach

If you’re streaming on YouTube, then the data has probably gone through dozens of switches/routers. If just one switch is audibly noisy, how come dozens don’t swamp the data with noise? Not to mention that YouTube uses lossy compression and loudness normalisation.

Alpha Audio is actually the only really good channel i found which try to combine objective and subjective matters …

This too is FALSE. What “objective matters” are you referring to, measurements taken from the wrong place and a blind test that isn’t a blind test? Do you know what “objective” means?

maybe they do it this way (not ABing) because they found its more effective …

Sure, it’s obviously more effective for spreading/supporting nonsense audiophile claims but it is NOT more effective for getting the actual facts! Which is why science demands DBT (not even just single blind testing, let alone test which aren’t even blind) and objective measurements.

(for me) they do a great service for the community

How is supporting nonsense audiophile claims/marketing, doing “a great service for the community”? They’re doing the exact opposite, a great disservice!

i dont think they just do it to push audiophile claims, i dont see the incentive …

You’re joking? You don’t see any financial incentive from advertising revenue (from YouTube views) or even from them actually asking for money/donations?

i understand it this way, ethernet switches are a heavly controversional topic.

Yep, so again it’s not the actual facts that are being disputed but your personal understanding! Ethernet switches are not even a slightly “controversial topic”, let alone a heavily controversial one! Where on earth did you get this nonsense claim from? Let me guess, audiophile marketing and the small number of audiophiles deluded by it. Not even many audiophiles think it’s controversial, let alone the pro-audio world or the hundreds of thousands of actual network engineers and scientists!

Again, it’s just all illogical, irrational anti-science nonsense, which you falsely claim is logical based on pretty much zero understanding of Ethernet, blind testing and “objective matters”. While this level of ignorance is surprising in itself, actually posting it in a sound science forum is mind boggling!

G

 

[Ghoostknight]

If you’re streaming on YouTube, then the data has probably gone through dozens of switches/routers. If just one switch is audibly noisy, how come dozens don’t swamp the data with noise? Not to mention that YouTube uses lossy compression and loudness normalisation.

thats why the most logical thing is noise, noise that can enter your music setup, i said before that i dont believe data integrity is a issue, thats why -if- there is a difference (and people hear this) the most logical thing is noise, specially from the last switch directly connected to your music setup

You’re joking? You don’t see any financial incentive from advertising revenue (from YouTube views) or even from them actually asking for money/donations?

and they build a youtube channel completely around fooling (and liying) to "potential customers" for other(!) companys? sure you are free to believe that
and i also think there are 100 better topcs for a youtube channel click-wise

---

well, i dont will get in a huge discussion beside that since it will end nowhere

 

[gregorio]

thats why the most logical thing is noise, noise that can enter your music setup

If it were noise, then the “most logical thing” is obviously that dozens of switches would make dozens of times more noise than just one switch. So it would be dozens of times noisier, dozens of times more obvious in measurements and dozens of times easier to hear. But obviously it is not, so obviously that is NOT “the most logical thing”!

the most logical thing is noise, specially from the last switch directly connected to your music setup

How does the last switch know it’s the last switch? And, even if it did know, why would it add more noise than all the previous switches? It’s just more nonsense, your “most logical thing” is actually the least logical thing!

and they build a youtube channel completely around fooling (and liying) to "potential customers" for other(!) companys?

Probably, although as I mentioned, they might actually be that ridiculously ignorant despite spending all that money on measuring equipment. There’s no easy way to be sure.

i also think there are 100 better topcs for a youtube channel click-wise

Probably but then there would probably be many more channels/competition for those more popular topics. Maybe you can’t see the obvious logic there either?

well, i dont will get in a huge discussion beside that since it will end nowhere

You’ve already got into a “huge discussion” by consistently stating falsehoods, consistently claiming the “most logical thing” while doing the exact opposite, posting nothing but a BS YouTube channel as supporting evidence and actually doing that in a sound science subforum!

Clearly it “will end nowhere” because you just keep repeating the same illogical nonsense while falsely claiming to be logical. You don’t seem to even know what logic is, let alone how to apply it!

G

 

[Ghoostknight]

Clearly it “will end nowhere” because you just keep repeating the same illogical nonsense while falsely claiming to be logical. You don’t seem to even know what logic is, let alone how to apply it!

its a free land, you can believe that

 

[castleofargh]

thats why the most logical thing is noise, noise that can enter your music setup, i said before that i dont believe data integrity is a issue, thats why -if- there is a difference (and people hear this) the most logical thing is noise, specially from the last switch directly connected to your music setup


and they build a youtube channel completely around fooling (and liying) to "potential customers" for other(!) companys? sure you are free to believe that
and i also think there are 100 better topcs for a youtube channel click-wise

---

well, i dont will get in a huge discussion beside that since it will end nowhere

The most logical is always human error. Assume human error, search for it and then if it turns out that the listener could really hear a difference(anybody can sit in front of music and have opinions about sound changes, that does not prove the sound did change!).
If audibility is demonstrated, then we look at the setup and wonder what might have caused the difference. The priority being to make sure it’s not the test setup itself introducing extra cues(a click sound only with one device, some longer lag to recover the signal...).
Taking shortcuts like deciding that if I feel like the sound changed then it did, that’s unacceptable for anything without a high magnitude sound change being measured with it. Last time I checked, people were wrong about a great many things several times a day. Our foolishness and overconfidence should not set the standards for facts.

 

[gregorio]

its a free land, you can believe that

It’s the ONLY thing you’ve demonstrated, so what else would I logically believe?

Look, it’s not difficult, if as the video claimed there is noise at ~52dB and as you claim, it’s is not removed at the Ethernet receiving end, then as noise sums at between 3dB and 6dB, depending on whether it’s correlated or not, then two switches would result in noise of -49dB to -46dB. A dozen or so switches would result in a signal that’s pretty much nothing but noise, so no data and no music, just full scale (or close to it) noise. Is that really what you claim to hear with downloaded/streamed audio and if it were, why would anyone ever try to download/stream audio if all they ever got was just noise? Or are you saying that all the other switches do in fact remove the noise but the last one somehow knows it’s the last one and for some reason the noise is not removed?

If you’re making either of these two clearly nonsense claims, then you’ve still got to explain how this noise gets through the data buffers!

G

 

[bigshot]

and they build a youtube channel completely around fooling (and liying) to "potential customers" for other(!) companys?

 

[Ghoostknight]

Look, it’s not difficult, if as the video claimed there is noise at ~52dB and as you claim, it’s is not removed at the Ethernet receiving end, then as noise sums at between 3dB and 6dB, depending on whether it’s correlated or not, then two switches would result in noise of -49dB to -46dB. A dozen or so switches would result in a signal that’s pretty much nothing but noise, so no data and no music, just full scale (or close to it) noise. Is that really what you claim to hear with downloaded/streamed audio and if it were, why would anyone ever try to download/stream audio if all they ever got was just noise? Or are you saying that all the other switches do in fact remove the noise but the last one somehow knows it’s the last one and for some reason the noise is not removed?

i think each network switch limits incoming noise to the noiselevel it was "designed"for, tho this is just an uneducated guess
no the last switch doesnt know its the last one, but the the noise doesnt gets lower than the switch noise itself... it isnt that hard to understand

 

[bigshot]

this is just an uneducated guess

 

[gregorio]

i think each network switch limits incoming noise to the noiselevel it was "designed"for, tho this is just an uneducated guess

Why rely on an “uneducated guess” when it’s not so difficult to find out how they actually work? Firstly, the data is sent using differential signalling, so any added noise is lowered by around 50-70dB through CMR (common mode rejection), which is why even high data rates can be achieved with unshielded cat 5 cable. The data is then buffered (and error checked), so there can be no added noise because a data buffer only has two states, on or off (representing one or zero). There are no other states, no partial on or off states, or any other states that can store sound or noise, so the conditions which allow noise no longer exist and therefore the additional noise cannot exist. When the signal gets to the switch’s output port it is then liable to noise, until the signal is received by the next switch and buffered again. So, you can have a pretty much endless series of switches because any noise effectively ceases to exist when the signal is received by each switch and the data buffered.

no the last switch doesnt know its the last one, but the the noise doesnt gets lower than the switch noise itself... it isnt that hard to understand

And where does the output from that last switch go? Do you plug the Ethernet cable into a wall socket, a balanced studio monitor or phono amp connection? Of course not, the Ethernet cable from the last switch obviously has to plug into an Ethernet port and this port obviously has to comply with the Ethernet protocol, otherwise by definition it’s not an Ethernet port. This means that this input port must apply CMR and the data must be buffered (and error checked) according to the Ethernet protocol, which means the noise must cease to exist, exactly the same as it does on the input of each switch!

You’re right that this really isn’t very difficult to learn or understand, so the obvious question is why haven’t you? Why come to a science forum with uneducated guesses and nonsense audiophile marketing BS? If you don’t know, then ask but if you just make up your own nonsense or recite audiophile marketing and just keep repeating that same BS over and over, in an actual sound science subforum, you give us no option but to conclude you must be an idiot, a troll or both.

G

 

[Ghoostknight]

Firstly, the data is sent using differential signalling, so any added noise is lowered by around 50-70dB through CMR (common mode rejection), which is why even high data rates can be achieved with unshielded cat 5 cable. The data is then buffered (and error checked), so there can be no added noise because a data buffer only has two states, on or off (representing one or zero). There are no other states, no partial on or off states, or any other states that can store sound or noise, so the conditions which allow noise no longer exist and therefore the additional noise cannot exist. When the signal gets to the switch’s output port it is then liable to noise, until the signal is received by the next switch and buffered again. So, you can have a pretty much endless series of switches because any noise effectively ceases to exist when the signal is received by each switch and the data buffered.

the Ethernet cable from the last switch obviously has to plug into an Ethernet port and this port obviously has to comply with the Ethernet protocol, otherwise by definition it’s not an Ethernet port. This means that this input port must apply CMR and the data must be buffered (and error checked) according to the Ethernet protocol, which means the noise must cease to exist, exactly the same as it does on the input of each switch!

what about the ground/shielding?

 

[bigshot]

You’re right that this really isn’t very difficult to learn or understand, so the obvious question is why haven’t you?

He isn’t here to understand. If he was, he would have understood a long time ago. He’s here because he’s in love with the sound of his own voice. His alter ego is doing the same thing. It’s great that you’re laying down solid info. I applaud that. But you’re addressing the lurkers now who want to understand better. You’re not addressing him any more.