The other crucial point only hinted at in the replies is that any metal cable is susceptible to interference from other electrical cables, particularly mains cables. IT installers are taught to avoid running lengths of ethernet parallel to any mains, and where they meet its preferable for them to cross at right angles. Also the installer will avoid sharp bends and tight angles which distort the cable.
1 Like
Shielded instead of non shielded could couple components and transfer noise. Donāt use shielded.
3 Likes
Hereās an interesting read from Galen Garis, the man credited with designing the vast majority of Ethernet and other cables for Belden for 35 years
2 Likes
Interesting post by Mr. Garis. Perhaps you can also post a link to version of the post that mere mortals without an electrical engineering degree might understand. Or maybe a decent summary of the main points. Thanks in advance!
1 Like
Iām afraid thatās about as simple as he knows how to communicate! Can you imagine when he speaks with other engineers? (they listen)
The take home is cat 6 is likely a better choice for our uses vs. 6a. Will it make a difference that is appreciable, probably not, but no need to pay more for 6a.
My take away, donāt sweat it, buy good cable that meets spec (tested and sold by Blue Jeans is a great choice). Iāve also had very good luck with patch cables from Monoprice over the years with very, very few failures in many thousands.
4 Likes
I am OK with relatively expensive cable where it does count, hence Shunyata Delta power cables.
Ethernet - properly made and tested seems all that is needed - so I use Blue Jeans for my audio and a 25m run of AudioQuest Pearl CAT7 to distribute video around the house. The Pearl CAT6 would probably have been sufficient, but we were taking up the floorboards and skirting so an extra Ā£100 was not a big deal.
Only if your existing cable is defective. (In my opinion.)
yes, to some extent. The better cables are stiffer, and larger in diameter than cheap cables, making them less likely to get sharp bends in them. Any ethernet cable can lose performance if there is a sharp bend in it. I sold Cat 5 and Cat 6 cable to installers for several years, and learned from the pros that sharp bends can reduce cat 5 performance to the performance of that of a Cat 3 cable (or worse). Also, the better cables provide better isolation from various sources of interference. But, that being said, if the cheap cable is installed and located optimally, the differences can be greatly minimized, sometimes to the point of being insignificant.
If you think the sound is different it is not because of proven theory go study CCNA or better and come back and make the same argument you wonāt. https://www.inetdaemon.com/tutorials/basic_concepts/network_models/TCP-IP_model/ note this is free training educate yourself.
By the time the music has reached the network cable it is packets of data not full bandwidth sound as in analog RCA cables. If a NIC cable was damaged (or poorly constructed) the amount of data (packets) being moved through it would be affected. See link above the only way you hear this is skipping (data stream affected), lower quality audio being played because the speed required isnāt supported (it would throttle down, if supported by Roon I wouldnāt know). Usually just doesnāt work at all cable map is tested in better equipment before link comes on. If it was still working you would notice this in higher ping times, slower FTP speeds (a quick test), speed tests between internal and external network devices that support it.
As for cables being stiffer the gauge and quality of copper is most critical if you need to make a sharp bend you use multi strand as apposed to solid core. (I worked at Nortel we regularly ran both network and optical fiber cables to stress test our equipment (beyond rated distance) most network cable today is 300ā before you need worry. This was for clients who did scientific experiments joining many super computers across country together. If there was a problem it would be evident immediately because it would go from taking 10 years to do computations to 20 years. When I left in 2004 there were projects that had been running for 25 years.
Anyway back to point Analog RCA cables make a difference as the actual sound you will hear transfers through them not just oneās and zeroās or packets of data. The the actual sound you hear can be rolled off or have a suckout as it passes threw depending on quality. Blind folded I can hear difference with better analog cables but after a certain point how much do you want to pay for 1-2% improvement in sound? Many affect sound in subtle ways, so how they interact with each piece of equipment in your audio chain is affected.
Both Digital and Network cable protocols have built in error correction. In extreme circumstances could affect audio quality but if loss is that extreme you are talking the difference between FM radio and AM radio as quality stepped down. Not little nuisances. Now if you are getting RMI, EMI, Electrical interference this can happen. That would not likely be the quality of network cable even cheapest stuff has basic testing done before it leaves factory this is all worked into design. I buy monoprice NIC cables off amazon havenāt let me down yet. over 20ā I make myself and test. I listen to Roon, while kid watches 4k streaming from Plex, wife watches 1080p tv shows from Plex. In my home I have a security Gateway, 2 x 8 port poe switches, 1 x 24 port switch, three wifi access points, a security camera system recorder. So video cameras are also streaming at same time. Have never heard sound affected occasionally my QNAP needs a reboot as ladies are complaining video gets choppy as a service is is taking too much CPU or Memory, Roon unaffected even when I am listing to 128 bit DOP files the Qnap is my server.
I have a few rolls of shielded network cable it is 2x the thickness of standard stuff I have used it grounding the shield at both ends not a spec of difference in audio, data transfer speed, that was tested using Fluke network tester that cost more than my car at time. Now this stuff is for labs and power sub stations that have 750v or higher power lines the thickness of a 5/8 garden hose (one cable, copper only, there are three that make up circuit) running close by.
As always everyone is allowed to have their own opinion, but please research, you will be much better served with better networking equipment then better networking cables. If you are in a multi user environment. Unless you need a run greater than 300ā. Then go to multimode fiber, with devices that supports it at either end then run nic cable from those devices to equipment.
Back to the music
5 Likes
I have to agree with @Jerry_B. My day job is telecoms (Nortel/Avaya/Cisco), international voice, etcā¦ and IMO, an Ethernet cable does not make any audible difference. It canāt. Any transmission errors are corrected by various mechanisms way before they ever get anywhere near being sound, and youāre probably not going to see any errors on a short patch cable anyway. In the analog domain OK, arguably S/PDIF too, but not with TCP/IP. But hey, if you hear a difference, more power to you 
10 Likes
Well, I am in theory a network engineer, so I should know what I am talking about. Conceptually and practically I have to say no, ethernet cables donāt make a difference (above cheap no-spec cables). I have a $1K ethernet cable hooked up to an isolator and I can not hear any difference from just going direct to the isolator or switch. Now before you jump on me, I have absolutely heard improvement with power cords, USB and especially interconnects. Ethernet in my system makes no difference.
But you should demo a few and see if it does for you. Every system is different.
2 Likes
Yup I will admit I hear difference with USB, bought a monolith from Amazon. I also needed a iFi Defender 3 for ground hum. Both added at same time. Now existing USB was pretty good too. Older Belkin cables were very thick and well made. Would often solve MFC problems.
Thought I was nuts when I heard difference with RCA SPDIF but did go from a skinny sound card cable to one thicker than a coax cable (Monster).
I have varying Digital SPDIF cables donāt hear more than 5% difference. So mainly purchase those on basis if they click or lock in, in some manner have had some issue with that.
Age of cable enters into this also unless you live in a bubble.
Here error correction and possibly noise enter into it causing higher error correction thus loss of detail.
As you know the network theory behind it.
I am no engineer just a tech but the network model has accounted for all of thisā¦so back on topic spend on equipment not network cable.
To answer the original question, Ethernet cables shouldnāt make a difference. But obviously some peopleās experiences suggest that they do. The reason is simple. No one who works in the wider industry connecting people, companies and even countries have to deal with $15 SOHO switches, power plugs and poorly implemented WiFi. Done properly, there should be nothing a cable needs to fix. But unfortunately that isnāt always the case. And to be clear, the issue isnāt data integrity. The problem is everything else that can potentially travel along that cable.
So to illustrate the problem, we know a CAT5 cable should be able to carry 100Mhz over 100 metres at various data rates. The reality is that if you take one of these cheap SOHO switches as the source, perfectly good cable will appear to underperform. So my assertion is the problem is the hardware, and not the cable.
My experience is 39 years mostly in Telecoms, the last 22 of which has been in voice and data. I am Cisco certified (the networking equivalent of painting by numbers) and have been a (small) part of major national networking projects here in the U.K.
1 Like
Love itā¦
2 Likes
That is categorically ridiculous. Nobody with a minimum of experience buys the snake oil marketing hype, but empirical listening tests between different ethernet cables between ISP modem/router, switch and streamer, including various combinations therein, do indeed produce different, audible results. If that defies the science around binary data flow, fine. The differences in ethernet cable performance and rendering is a clear reality. Itās about the sound, not the science.
Iām sorry you find engineering truths to be categorically ridiculous. Yes, your opinion defies the science and engineering around data flow via ethernet cables. Such data flow is not āsoundā yet, itās a stream of data. Others have pointed out better than me the engineering reasons for this in posts above. This said, Iām perfectly happy for you to believe whatever you want. We now live in a world of alternative facts where people can choose their own reality.
4 Likes
Hopefully, that will end soon.
2 Likes
Hi Gary, no need to be sorry. I didnāt require your permission to believe what I want. Some folks believe what they read, what they see or what they hear in this case. I prefer to decide for myself. Specifically, changing ethernet cables in a controlled environment with all other system elements in place, produces clear sound differencesā¦for those who can actually hear. I will be the last audiophile to jump on a never-ending quest for the ultimate 1000 euro ethernet cable. I have now what suits me after empirical testing, listening and re-testing. In fact, I can tell you that there are some ethernet cables that will never, ever come back into my systems due to listening preferences. If this is in direct conflict with electrical engineering data, then Iāmānot sorry. I donāt listen to theory, I listen to music. Not trying to convince you, but your scientific certainty
cannot supercede my audible perception. Youāre happy, Iām happy.
If it sounds better to you, then it sounds better. What else matters?