AudioQuest Ethernet Cables

Interesti - it could be a limitation of the ND5XS of course. I assume you are sending PCM and not FLAC encoded files to the ND5XS?

Simon

I'm using Synology's server application to transcode the FLACs to WAVs before sending them to ND5XS. I'm not saying the disks sound noticeably worse via ND5XS, just different. A bit more hi-fi, more sparking clean, less "bloomy", smaller soundstage (comparing to direct CD5XS replay). But then I had identical observations when comparing CD5XS analogue outputs with S/PDIF going to ND5XS... So it may not be the ethernet piece that's causing this. Still, won't hurt to try that cable. I just repositioned the router a bit in the living room and now I could get away with 1,5 meters of AQ Vodka, which is something I could probably afford in a salary or two...

The drives in my QNAP NAS are relatively noisy, so I keep it well away from my audio setup. It's connected to a local switch in a "data closet", and from there I run a 30' Cat6 cable in-wall to the NDS in my listening room. I don't even use a wall plug -- it's simply fished through a small hole in the baseboard. I did follow Simon's advice, and added a few ferrite chokes to the end of the cable, very near the NDS's ethernet port.

As an experiment, I tried connecting the 30' Cat6 cable to a small switch in my listening room, and then ran a 4.5' Cat7 cable from the switch to my NDS. I also moved the ferrite chokes to the Cat7 cable. I could not hear a difference between this and my original setup, so I've gone back to a single 30' Cat6 run.

Hook

PS - Bought the Cat6 on sale a few years back. The price was close to Cat 5e, and I figured the extra twists per inch might do a better job of fighting RFI.

As I gathered from several folks' experience here, changing Cat5e to Cat6 or Cat7 doesn't do too much in terms of sound quality... Now if you repeated this experiment, but for that last 4.5' used a high-end AudioQuest cable, you might be surprised. Or not. But several members of this forum WERE surprised...

Hi Marcin -

Perhaps, but it would nice to understand why that is so.

I doubt it can anything to do with the actual transmission of network data packets.  Just logged into my NAS, and I see 222 million sent and 181 million recieved without an error (it's a relatively new QNAP, used almost exclusively for listening to music).  So while I grant you that AQ uses very nice materials, and while the cables appear to be well built, they would also appear to be serious overkill in terms of delivering high-speed, bit-perfect data transmission according to the Cat7 spec!

That leaves RFI rejection as the only area of explanation for improved audio quality that I can think of, and I suspect I'm already getting very good results just by using inexpensive ferrite chokes.  It isn't a huge day and night kind of difference, but adding three chokes did bring a small, but noticeable improvement in audio quality.  

If I am understanding the AQ product line, Vodka is the first level at which they attempt noise rejection (beyond Cat7 shielding). This might bring some small incremental benefit beyond simply using ferrite chokes, but if it does, it does so at quite a high cost. TBH, I am skeptical that there is much more to be gained in this area.

If the opportunity to demo one of these AQ cables presents itself, I may take a shot. But it is just not a high priority area for me.

ATB.

Hook

Does this all mean that wireless would be the better option to stream music if there was no drop out?

SJB

Hi Hook,

it's best not to spend too much time rationalising where these things are concerned. Just borrow a some and listen. I don't think AQ understand why they make such a difference.

Keith 

Quite right Mango.

Not necessaily. Yes, you eliminate the "antenna" effect from connecting a wire, but the processing of wireless signals may itself generate considerable noise.

ATB.

Hook

Peace of mind?

It's a trap.

This is what I'm thinking, i.e. replace the Ethernet cable from Airport Express to NDX.  What about from the NAS to the AirPort Extreme?

Bud

The only issue is that the airport express will be polluting the mains. ;-)

Huh?

Allen, ok I see - yes I did mis understand you. I can't see there would be much of an advantage, and generally I think its better to have longer ethernet patch leads - but you never know....

Having a separate SSID for effectively point to point wifi network will clearly help at collision avoidance at the data frame level, and so should provide the best throughput. However at the physical wireless level the SSID still needs to share the spectrum with other SSIDs unless there are clear channels and so if many wifi networks around the benefit might be reduced.

As I have said elsewhere when and if Naim support single/multimode fibre ethernet connections - then  many of these issues and side effects we sometimes see and hear with EMI will disappear. Fibre patch leads are more costly that Cat5e but cheaper than many boutique wired ethernet patch leads as far as I can see. However as night follows day I am sure someone will 'invent' an audiophile fibre patch lead as well as and when fibre gets adopted.

Simon

Allen - indeed i got you - my point was it is IMO better to use longer ethernet wired patch leads than shorter ones for our audio uses.

So what are you trying to achieve using the Apple Airport device as a bridge? The link ultimately will not be as reliable and will be slower (and be half duplex) compared to a fully wired connection.

Simon

Originally Posted by Hook:

I doubt it can anything to do with the actual transmission of network data packets. Just logged into my NAS, and I see 222 million sent and 181 million recieved without an error (it's a relatively new QNAP, used almost exclusively for listening to music).

Yes, but the question is: does the NAS software report the No of packets sent/received after or before any error correction that may happen along the line and put a strain on the CPU?

Just thinkin' aloud...

It does not matter, TCP handles it & if MTU (maximum transfer unit) size is 1500 bytes & if one packet from millions is resend you can forget this if the network is really * really * really ^ 100 rubbish. You can get this same statistics also from router/switch = same story what Hook already mentioned. Packets failing is not an issue in home networks.

How should I prioritize the following connections using Ethernet cables?

NAS to router.  UServe to router.  Airport Express to NDX?  Modem to router?

Bud

Gandalf_fi - you are right for TCP which is what is used in the media transfer, but remember UPnP uses the multicast IP addresses as well, and since this is UDP it is fire and forget there is no resend. UDP discovery messages occasionally disappearing will make UPnP servers/control points disappear and reappear. So it makes sense to have reasonably reliable network for the UDP. (I have experienced some reliability issues with UDP with some home wifi access points)

Simon

Yes, good point. I just said that typically wired home networks are working like trains toilets which can be seen also from router/switch metrics & if no resends then no worries. In wifi case this is totally different due it is quite unreliable on many cases.

Worth auditioning Cat7 before spending that much on Vodka?

The Audioquest Cinnamon arrived two days ago and now runs between the switch and NDX in my system. Struth! What a difference. No ideas as to the technical explanation for this but my scepticism has been banished. Halibut commented on a smoother more natural sound, but my results are quite different. PRAT! The scale and tonal balance of the system are just the same, but it now has an infectious rhythmic clarity. In addition to this, are improvements in texture and focus. It's as if the system has gained some of that addictive rightness of the Nait 2. I'm sure many here will know just what I mean. This is a good thing.

I'm tempted to say the difference brought by the new cable is stunning, but I complicated things after one day by adding a Powerline to the NAPSC. Now for two "upgrades" of supposedly rather dubious merit... the difference is outstanding. I never knew there was so much music below 100hz as the tune and texture of the bass is now more articulate. I can actually hear the vibration of low bass guitar lines rather than just a relatively amorphous diffuse bass as was the case previously. Hard to say how much is due to the Cinnamon and how much the PL, my suspicion is it's mostly the Cinnamon (and potentially all), but for a total outlay of less than half a Flatcap, I've probably never spent my money better. I'd say the change is easily greater and somehow seems more fundamental than last years change from the 122x to the 202. One more note... this just furthers my admiration for the NDX. What a machine!  

I think Gajdzin is doing his A/B comparison today, which should be interesting.

Keith 

He was SUPPOSED to be doing them this week, but his guinea pig (brother Paul of the Golden Ear) is down with flu. Half the town is down with flu...

My own listening tests reveal that I can't hear the difference that ONE cable produces. However, I can very clearly hear the difference that changing a lot of cables produce There's absolutely no doubt that adding 3 PowerLines, 1 Hi-Line and 1 AudioQuest Vodka to my setup produced a massive change. But just swapping the Vodka for an old Cat5 is not something I can hear, or at least I think I can't hear it - the fact I know the price difference between the 2 cables colours my judgement too much... My brother won't even know what has been changed and has never even heard of "audiophile" ethernet cables. So he's the right guinea pig for this experiment