Very true, but no one’s commenting on the musicality of that signal. 
There is no noise on 0s and 1s.
Ethernet cables transport digital packages and the TCP protocol ensures that you receive to sended data in unchanged form.
If switches would add noise, a service like Tidal wouldn’t be possible. A Tidal stream passes thousands of switches on his way around the world till it get to you. And believe me, internet nodes don’t waste money on audiophile switches 
1 Like
The following article does not directly address the question of Ethernet cable , but does talk, in a slightly round about way of the value of eliminating switches.
I think the case for Ethernet cabling is focused on whether retries are less with more expensive cable.
BTW - Just so my motivations are not misunderstood, my old ears can’t tell the difference between Amazon Basic and Audioquest Cinnamon. I’ve got plenty of time and some discretionary income, so I come to this discussion purely on a tinkering quest.
The problem I have is with the way the argument is parsed.
1 Like
The following article talks, in a slightly round about way, of the value of eliminating switches.
As to your observation about Tidal, I thinking that the majority of transmission occurs over fiber optic media and switches. Once and every time fiber optic transmission occurs the integrity of that transmission is reset.
Even if that is a possibility, how does that affect sound quality. The data handling capabilities are way in excess of what is necessary to supply a stream over Ethernet for music reproduction.
I believe it is impossible for any Ethernet cable to sound different to another. Digital data has no sound quality. Proving my belief wrong will not be established by someone saying that they ‘heard’ a difference.
1 Like
I’m listening to Brent Cobb via Tidal in MQA on my Bluesound and the MQA is Authenticated with normal cat5e cable and a couple of switches.
This is the Studio Sound (as best the Bluesound Pulse 2 can manage) What am I not getting by having standard cable?
5 Likes
Here is what I did:
Cisco SG 200-8 managed layer 3 switch with ports 7/8 in a LAG group (one MAC address that routes at layer two for both ports). This means that the frame on the switch is addressed to a single MAC address and either port will get it with whatever cable is currently connected to either 7 or 8.
Two setups:
Computer with Intel Dual Port Server NIC to Emotiva Stealth DC-1 balanced out to RME ADC UFX (extremely high end ADC).
Cary Audio DMS-500 streamer with balanced out to same RME.
315 foot bulk (BerkTek Hyper E) CAT5e cable at $0.30 a foot. So a $90 cable that I terminated.
3 foot Audio Quest Vodka at $340
12 foot WireWorld Starlight at $350
3 foot Nordost Heimdall 2 at $699.
For added measure I tossed the 315 footer underneath a running microwave while I did the captures. I posted the captures for download and just labeled track 1, track 2. So I blinded the tracks. No one to date has ever reliably been able to tell any difference and neither could I.
Here are two good links to read about CAT cabling and how absolutely immune they are going to be to external noise sources:
Siemons “The Antenna Myth”
http://www.siemon.com/us/standards/Screened_and_Shielded_Guide_7_Antenna_Myth.asp
T.I.'s 'Reducing the Radiated Emissions of 10/100 LAN Cabling
This Siemons paper shows how resistant to EMI LAN cabling is:
http://www.siemon.com/us/white_papers/97-10-02-presentation.asp
I personally use wireless. For ~$80 I setup a dedicated 1200AC router in WAP mode and a USB AC600 2dBi network adapter and get routinely 36MB/second throughput. Not even upsampled DSD could saturate $80 worth of wireless and I have total isolation from the network WRT ground plane and EMI.
1 Like
One 315 foot of Generic cabling:
- Sub 1ms ping times
- Out of 1000 pings (ping 192.168.0.1 -n 1000) I had zero loss, high of 12ms, average 1ms.
- On transfer of 280GB, according to the switch I had zero errors. No retransmits. No nothing.
In a home network a $700 cable isn’t going to improve any of this. Just get cables that you know hit spec.
And yet –
You see, that’s really the point. If people ‘hear’ a difference with expensive add-ons, then since anyone’s hearing is by necessity a subjective experience it follows that the expensive cables made a difference (however it’s perceived) in that person’s listening enjoyment.
For example, MQA is first and foremost meant to be a new way to do lossless, for transmission purposes, but folks also ‘hear’ an improved difference over other lossless formats.
My theory is that there is zero difference and it’s all in your head. Why you aren’t entertaining confirmation bias as a source of the improvement is beyond me.
Cable that passes spec, and are built roughly with the same construction, shield the same way, are going to sound identical.
I’ve got $2000 to your $500 plus loser pays travel expenses that once I connect my layer 3 switch with LAG and a $12 CAT6 cable and whatever cable you are using, you won’t be able to tell the difference any longer.
Your setup, your room, your music. You just won’t know which is which. Can use foobar abx utility play the same tracks copied twice to two difference SMB shares and a dedicated network link for each.
4 Likes
And Darko, Van Es, Lavorgna, Hull, Plaskin are all welcome to sit down and evaluate $120 foot cabling vs $0.30 a foot blinded.
Honestly their Ethernet cable reviews are BS. They are 100% making stuff up which means they will make anything up at anytime. I wouldn’t trust a review of anything by any of them.
4 Likes
The wise man speaks.
Ending my participation, in this discussion.
1 Like
That is correct. And by short runs, I mean anything under 100 meters.
Absolutely correct. Carrier grade provider § and provider edge (PE) devices are designed to pass packets as quickly as possible. These are devices with back plane bandwidths in the millions of GB per second. If there are any quality of service (QOS) settings applied, it is based on paying customer tier, not necessarily payload. And QOS is typically applied at the customer edge (CE) devices anyway.
This is due to transmission distances and bandwidth requirements. A carrier grade transport layer optical device has to send traffic for many miles/kilometers between optical repeaters. Copper just can’t do that. Single mode fiber with various multiplexing techniques are required to be able to handle 100Gig WAN PHY or OC-768 Sonet. That is a completely different beast than local area networking.
The sound that produced by your audio system is created from the payload of the IP packet, not the packet itself. So as long as the payload makes it from end point to end point, then there is absolutely no impact on the quality of the sound.
7 Likes
Here is another item that many don’t realize. These are non-realtime systems. My setup has 20 seconds of buffer. I can literally pull the plug/disable the wifi and music still plays.
Did my sound get better? Get worse? Stay the same?
2 Likes
It’s on the internet, so it must be true. What other prove do you need?
2 Likes
I understand. The OP was making the argument that if switches are a factor then TIdal couldn’t work. I’m simply saying that this argument can’t be used because probably for the most part the media and switches are fiber optic, not Ethernet centered. I make no reference as to why this is case.
Can you restate that differently as I’m not sure what you are getting at.
No one is suggesting that. Ironically, I hold similar beliefs about audio reviewers. One of the names mentioned has banned me from his forum for questioning his reviews. I included this article because there are alternative interpretations based on noise, but you’d have to read the article to know that.
Your disrespectful snarky comment adds nothing to the topic.