Apologies, but even as a relative newbie I need to take issue here. Nicely and supportively, of course!
You say “Logically, ethernet switches, ethernet cables, power cords, even clock cables shouldn’t make an audible difference but I’ve personally heard that they all do.”
This not actually logical at all. It comes down to noise (RFI/EMI) and shielding from same and if you/the manufacturer address this correctly then there is a good chance you will hear a difference:
Ethernet switch: any switch will break the noise “chain” and stop noise picked up on your ethernet cable from reaching your equipment. It’s true, the boards used inside most “audiophile” switches are stock boards you can find in many a budget switch. They often have a super-accurate clock added but if you know how ethernet works, this will have zero effect on sound quality unless, theoeretically, it replaces a noisy stock clock. The key to superior performance is shielding: instead of using a plastic case or an aluminium case with vent holes in it, wrap it in milled metal and it will reject noise far more effectively.
Ethernet cables: I think others may have pointed out that you should use CAT cables not proprietary “ethernet” cables sold by boutique manufacturers as these often have metal rather than plastic plugs and they stop the galvanic isolation in a typical switch board from doing its job. The noise they let through can initially be perceived as more detail but it isn’t.
Power cables: screen them to keep the noise inside them and thereby away from sensitive signal cables.
Clock cables: no idea or experience but I’d suggest if any difference is heard then it’s noise-related.
I don’t personally buy any of the guff which suggests there is anything magical going on in the digital sphere, at any level of the ethernet protocol, and neither should anyone else. It’s analogue noise, pure and simple; the clever bit is how you reduce this noise to its lowest possible level.
Totally not a problem Nigel - you had something to add to the discussion and these discussions are always findable on searching (as is common with things on’tinterweb) so IMO your addition to the topic is totally justified…
My ears/brain are the final arbiter. Using solely Ethernet from source to dCS Upsampler for years. My experience with stock switches and cabling and then testing various alternatives has revealed that “noise” does somehow creep its way into the system. And I’m thinking the sources of noise attack like a tenacious virus into all points in the chain, from the power company through the very atmosphere.
I won’t make any claim on the technical side but only that it has the effect of letting the music emerge from a “dirty” sonic haze once deployed.
Also, after many years of testing and using various digital and analog interconnects of various designs and technologies, I’ve found the noes designed by Shunyata Research to be the most effective. Not only at dealing with “noise” but in what I can only describe as ensuring the proper flow of electric signals, or whatever. The net effect: a lower noise floor and musical flow. Again it’s all in my head and heart.
I thought I would add my experiences here, without wanting to justify my choices or get into a debate about why the changes could /should have made any difference.
I use a Rossini + Clock supplied by an Innous Zenith MK3 server, which in itself is served directly from my Virgin media hub which is coupled to the internet by fibre optic cable into my house. I also run Cat 6 floating screen(s) ethernet cable (Design Cable co.) between the Zenith and My Rossini.
With this rather direct arrangement I didn’t really expect to make much of an improvement - but I have now learn’t that the ethernet input performance is heavily affected by insidious noise riding on the cable - just like USB.
First I added a Chord Ground USB array to the server and to my surprise the Ethernet input on the Rossini sounded better. Next I installed an ADOT fibre optic kit (dual SPF) with linear PSU (with splitter cable) between my Virgin hub and the Zenith - this resulted in a very substantial drop in the noise floor when streaming from Qobuz. Now, the Qobuz streamed performance was much closer to that provided from the solid state memory from the Zenith. Qobuz was starting to be engaging. PS: The best performance possible is to load a file into the Rossini buffer from the server and then disconnect the ethernet input - of course this is not practical for extended listening but does provide a baseline to make judgments about the effects of the noise riding in/on Ethernet cables.
I was intrigued - could this audio improvement also apply on my set-top box which outputs to my Rossini (Coax) and TV (HDMI)? [The set top box uses a separate Ethernet link back to the Virgin media hub]. The answer was a resounding Yes!
I was initially concerned about the intentionally ‘low bit’ rate for the ADOT switch but this was not a limitation at all when streaming HD files on bbc iplayer. I ran this combination for two weeks to check and when satisfied bought a second ADOT dual SPF kit for specifically for my TV. Once you have seen the visual improvement and heard the improved dynamics you cannot go back.
So now with four SPF switches I wondered if better power supply would improve them further - just as expected the noise floor of the ethernet input to Rossini dropped again with all the usual audio benefits when I added a Plixxir dual output PSU just feeding the Hifi side SPF switches. Another purchase!
My spare Linear plug top supply was then re-purposed for the second SPF switch on my set-top box pair - another audio visual improvement!
Next I learnt about a new single mode fibre cable upgrade - this yellow cable is thinner (lower optical losses?) than the standard orange cable and just as with the PSU upgrades appears to improve the sound. I swapped between my two sets of switches (one with Yellow and one with orange fibre cables and the differences were easy to discern. So now I use Single mode cables for my Hifi and TV.
Whilst this approach is not inexpensive (or pretty!) the benefits are profound and represent very good value for money in the context of a DCS based system. Ethernet it seems is not immune to noise while exotic Ethernet cables still won’t provide isolation from noise transmitted / passed from connected devices. Admittedly the SPF’s are not completely noise free as proven with the PSU upgrades I made, however from my perspective they represent a credible and highly effective approach to improving streamed music sound quality.
Basically two different approaches to tackling the same challenge of stopping RFI/EMI noise reaching the DAC. As you’ve discovered, optical’s theoretical advantage can be limited in practice due to noise generated by the media converters themselves (and I include their power supplies here). In neither case is there any digital magic, both fibre optic and ethernet switches are digital boxes deployed to analogue effect. No bits are harmed or enhanced in the process!
I have tested the adot fiber with the Melco S-100 switch but it didn’t do much compared to my existing setup at the time.
I have however recently added a different fiber optic isolation and it made a HUGE difference.
It was literally a tipping point and what brought the whole system together.
A friend of mine experimented with the sonore optical modules and also got results he’s very happy with.
I had problems with dCS Rossini and Ansuz switch (NAS wasn‘t detected, roon didn‘t work). Problem was that Ansuz had some blocking/filtering (IGMP issues?) that didn‘t work. So I decided to sell Ansuz (but it really sounds great) and got an Innuos. Now does one know if APEX changes something streaming/networking wise? Or has one of you folks an Ansuz running with Rossini and Roon? Thanks!
Ok Vlade, that’s the point- you use the Statment LAN port to connect dCS directly. I used the Ansuz LAN port to connect dCS as well and that didn’t work. Had a chat with the dCS distrubutor 10min ago and they confirmed that Ansuz switches are still “blocking” …
Great - thanks Vlade! Please leave it for some hours if possible, it takes some time and then it blocks (at least that was what my dealer and I found out)- again THANKS!
Many thanks Vlade!!! That‘s really great …and drives me nuts because it‘s exactly my dream combination you have there!
I couldn‘t get Ansuz switch running with dCS even wit NAS systems from Qnap, Synology and Melco…it blocks the detection of upnp and roon services means neither Mosaic nor roon could play something.
I really don’t know what the problem is. Last year I used an Ansuz C2 switch, and it also worked perfectly well. You have to try something else.
After almost five hours the problem would appear, that’s for sure.
Ok, roon works with Ansuz the problem is the upnp service and Mosaic. This service is blocked by the Ansuz switch‘s IGMP spoofing. Had a call with my dealer again, he confiremd his findings that matches mine. Ansuz uses a D-link switch engine in its devices.
