The degree of added value of an expensive high-end network server partly depends on the method of integration into the system. Absolutely no regrets over here after investing in an Aurender w20se, now directly into my Vivaldi Apex with dual aes and use of Vivaldi’s wordclock output to sync it’s clock…
1 Like
In my experience cables and servers do make a sonic difference when using an ethernet connection (and USB for that matter), but only because of their ability to reject or admit high frequency noise which rides on top of the ‘perfect’ data packet and gets into the DAC raising the noise floor. Like many on this forum I have experimented with NAS drives, data switches, cables and optical links and ultra clean DC supplies to power media converters. The data packet arriving at the dac is the same in each case, what differs is the amount of noise that is, generated, passed-though or rejected by the upstream hardware. It is quite likely that the noise rejection mitigations in the Vivaldi up-sampler are so effective that noise from upstream hardware never makes it through to the Dac circuitry, which of course would then support the notion that servers make no difference.
2 Likes
This is patently untrue. The two different servers can allow different amounts of conducted noise to travel along the conductors or shield alongside the data. Ethernet is ethernet and is bit-perfect, error-corrected, jitter-free etc, we all know this.
As we say in the courts here in England, the truth needs to be “the truth, the whole truth and nothing but the truth”.
Your scientific fact is indeed fact, but it is not the whole truth.
1 Like
Absolutely this. Thank you.
I do get tired of people insisting that ethernet is ethernet, as if we hifi and music enthusiasts were interested only in data integrity. Some of the most cynical and uninformed comments on streaming come from those who work professionally in corporate networking. RFI noise is simply not an issue in the corporate world, all that matters is data integrity, and in that specific respect, @Anupc is more knowledgeable than many. But hifi systems connect to analog(ue) components where noise is anything but a non-issue; indeed, it’s the only explanation for the self-evident differences in audio performance of different network servers.
1 Like
Oh dear, here we go again. How many times have we been over this Nigel; noise doesn’t travel over fibre. 
4 Likes
Oh dear, did we limit the conversation to fibre at some point? I’m sure I’m not the only person who missed this.
I also missed the part where your own responses said you were referring only to fibre connections.
Agreed 100% that noise does not travel over fibre. I have always said this.
Noise can and does however travel on conductors and on shields, and I’m pretty sure this is why people are describing differences they’ve heard.
1 Like
No. That’s just in my case.
Yes they do….
HOWEVER, when two different Servers are connected to an Ethernet Switch which also has a Streamer connected to, any noise generated by each Server has absolutely no bearing on the Ethernet port that the Streamer is connected to.
Even the most basic Ethernet Switch will ensure that, no one needs to buy a Reiki Switch 
5 Likes
You are patently, audibly, demonstrably wrong on your first point.
You are also patently, audibly, demonstrably wrong on your second point. My own points are not and never have been about my own company’s products, award-winning that they are. 
You asked how many times we would have to have this conversation. We will have it every time you make sweeping and inaccurate generalisations.
1 Like
Media converters generate noise. Converting light pulses back to electrical pulses requires power (DC usually which has to be rectified from AC) which can be a source of noise while the converter processing itself may be inherently noisy. In my experience using ultra clean DC supplies on the media converters ahead of the DAC yields obvious audible benefits.
2 Likes
I would like to keep this discussion on track, politely, please.
It sounds like their is broad agreement that Ethernet is bit-perfect and we don’t have to worry about loss of musical information along the cables. If this were not the case, the world wouldn’t work they way it does today, as emails would be missing words, and financial transactions would be missing $/€ etc! I’m glad that matter appears to be resolved.
There also seems to be agreement that RF noise is not able to travel along fiber, and for those who have a Ethernet-to-fiber signal pathways, as @Anup does, then the noise issue is moot.
This leaves, in my simple mind, two downstream issues to explore for those of us who don’t have fiber in the chain, and use a pure Ethernet pathway into the DAC:
(a) Can RF noise travel along the Ethernet cable with the information packets?
And
(b) If (a) is Yes, can such noise negatively impact the D/A conversion process and/or the integrity of the output analog signal in a dCS DAC?
Is this at least the right (or one right) way to frame the issue?
Thx, also for any thoughtful responses!
Richard
4 Likes
It’s really quite straight forward; by design, the Ethernet physical layer (the IEEE 802.3 standards family) specifies strict performance requirements for the Physical Layer (media, cables, connectors, etc.) to ensure reliable operation despite noisy, to both physically and logically isolate and abstracted from the actual communication stack above it. By the way, Fibre connections aren’t a necessity by any means except for reach and/or industrial-use where high-current interference are common.
More important though, the work that the IEEE does is not just theatrical, it’s practical and we live it everyday. If an Ethernet Switch didn’t do what it does and instead allows different Server noise to impact every downstream device differently, the whole World would come crashing down pretty quickly 
3 Likes
Anup, you know I greatly value your contributions and that you and I are aligned on perhaps 90+% of the forum topics. As such, I’m not challenging you, but would still like to get to what I think is the heart of the matter.
You write about IEEE standards and as a (very bad) engineer myself, I understand this, and agree that these standards would stipulate very precise technical and operating requirements that would need to be accepted and tested worldwide, and verified by third-parties to make sure they work. That makes a lot of sense to me.
What “the other side” would say, I think, paraphrasing, is that a corporate Ethernet environment has no concern for sensitive analog equipment on the other end, and therefore, perhaps, this design criterium is not in the IEEE worldview.
As such, I think the question about RF noise (a) traveling along the ethernet signal, and (b) impacting sensitive D/A conversion equipment downstream remains open and valid and if the simple answer is No, this is not a concern, then I, for one, would greatly appreciate understanding why not, or why yes, either way.
(This comes from someone who has followed dCS’s advise and uses only spec, unshielded, $10 BJ ethernet cables!)
Thx!
2 Likes
To be honest, Audiophiles imagining that HiFi Audio equipment is the most sensitive when compared to commercial platforms is really quite hilarious.
There are medical and industrial applications of Ethernet with far more sensitive downstream elements, and in some cases involve life & death.
This is not rocket-science, it’s dead easy to measure the impact of the data stream coming in from an Ethernet interface. None of the “Super Server” vendors can show a single example of how they can improve sound quality.
1 Like
Having ZERO engineering experience my comments will be extremely short.
When it comes to noise traveling down the Ethernet cable and reaching our dCS units, potentially harming performance, I defer to dCS themselves who have addressed this issue. They have stated our units have been designed to not be susceptible to this issue and perform as expected.
Enough said to satisfy any concerns (for me)
PS - I was going to mention something about galvanic isolation however I think I will leave it as written above
4 Likes
As per the way the world, this thread demonstrates perfectly that some individuals really struggle to come to terms with the fact that other people have different experiences and therefore opinions, and in attempting to get people to bow to “the truth” actually undermine their own credibility as a voice of reason on other matters.
Social media has enabled those that think its ok to be rude and dismissive of other experiences. It’s a shame - but we all know something has been lost with how people politely interact, debate and discuss since the rise of social media.
I have a music server. It makes a big positive difference to the sound of my system, because it’s designed to be better at keeping unwanted noise out of the time critical music chain than a normal NAS, and my experience of living with both it’s pretty black and white better having it in there than not. It was worth the investment to me , and there is no quasi science underpinning why it’s better - it makes perfect sense and delivers.
There are a lot of people who have the same experience - and it’s very simply and easily demonstrable. Those that disagree and think we are mad, delusional and “subjectivists” (what a strange term - aren’t we all driven by subjectivism? …how else do we choose our speakers?) then fine - but it’s you that need to deal with it or just ignore it, as repeatedly telling us we are idiots isn’t the way,
4 Likes
Science & Facts vs Beliefs
=> what can go wrong?!!!
2 Likes
The preceding posts have been technical in nature. No one is attempting to get people to “bow to the truth”. Everyone here is a big boy and girl and are free to spend their money as they see fit and come to their own conclusions. Likewise others are entitled to put forward alternate opinions as well.
Next…..
2 Likes
You’re clearly reading a different thread to me, as there is definitely a view being offered that there is no room for opinion. If that’s not the definition of “bow to the truth” I don’t know what is.
Most certainly. And, some of the audio we play back every day has passed through network interfaces using Dante or RAVENNA.
1 Like