2 streaming associated questions

1.  There are a lot of words typed here on the forum about RFI, but I suspect that evidence would show that those who expend any effort to reduce it among the customer bases of most hi fi brands is exceedingly small.  I do truly believe that this is a forum 'artifact' as opposed to what any large number of users experience out in the real world and do anything about.  My local hi fi dealer who sells many brands of digital streaming sources, including that $109,000 dCS system, do not have any recommendations to reduce RFI other than buying good quality cables.  I know for a fact they do not themselves, or encourage their customers to, install ferrite chokes along the length of cables.  (I am not editorializing as to the aural merits of this approach; merely trying to point up that the number of posts devoted to a topic here may not correlate to anything other than the number of posts devoted to the topic . . . )

2.  The 1's and 0's don't degrade; this has been established.  If the path from server to player 'picks up' something additional that the server didn't serve, then something is either wrong, or <see answer 1 above>.

As far as I  understand, it's the sp/dif connection that's the real issue as it is generally poorly implemented.

Optical would isolate the components electrically and would not succumb to RFI so seems to be the obvious choice. But reportedly (anecdotally?) people prefer coax. Most manufacturers seem to promote coax over optical too.

Just saying 

I think the RFI & streaming was more to do with NAS to Network to Player/streamer than to DAC

Howver I'm not sure optical has all the answers, especially in a small home network - other than optical IC -vs- Coax.  Plus optical networks are not that cost effective as far as I understand.   

100% agree on more words than enough on RFI, but thats forums. The interested & participating readers have/are all gone/going thru learning curves & it all takes time & - like the same old questions we get with subjects like speaker cables & Fraim building - it will all continue and/or start again in the future.

In defence of the many words from the ferrite-fan-club,  we should not forget EMI alongside RFI,  or how the effects of inter-pair crosstalk can be helped with it.

I agree 0's & 1's don't degrade, but I believe stuff like crosstalk can degrade the signal TX & RX process by adding interference.

Seems we all need to follow our own agenda in this area. I have friends with mid/high end systems that are very happy with low cost Cat5, some others are 180 degree the other way

One problem is this streaming network stuff is relatively new & comes with a new techno-babble language that many find a turn off. 

Another problem is changing an Ethernet from Cat5 to Cat7 or adding some ferrite is nothing like the obvious SQ changes we get by changing some cheapo speaker cable to NACA5

= read = it works pretty well without the fine tuning,  but little or no harm is done by using over spec'd cables & a few ferrites in the right places. 

1. IIRC it has been postulated here that the transduction necessary to make a modulated light signal into an electrical one puts load pressure onto the device and can elevate the noise floor. Plus, cheap plastic optical cables may also induce more processing (more errors ? less well-formed signal ?). Don't take that as a fact, it was just my understanding at the time of the discussion. So it seems to appear that, despite S/PDIF limitations it remains able to have the edge over optical, at least in some cases (and apparently many, unless you have other problems that could be cured by a non-electrical connection).

2. It has clearly been stated here, through experience but also directly by Naim I think (perhaps not directly on the forum but during a discussion at a bash with Naim personnel, I don't remember), that the S/PDIF output on the UnitiServe was added as a convenience and was not a primary tuning target. It seems to be the least good of any digital output on any current Naim device (maybe not the first UnitiQute, I have a doubt here). Not saying that it cannot sound good, only that it can be, and actually is, bettered.

More knowledgeable people will certainely add to, or amend, these recollections.

HTH

Maurice

I'm no streaming expert (it wasn't invented when I got my degree!) but the inherent 'share-ability' of data over a streaming network means packets are sent instead of a continuous train of data and there are safeguards against corruption (resends) that would not happen in a simple serial ' from here to there' digital system, obviously traffic is not pure 'audio' data, which may be deemed inferior in hifi terms but the ability to 'toss' away bad data is an advantage. Also you can use the system bandwidth to play different CD quality albums at half a dozen separate locations from the same NAS (try swapping CDs at that rate!)

One advantage with a networked connection vs. direct connection is that with a networked solution you can locate the 'server' far away from the hi fi; all you need is an ethernet (best) or wi fi (not so best) connection.

Question 1

RFI affects affects both digital and analogue circuitry, but does so in different ways.  S/PDIF over optical links doesn't conduct the RFI, but RFI still gets into the system via all the cables including the mains cable.  The real question is "Is there enough RFI getting in to degrade the sound quality?"  If the answer is no, then you don't need to do anything about it.

The problem with optical links is not so obvious.  The standard toslink uses a different type of fibre than the type used in telecommunications and unfortunately the optical emitters and receivers available for this are not of the same quality.  This lack of quality leads to jitter and that limits the sound quality.

Question 2

As others have said, the UServe wasn't optimised for S/PDIF, hence with a good streamer, the quality via Ethernet can be better than that via. the coax connection.

The UnitiServe uses the same Naim-designed single-zone PCI audio card as the HDX (and later NS01s); the big difference is in the power supply...also, it can be said that Ethernet streaming is asynchronous, so less subject to (but not immune from) issues of source (power supply) jitter than a direct SPDIF connection.

Thank you David. At last we have a clear technical reply on the quality of the UnitiServe's S/PDIF output. To my ears, the addition of a linear PS to the Serve removes much of the difference in sound quality between S/PDIF and UPnP playback, that can heard when using the stock switching supply. 

Jan

I suspect there are a number of SPMS in the UnitiServe that you'll never get rid of

The DC-DC converters for the 5V, 3.3V, 1.8V and possibly -12V rails.

There may be others, it's a desktop computer.

That's not to say that providing a lower noise external 12V supply isn't worth while; but there may well be a limit to the degree of improvement achievable from the S/PDIF.

Perhaps if you want to find the limit, fit a high current ultra low noise 12V PSU, then Hugo and find out!

I have only ever used the Ethernet output, and found a very significant improvement with the furry bedtime companion. Maybe it is bigger still if using the coax output.