I am using Duall Switch by PPA Studio. And then the signal goes to GigafoilV4 and to the DAC. I did not notice any interruptions in the music streaming, either in DXD or DSD128 quality. If we have an efficient data server like the slightly more expensive Qnap models, the 1000 MBS network should not be a bottleneck. Streaming local files is very stable. DCS devices have a built-in data buffer and it seems that it is sufficient for correct, continuous transmission. There are exceptions like 24 / 192Mhz file streaming but the reason is related to a bug in the DCS BartĆ³k software.
To give you my answer, I would say unequivocally yes. And in my experience significantly more that just āslightā. But of course, like everything else, the amount of improvement is subjective. A reason itās such a big thing for me is that the resolution of the soundstage is so much better now. Not just width and depth, but the tangibility and three-dimensionality of the performers on that stage. Also the tonality of acoustic instruments and voices is more accurate. But I guess if youāre into āwall of soundā type electronic pop, perhaps the improvement would indeed be merely āslightā.
Iām a great opera fan and these factors are key to creating a sense of being in the theatre - the physicality of the performance. Qobuz have recently added a remastered hi-rez version of the classic 1965 Karajan performance of Cavalleria Rusticana and Pagliacci. Itās a recording I know well and have had the CDs for many years.
Optical isolation effectively seems to wash the sound clean and allows the hi-rez to sparkle. The sound totally envelopes you, as indeed it does in the theatre. To be so palpably confronted with these performers long since turned to dust. Carlo Bergonzi, at his peak then, such an incredible voice. To be honest I got a bit too emotional, bringing tears to my eyesā¦
1 Like
Recently I swapped the Oyaide digital cables for the Vivaldi stack out and Black Cat Tron in. Tonight I swapped the Melco S100 in and the English Electric 8Switch out. The LPS for the switch remained the same - Ferrum Hypsos. The power conditioning for the switch went from Furutech Flux-50 to Furutech Daytona 303.
The self delusion I am under ever since is very pleasant, captivating and convincing. I seem to be good at switching on and sustaining an illusion effortlessly. Master of illusions? 
It has been a long road going from an endgame Naim system with speakers to a comparable music presentation and listening experience with dCS and cans. As of now, itās achieved. The meditative immediacy is there again. And then the qualities of dCS, which Naim sources never possessed.
To me the illusion / imagination argument when presented as universal is bogus. To fit the observations from forums and from dealer conversations to a generalizationā¦ Not all listeners react to a change in switch. Some do. Network setups react differently too. The science argument cuts both ways. It can render effects with unexplained causes as illusion. On the other side, if I am certain a difference is impossible, my threshold of perceiving a change others report can be higher than theirs.
Scientific reasoning is a valid selection tool for upgrades. Only few want to waste money or time. There are other pathways too. One that served me well is looking at the broader hifi preferences of people who recommend upgrades. If their system choices or the audiophile taste behind these choices is similar to mine, I listen more to their recommendations.
IMHO, a manufacturer of music reproduction devices has little incentive to take a contested position on a controversial issue. The upside is small, the downside of taking flak or damaging reputation is large. Similar situation for participants in online forums. Flak is unpleasant, risking reputation is undesired.
Thank you, @poseidon77. Three of the books I have in my professional library. I look at the others.
My occupation, capital markets investing, benefits from valid cognition. It concerns the future though. So my task there is to be right more often than wrong or right with more capital than wrong as the future is unknowable. There are precious few mental models that remain profitable for long. Or others look sound, sporting fancy mathematics and lots of statistical evidence. Like modern portfolio theory. The rub with that pervasive model is, you can optimize for low risk mathematically all you want when markets tank, correlations between asset classes have the habit of turning to one. And then, the insurance you think you had is out the window. What Iām saying is, when things become complex enough, reductionistic models have a stubborn tendency to fail. And as said above, hifi with the listener included is a complex system. I donāt aim to be right - just an attempt at fleshing things out more.
The invisible gorilla experiment is hilarious!
Popper may not have very much to say re the more mysterious workings of ethernet switches and cables (according to some hifi enthusiasts). 1) Popper is beholden to a positivist understanding of meaningful claims ā what is meaningful is basically reduced to what is scientifically verifiable (thatās something you also have in early Wittgenstein, before his linguistic turn). That leaves a lot of āmeaningfulā stuff out (even early Wittgenstein admitted as much: nothing to say about music, the arts, literature, religion ā¦). 2) Popperās heuristic tool creates as many problems as it purportedly solves. Popper is useless when it comes to theory choice (due to the issue of underdetermination of theory by observation); moreover, most scientists arenāt particularly interested in falsifying claims so much as corroborating them. All of this has led to Popper having been quietly removed to the sidelines in philosophy of science quarters. There are other problems, too.
I guess it all depends on how reductionist the model is, but I get your point.
Perhaps a way of having the best of both worlds (the need for a) objectively verifiable rationality as well as the need to b) remain open to certain possibilities that point beyond the more narrow parameters of a)) ā is by way of qualia (the intrinsically and utterly personal experience of a perception/ sensation). A great book on that would be Thomas Nagelās What is it like to be a Bat? surely the most wonderful title ever for a book on a complex philosophical issue.
Science has very little to say about qualia ā canāt be measured or mathematically codified. And yet, no one would seriously claim that your individual perception of, say, the colour blue us somewhat unreal. Pain is a good example, too: who can truly measure the level of pain you feel? Does that mean that your pain is an illusion? (well, maybe, if you want to go down the Buddhist line).
What is important about the issue of qualia is that it circumscribes a radically personal and individualistic thisness (in medieval terms, the haecceity rather than quiddity of something). Only you are privy to your qualia.
So, to get back to our topic, if you experience an improvement because of this or that tweak to your hifi system, then this is your experience, and nobody could possibly call that into doubt as far as the qualia reality of that experience is concerned.
You, as the one who has that experience, may wonder though whether itās real. Or perhaps not: I perceive the difference the tweak has made, you might say, and thatās good enough for me.
Or not, and then you will have to take a step back, take your own experience as an objective event (it is now endowed with objecthood: itās something not only in your mind) and run a number of tests on it to determine, not whether your experience is real, because it is (how could it not be?), but to determine whether your experience is possibly shareable, which means replicable (for and by other people).
The problem is of course: what tests? how will I set them up? But thatās not our concern for now.
What I have just said seems to run counter to some of my earlier contributions to this thread. No, not really ā I have just given greater consideration to the thisness aspect of an experience. When it comes to shareable experiences (those that fall in the public domain), I still hold that a scientific attitude is our best chance to cut through bullshit, fake news, and manipulation. Because what is deemed shareable, and especially those things that need to be shared (such as ways to control a pandemic for instance), have nothing to do with qualia and the solitary chambers of your mind.
Last point: qualia demonstrates that not everything can be measured! What you make of that, and how far you want to take it, is up to you
I agree, but Popper also has his merits. His rational alternative to inductivism, for instance. As a mathematician, I study a language, something different from a science. This puts me in a position from which I appreciate the merits and the limits of Popperās ideas.
As I always tell my students, theories, methods, ideas are great when you use them as tools. You must use the right tool for your aim, and know the limits of that tool: what you can and cannot do with it.
A.
In my view; keep your Roon Server far away from your Rossini, ideally in another room and especially on a different AC line, and interconnect them with Ethernet over Fiber. For the Media Converter at your Rossini end, Iād recommend using a good Power Supply (so as to not pollute the AC line near the Rossini), and an unshielded CAT5E/CAT6 cable.
I used to have a couple of different elements in the past, but none now, just stock standard Enterprise-grade hardware. The only remaining āupgradeā, if you could call it that, is a Keces linear PSU for the Media Converter at my Vivaldi end. 
With the exception of the EtherREGEN, those āsolutions looking for a problemā are actually priced-inflated OEMāed/re-cased versions of low-end Ethernet Switches; that includes the EE, Silent Angel, Melco S100, Synergistic EUF, Ansuz PowerSwitch, and a whole bunch of others.
The Phoenix is a clean design, or so it appears to be based on the product brochure (I havenāt seen the real unit, yet), but the fundamental āproblemsā Innuos claims to address are not problems in the real-world. So, itās just an expensive Ethernet Switch. 
Coincidentally, I just read the Organikās review in the current issue of The Absolute Sound (by the same guy who wrote the EtherREGEN review discussed in this thread).
IMHO, itās actually a smart decision to incorporate an SFP cage into the unit (like Lumin first did). Even though it increases the cost and complexity of the design, it gives Users complete flexibility in the type of physical media they can use for Ethernet - Fiber, Copper, or even Coax, with just an appropriate SFP. Plus, using Fiber eliminates the potential for ground noise issues, i.e. the whole shielded vs. unshielded Ethernet cable debate. As for the āslightly improved performanceā comments, it kind of depends on the circumstances, for example, the ground noise issue.
Well, since you already have a switch with an SFP cage, it wonāt cost a lot to add a Media Converter, PSU, and a pair of SFPs to try an optical connection to the Rossini. And you can experience first hand if thereās anything to fuss about. At the very least itāll be a learning experience. 
5 Likes
@Anupc
Thank you for the replies and advice. I just realized the Cisco 2960 switches I currently have are the WS-C2960CG-8TC-L models which do in fact have a pair of SFP cages. I just ordered a pair of original Cisco GLC-SX-MM transceivers and have decided to forgo a pair of external Trendnetās and their associated switch mode PSās and instead will use a pair of my Cisco switches. If I may kindly ask you to review my plan:
Router >>>
Cisco 2960 which will have other networked devices plugged in via ethernet >>>
Multimode Duplex Fiber Optic Cable (62.5/125) - LC to LC >>>
Cisco 2960 >>>
Cat 6A unshielded >>>
Rossini
The original Cisco GLC-SX-MM Transceiver was recommended on a similar thread on the Naim forum. The Fiber Optic cable comes from a thread on the dCS forum - is the specification correct 62.5/125? This is the Amazon link
https://www.amazon.com/gp/product/B079G56L38/ref=ox_sc_act_title_5?smid=A27DUHPPTQX1VZ&psc=1
Is there a recommended minimum length for the fiber cable or is shortest the best (7", 1 foot etc)? I may be mistaken however I came across a post indicating that a longer run of fiber may be preferable?
Finally is it recommended for the pair of Cisco switches to be on the other side of my dedicated listening room or can they remain physically close to the Rossini+Clock? The Cisco switches would be plugged into the regular house mains. The Naim/dCS system is fed by a Torus Isolation Transformer which in turn is powered by a separate dedicated 20 amp line run directly back to the panel.
Thank you for your help
Best
Gregg
Thatāll work. 
By the way, the two end-points of a fibre link donāt have to be matched devices; you can have the Cisco at one end, and a TrendNet Media Converter at the other end, as long as the SFPs are the same (or compatible).
Good idea to stick to a Cisco SFP, for compatibility. The electrical specs of SFPs are very tight, but SFPs are programmable to contain vendor specific settings that can occasionally make them āincompatibleā with some Switches.
Some Switch vendors also go as far as keying their software exclusively to their (heavily marked-up) SFPs. FYI, almost no Switch manufacturer makes their own SFPs, even the Cisco SFPs are OEMāed (from II-VI if I recall correctly).
Also, donāt believe any online posts suggesting SFPs affect the sound. Thatās as absurd as suggesting a Web-browser you use to access your online bank account can affect your bank balance (Audiophiles with overactive imaginations 
)
SFPs feed into the Media Independent Interface and then into a switch chip and eventually into the buffered/error-corrected TCP stack on your Rossini. Unplug the SFP and your music will continue streaming for a good 30 seconds or more.
Thatāll work . As long as you match multimode SFP transceivers with multimode fibre.
Thatās correct in broad terms. Itās not going to affect the sound either way, but the receiver in the SFPs operate best when the signal is not too hot (i.e. the laser is not too bright). A good length of fibre attenuates the signal. A meter or more probably works best.
The Torus should do the trick, so, placing the Cisco switches near the dCS shouldnāt be a problem 
2 Likes
Another question for @Anupc 
I have six wall outlets. Four of them are on a normal service AC line. The other two are on a dedicated AC line for my hifi. One of the hifi outlets feeds a honest (Ensemble) distributor which feeds in turn my SPEC amplifier and dCS Rossini+clock.
When I add an optical transceiver, should I plug it:
a) into the generic wall outlet
b) into the other hifi wall outlet
c) into the distributor where amplifier, Rossini and clock are already?
Thank you
A.
@Anupc As always thank you for the replies.
The quote above is classic. With your permission I am going to start posting it whenever I see the typical Audiophoolery claims posted on various threads.
I was wondering though - if I purchase a Chord Music Ethernet cable, or perhaps even one of those uber expensive Nordost ones, and use it to connect my iMac to the internet, do you think my online bank account balance will grow enough so I can get rid of my Rossini and move to a Vivaldi? May be worth a try 
3 Likes
My guess would be this one in order to keep it away from polluting your dedicated line
That would also be my best bet, but I have seen and heard funny things in this hobby, so I thought Iād ask some expert ā¦
A.
Sounds like the start of a Math joke
As youāve guessed, the main issue is that Media Converters typically come with cheap switching power supply units that can spew crap back into your AC line. So, do whatever you can to mitigate that issue, b plugging it into a different AC line, or by using a better power supply, or using a power conditioner for your DAC etc.
Be my guest!
1 Like
I am using a Melco S100 connected to my LAN via Meicord unshielded ethernet. Subjectively for me the Melco is an improvement over the Netgear FS108, so I would like to keep it.
The Melco s100 has an uplink SFP port. I am aware that I could link that to a Cisco 2960 to isolate the Melco from the rest of the LAN. Is there a good solution to achieve the same with a smaller footprint than the Cisco 2960? Maybe a TP Link MC220L? Which SFP modules and fibre (1m)?
Who benefits from bullying? The supposed in-group and an eventual clapping audience. Beyond that it creates significant problems for the cohesion of a community. In the end it comes down to what our motivation is. How much of our intent is driven by self-cherishing and how much by the will to benefit others.
Any Ethernet Switch or Media Converter with an SFP cage will work, itās all standard Ethernet, just pair them with identical SFPs.
The Melco S100 is actually a Buffalo Switch, Buffalo doesnāt have many of their own branded SFPs, which suggest common SFPs will work.
Trendnet Media Converters tends to be a little more reliable then TP-Link. Pair with any SFPs, multimode is fine, and likewise with any multimode fibre of any reasonable length.
Thanks @Anupc, I will look into Trendnet.
I see they have models with integrated multimode optical transmitters
- TFC-1000MSC (compatible with TRENDnetās TFC-1600 fiber chassis system)
- TFC-GMSC (not compatible with TRENDnetās TFC-1600 fiber chassis system)
Any reason not to chose one of these over the ones with SFP cages. I assume one would need to put a multimode Trendnet SFP module at the Melco S100 end.
I will be the first to admit some claims go too farā¦ but I am also willing to listen and see if they do in fact make a difference. Logically, ethernet switches, ethernet cables, power cords, even clock cables shouldnāt make an audible difference but Iāve personally heard that they all do.
The question then becomes one of why. Anupc explained why people who say PHYs can make a difference are deluded and wrong, and a designer of PHYs for decades says they can. Whatever.
The question is can you hear an improvement with the alternative equipment in the system? If so, I donāt care why.
If youāre the type of person who says āthis review said this is the bestā and is swayed by that, thatās one thing; Iāve not purchased more highly reviewed equipment when I couldnāt hear an improvement than I can count, and I also have owned much equipment that didnāt measure particularly well but sounded absolutely incredible.
Wadia Digital was the primary vendor along these lines and had Fine Sounds/McIntosh Group not killed their product lines, Iād still own their products rather than dCS.
Iāve stated before that at least for me, it took the combination of a Rossini and the Rossini Clock for the dCS sound to surpass that of a Wadia S7i; the Rossini alone did not and the Bartok certainly didnāt.
It would work, but thereāre no benefits, and I would not recommend it, for a couple of reasons;
-
Generally, those fixed configuration Media Converters are designed to work in pairs, so, itās not ideal if you want to pair one with your Melco S100 SFP port.
-
If you didnāt notice, the optical ports on them are with SC connectors, whereas SFPs are LC connectors. In order to interconnect one to a Melco S100 SFP, youāll need an LC-to-SC fibre cable.
-
Finding a Melco compatible SFP thatās fully compatible with the fixed optical port on those Media Converters could be tricky.
Best to stick with a Media Converter that has an SFP cage