Not only that, dCS uses among the best discrete Magnetics in the market for that coupling-transformer;
3 Likes
I’m not suggesting this. I’m pointing out the fact (and I think this is a fact we can agree on) that jitter and timing matters and servers are designed to reduce it and pass a better data stream to a DAC. Any DAC.
Don’t know about your Melco NAS - but current Melco Melco servers reclock the data. They don’t rely on an external clock input to do this.
No. Melco Servers have better linear power supplies, better disks, internal clocking, etc etc….
If you just wanted silence, there are a million different manufacturers make a silent NAS.
You bought into the very benefits of the Melco you’re so desperately keen to deny. The contradiction in your stance and your actual behaviour is undeniable…and yet!
Not one of the “primary ones” but a reason nonetheless. So I wonder what benefits of fibre you DID select its use over standard cat6 for?
As far as Ethernet streaming-based Servers are concerned, “jitter and timing” have absolutely nothing to do with music playback as it’s completely asynchronous.
In fact, there’s significant jitter involved, packet jitter that is, not the jitter associated with synchronous systems. And timing is a non issue as Ethernet streaming is fast enough to both error-correct and fill playback buffers (which are separately clocked-out synchronously).
Not when the Melco’s were first launched, it was the only fanless NAS on the market at the time. I’m guessing you’re new to most of this?
5 Likes
The ONLY fanless server? …no - they’ve been around a long long time. Melco didn’t invent them.
And no Anup - I’m not new to it al all - that superior tone you have coming through just a little there.
No - I have a long memory, and had a device called a RIPNAS Statement in 2009 - a silent NAS. Not much more than an ATOM computer in a fancy case with storage - actually - sounds perfect for you. Theres one on eBay.
When did you buy your Melco out of interest? …I’m guessing after 2009 
You’re once again trying the slightly redefine and narrow the question to suit your argument. Let me spell it out again in simpler terms for you - in audio systems jitter matters - a server that reduces jitter is a good thing, because it makes life easier up stream. You should know that, because you own one.
Seriously? ….youre still unwilling to accept that the fact you own and use one of the devices that you say are utterly pointless isn’t a teensy bit contradictory?
That the fact that you use fibre connections (now for non specific but definitely NOT for anything to do with noise reasons) is also a little conttradticroty to your stance?
I can certainly commend you for the mental agility required to hold true to both these positions in the face of your actual behavior. Do as I say, not as I do. Bravo.
Anup - with the greatest respect - you keep making statements inconsistent with your own behavior - and littered with half truths and partial science to support your newly arrived at view - I’ll keep pointing them out.
Or you could open your mind to accept that people who have made exactly the same purchasing choices you made - have done so for both good reason, and reasons that have some basis is good engineering practice. And we can move on.
Well, it sure reads that way from your post in this thread.
For a Server that is connected over Ethernet, no, it absolutely does not.
The fact that you still don’t understand the distinction between a Server that streams asynchronously versus the typical legacy audiophile discussions around reducing jitter explains why you keep insist[ing] a Server can make a sonic difference.
Until you fix your knowledge gap on that, we’re going to go round in circles 
4 Likes
Again, keeping things polite:
What I personally struggle with is if (a) we can all now agree that Ethernet is bit perfect, and if (b) dCS is telling us that they isolate the Ethernet port on the DAC from RF noise, then how or why would any upstream server or streamer matter?
(And for me, since I have yet to hear, to my mind–and pun intended!–a compelling answer to this question, I use a fan-less, SD-drive Roon Nucleus, and $10 spec unshielded Ethernet cables … That’s my choice, and I’m not trying to convince anyone otherwise).
6 Likes
But thats the crux of the debate. It does not matter; a $20,000 Server is not going to sound different than a $2,000 Server when both are streaming over Ethernet.
ps: hence why dCS shouldn’t bother getting into the Server business.
7 Likes
I “slum it” even more. Ripped files stored on a kitchen iMac located downstairs in the….kitchen 
It is not even an iMac dedicated for my audio system - it is used by the entire family for everyday tasks.
4 Likes
When you have no more points to make that don’t further expose your own glaring and staggering inconsistency, when your half truths and inaccuracies have been taken down, when your own lack of knowledge and blinkered nature is exposed, when there are no more cherries for you to pick and you can’t or won’t address any of the questions raised….just revert to character and be rude, dismissive and superior.
You’re very well known for it in these parts Anup.
But fine - I accept your resignation from this discussion on the basis of your inability to respond to any of my points. But I’m certainly here for next time.
EDIT: I have replaced the term server with streamer, which is the main subject of this thread.
The title of this thread is not a question, it’s a statement, namely “DCS Should back into the network server market”. No wonder it is causing a heated debate.
Let’s look at what may well be happening in this thread, which is a good example of what happens in many such discussions.
On one side we have @Anupc claiming that there cannot be a difference in sound quality when an external streamer is added into the equation.
In this post @Anupc has demonstrated with his Vivaldi setup, that the bitstream within the dCS DAC (before the DAC stage) is identical whether it is fed from Roon or via Mosaic (another heated discussion). In fairness I have to admit that the test setup did not prove explicitly whether the addition of an external streamer from brand XYZ does alter the bitstream or not, but very likely it will not, as all these servers claim to be bitperfect.
So “science” would suggest that there is physically no difference by adding an external server, as all the elements after the bitstream in the dCS unit (output stages etc.) are identical in the two scenarios.
On the other side we have @MattC who is sure he does hear a difference between the setup with and without an external streamer. Well “science” says that he does indeed hear a difference as well as do many others. The difference is heard and experienced by him, that is for sure. As for the reasons why he is hearing a difference there could be several. It could be that the two setups actually do produce a different sound, because something in the test setup is different, such as volume or indeed the bitstreams could be different. I would regard this as unlikely.
However what could well come into play is the placebo effect. The addition of an expensive piece of equipment (which carries a flowery description of what it is supposed to achieve) simply must sound better than the dCS gear stand alone. That is the problem with sighted comparisons. Placebo has been proven to work in medicine, even if the patient knows that he is given a placebo.
How does it feel to be accused by someone of wasting money, because one has succumbed to the placebo effect? Well, not very good. Does the fact that the perceived difference in sound quality is due to placebo change anything about the perceived improvement? Not at all. So as long as the buyer of the external server hears a difference, the purchase was worth the money. Same goes for a Taiko extreme by the way. Placebo or not.
3 Likes
Not sure what you meant by “addition of an external Server” 
… Whether you use Roon or Mosaic, you will need a Server on your network anyway, so addition of an external Server is a given. Unless you meant 1 Server versus 2?
That test set-up was originally put together to debunk HiFICritic’s contention about Ethernet Switches & Ethernet Cable, hence the dotted-line box “System Under Test” in the diagram. I subsequent used the same setup to validate that Roon and Mosaic are sonically identical when streaming the same locally stored file.
If you also notice in the diagram, the test actually involved comparing two different Servers, one running Roon, the other MinimServer, as my Roon Server is a dedicated machine with no UPnP Server Application on it. MinimServer was actually running on my Synology NAS (my Melco runs Twonky). So, I was in fact comparing a $4,000 Mac against a $800 Synology NAS.
That test actually objectively proves a couple of things, all at the same time;
- That Roon and Mosaic deliver the same bit-identical end-results out of the Upsampler - with no noise contamination of any sorts
- That Roon RAAT is no worse (or better) than HTTP (UPnP) as a music streaming transport protocol - not a surprise considering they both use TCP as the underlying packet transport
- Servers make absolutely no difference whatsoever - even when running different serving Applications!
The results speaks for itself, and the test setup is easily replicated by anybody (no need to take my word for it ) - if one doesn’t have access to a Vivaldi Upsampler, even a dCS Network Bridge will produce the exact same results.
4 Likes
Sorry however a post like this is totally uncalled for.
Anup has not attacked or been derogatory towards you in any way. You on the other hand are now resorting to personal attacks.
These are the kind of posts which take place on the Naim forum - we do not want or encourage them here.
4 Likes
Yes he has - he’s been dismissive of experience by suggesting I must be “new to most of this” and further suggested from his apparently superior position that I need to “fix gaps in my knowledge” and avoided reasonable challenges back - I stand by what I said. I think the escalations and provocations repeatedly came from him and not me.
I agree with the principal - but in suggesting that it’s my issue, and Anup had no role to play….well….sorry…just because you agree with him on this subject, doesn’t mean you should by default agree with his approach to discussion.
Hi Guys,
I’ve been watching this but keeping out of the way however someone has now flagged it so can I very politely and respectfully ask that we all virtually sit down round the campfire with a glass of our individual favourite beverage by our sides and have a Kumbayah moment.
I get that we are all very passionate but can we chill this one down a little please?
Thanks
Phil
10 Likes
Crikey, we agree on something! Let’s riff on that…
Clock accuracy really has no role in sound quality when it comes to streamed audio.
And here I find myself in complete and joyous agreement with @Anupc. This is precisely what ethernet was designed, quite brilliantly in my opinion, to do decades ago. The fact that it continues to deliver (take that as literally as you wish) despite huge technological advances in that time, is astonishing and a huge testament to its designers.
And this is where I find myself in disagreement with you. There is far more to science than digital science.
If anyone here is serious about understanding why/how a network server can affect sound quality, they must look beyond the digital. If they don’t, it’s simply an invitation for tropes to ve dusted off and wheeled out, particularly by those of us who live (or have lived) a substantial portion of our lives in the digital world and are therefore prone to being digitally blinkered.
I fully agree with you that there is more to reproducing music than digital. I have myself had a discussion with Anup about the influence of Ethernet noise. And I have a Melco switch and a FO decoupling in my network (just to be sure).
However, Anup’s comparison setup is looking at the bitstream within the dCS Vivaldi, in an environment where this bitstream is shielded from noise and clocked by dCS’s good stuff. So no opportunity for jitter to do its dirty work, nor for electrical HF noise to do damage. In fact HF Ethernet noise in the Vivaldi case would have to travel through the upsampler and then through the proprietary (designed by dCS) link between the upsampler and the DAC.
Unless of course, dCS have done a lousy job of keeping the inside of Vivaldi Stack clean.
If there were a difference in SQ due to (HF induced) noise, which the external streamer manages to keep away, then I think dCS should really go back to the drawing board. Then some fundamental things would be amiss with their design.
I don’t really want to get involved in all this but, I have experienced this at the recent hifi show i did.
At home I have spent a fair bit of time and money getting my ethernet connections to my Vivaldi upsampler as good as i can, for a sensible outlay. This involves a innuos phoenixnet switch, shunyata omega ethernet cable, a main switch doing all the house stuff, etc. In this set up my streaming is, as far as I can hear the same as playing a stored file on my melco.
At the show I used a much simpler ethernet setup, this consistent of a cheap wireless box, picking up the hotel Internet, a 10 meter cheap ethernet cable running halfway round the room, then into my innuos phoenixnet, and the same shunyata omega ethernet cable between phoenixnet and vivaldi upsampler. In this set up, there was a difference between the stored melco and streaming the same file. With the streaming version being slightly worse. It just didn’t have the same sparkle, it still sounded great, and we used it mainly all day, but it wasn’t the same as I get at home.
Is that just down to the hotel, or the cheap wireless receiver, the cheap 10 meter cable, all, or what ?. But it certainly wasn’t as good. Also at home I prefer the sound when I don’t run my ethernet into the melco, and then from the melco player port to the vivaldi upsampler, I much prefer the melco just connected to the phoenixnet, and the vivaldi upsampler also connected to the phoenixnet. Not sure what is happening in all the above statements, but it’s how I hear it, and not just me.
Cheers dunc
3 Likes
HF noise impacts the analogue output of the DAC.
HF noise doesn’t/can’t have any impact on the bitstream, so it’s hardly surprising that if you restrict your testing and analysis to the bitstream you will “prove” no difference. I would have expected exactly the result Anup reports.
HF noise doesn’t affect timing so won’t introduce jitter, so I’m similarly unsurprised.
HF noise could indeed travel via the route you mention. I’m a lifelong dCS user and fanboy, hence my involvement in this forum, but I wouldn’t rule this out as a possibility. When it is proven/demonstrated that this doesn’t happen, I’ll doff my cap as a salute to the extraordinarily talented designers at dCS. But until then, I won’t let only partially relevant theories get in the way of practice.
So what you are saying is that HF noise will not make it through a 2k Euro streamer, but it will make it through a 20- 60k (depending on model) dCS piece of kit?
HF noise can be measured, if it is of significant level. The streamer manufacturers (and anyone else) should easily be able to show measurements by how much their equipment attenuates HF noise at the analogue output of a high quality DAC.
Don’t get me wrong. I have also advocated the position in the past that HF noise influences the analogue circuits in a device. That may be true for low cost devices, but in the meantime I doubt it is for devices of dCS calibre.
For the time being I’ll stick to my theory of placebo, which is not a bad thing in itself. The differences people claim to hear are real to them. I hear differences too. Using the dCS as a streamer via Ethernet definitely sounds better to me than connecting a Raspberry to the dCS’ USB port. I also hear differences between USB cables. But that’s subjective and all sighted, not double blind. I can live with the fact that what I am comparing may actually sound identical if scrutinized scientifically. After all I’m in this hobby to enjoy listening to music.