Thread: New ethernet cable - surprised at improvement

Just installed a new ethernet cable - Wireworld Starlight 8 Twinax Ethernet Cable for Sale - between the Lumin L1 and Lumin X1.

Generally I'm at least a little skeptical about improvement from this kind of thing, particularly when it comes to the network side of things (I mean "cat 8 is cat 8" right?), but also generally willing to keep an open mind and keep trying.

Maybe it's just confirmation bias or something else at play, but I was hearing not-subtle improvement in a few areas. Improved attack and an overall sense of crispness/resolution of small detail, better separation and layering, voice and instrument timbres were more realistic (guitar, particularly acoustic guitar, double bass, horns, piano, percussion), and a slightly increased soundstage in all directions.

I'm pretty attuned to instrument timbre especially guitar and sax since I've played both for many years, and it went from "eh, it sounds close but not all the way there" to "yeah, that sounds like a real, live acoustic guitar is supposed to sound".

Again just kinda surprised at how much it seemed to make a difference. And this is a pretty inexpensive ethernet cable.

Jitter, jitter, jitter.

Originally Posted by BlueFox

Jitter, jitter, jitter.

In LAN/Ethernet connection?
Nope.

Timing is not a problem with ethernet connection - the data is transferred just like any other computer files using TCP/UDP protocols and buffered on streamer/renderer.

It's completely different story compared with the connection between transport and DAC where jitter/timing does have an effect.

Electrical signal prone to jitter is between streamer/renderer and DAC - not before streamer. The data that came through ethernet connection is 100% correct in memory buffer - before clock is even used to create that electrical signal - for DAC.

If streamer/renderer is not implemented properly and with proper galvanic isolation of ethernet port, electrical noise/EMI could affect the generated electrical signal on streamer output - toward DAC.

Originally Posted by Puma Cat

TCP/IP is just transmission protocol, nothing more.

The actual signal is an analog square wave voltage.

It's noise on the ground plane in the form of high-source impedance leakage current on Ethernet and USB cables that causes the problem. This voltage is impacted by the high-source impedance leakage current, and results in a specific form of jitter known as threshold jitter, which concomitantly impacts timing.

Yes. Ethernet data is transmitted as an analog signal. The receiving device converts it to digital.

Is an ethernet cable digital or analogue? - Quora

Thought her explanation was helpful, as well some of the ones further down.

In a separate thread I finally - after Puma patiently schooled me lol - realized I'd forgotten this part of networking theory but it is in fact what I was taught. In my probably over-simplified version the data (songs represented by a bunch of 0s and 1s collected to a file stored at rest on a hard drive somewhere) is digital in the sense that the data is encoded as 0s and 1s. That file is transmitted in the form of ethernet packets ("ethernet" being one of many possible definitions of the arrangement/structure of that "data" required to ensure the data gets from point A to point B as intended). On the receiving device side it's unpacked/decoded from bits to something the intended device interprets - via software - for it's purpose (audio video websites whatever).

But the way is gets from device A to device B is "analog" in the sense the 0s and 1s are represent by electrical voltage changes over that piece of wire. When Puma and others talk about why ethernet data transmission is subject to types of interference causing jitter and timing problems it's the analog/electrical nature of that transmission they're referring to.

Hopefully I simplified that in a mostly correct way. Ironically when I was studying this kind of networking theory I was sort of like "who cares move on this makes no difference to me some EE somewhere handled it". And now, I care!

Originally Posted by Puma Cat

a specific form of jitter known as threshold jitter, which concomitantly impacts timing.

reminds me of the good old days when tuning up my car (points ignition), dwell impacts timing yet timing does not impact dwell.

Sorry Stephen, I couldn't resist .....

Originally Posted by Puma Cat

Yep, this is what as known, Dave, as a statistically-significant interaction of control factors influencing a functional response.

A classic one in baking, as well as chemistry (of which baking is one form of..), is the Time*Temp interaction.

Yet, science students at Uni are not taught this, students are instead taught OFAT....One Factor At a Time.

Regarding baking, the wife noted you left out one of the most important considerations, 'measurements", if your measurements are off, your times and temp are off as well. carry on.

Originally Posted by Puma Cat

Before you would characterize any statistically valid functional interactions with DOE, it goes without saying you would do an statistically valid measurement systems analyis, aka, MSA...to qualify the measurement system before doing DOEs to characterize interaction studies.





I didn't forget about it at all, I just didn't want to get into presenting an entire course on Design for Six Sigma here. (FWIW, I used to do this professionally as a DFSS Master Black Belt).

Carry on.

What you noted has zip to do with my comment on baking , Define the problem, no measurements . But interesting., .

Originally Posted by Puma Cat

Ethernet cables make a BIG difference. It's not as simple as matter of "Cat 8 is Cat 8"....I could write a white paper as to why, suffice it to say there a number of reasons why they matter.

You should rent in a Shunyata Ethernet cable some time, even the entry level Venom blows the WW Starlight Cat8 into the weeds.

There are other cables then just Shunyata, it's all very system dependent, not just our own systems.

Originally Posted by imprezap2

There are other cables then just Shunyata, it's all very system dependent, not just our own systems.

@imprezap2 what do you like for ethernet?

I have tried a few different brands in my system (also Shunyata), I stream with my dcs Rossini and Aurender N10 and prefer different cables for both of them.
My preference of ethernet cables changed after adding a Melco S100 switch (with Plixir power supply), therefore I think it is useless to give "hard" recommendations as there are so many variables. Best is to fix a budget and try a few different brands.

Originally Posted by Puma Cat

What virtually none of the other Ethernet cables mitigate, which is quite important, is common-mode noise.

With respect to conducting experiments examining many variables, this is easily handled by conducting a "screening DOE" (Design of Experiments). JMP's Definitive Screening Design DOE platform is perfect for this. The functional response should likely be threshold jitter and phase noise.

Using Definitive Screening Designs to Get More Information from Fewer Trials - JMP User Community

You can make it as scientific as you want, as far as I understand you conducted your ethernet cable test in your own system and that's fine, but that does not mean that your "best" ethernet cable is the best suitable cable for other systems.

Originally Posted by Devg

If I were given a choice between a audiophile grade network cable vs audiophile network switch, I will opt for the later.

In my implementation involving a copper ethernet cable, the switch doesn't come into play. This is a direct ethernet connection between the NAS and the streamer/DAC.

Originally Posted by nicoff

Some folks will claim to hear differences/improvements.
While others will rely on test results to show that ethernet cables do not make an audible difference (for example, Archimago's Musings: MEASUREMENTS: Ethernet Cables and Audio...)

Except that's not what he found at all. All he showed was that he was unable to construct a test that could detect the differences people report that they are hearing. Anyone who understands science knows that broad conclusions cannot be drawn from tests that find no differences. His hypothesis was that if there were differences they would show up in the measurements he selected. The results didn't align so he would have to construct a different hypothesis as shown in the flow chart.

After the powercords in my system, the ethernet cable between switch and streamers makes the biggest difference. My system seems to be less sensitive regarding speakercable and interconnects.

dCS is recommending a standard unshielded ethernetcable, but I prefer a shielded or a floating shield cable in that position.
Differences are easy to hear, no "golden" ears required.

4G/5G, WiFi, Bluetooth, DECT Phones, power supplies, digital gear, power grid, all can have their influence IMHO, you just have to find the right cable for your system.