A novel way to massively improve the SQ of computer audio streaming

[kyrill]

I may have missed the solution buried in this massive thread how to turn WS2016 into a music server. Officially there is no

media pack for WS2016, but too many have this OS in their audiophile server set up

 

I manage with bridging the following:

The red and the blue ports of the NUC are bridged with a fixed ip number in WS2016

the mobile phone ( with its simple DLNA player app) sees the UltraRendu and the other (windows 10) PC  connected to router and its library ( media streaming allowed)

But it does not see the NUC or its shared folders

Any help how to accomplish it?  When it works then I can apply AO 2.2 hopefully

 

 

[kyrill]

yes Gui mode

the other computer (PC) does  see  itself and the NUC under computers, when NUC is clicked the , NUC asks for its psswrd and name. I have not set that up

when the PC clicks itself it shows its shared folders in library. The PC sees also the Urendu under media devices

The NUC itself sees itself now under computers and  when clicked itself it shows empty folder

It does not see the PC under computers but under media devices. The NUC also sees the Urendu under media devices

The mobile phone sees only  the PC ( and its library) under media server and not the NUC. It sees the Urendu and plays teh PC library or shoutcast to the Urendu

1) with ping in cmd how do i find the ip number of the NUC in order to ping it? is it the manual ip number given to the bridge? That can be pinged

thx for the response

[kyrill]

Ok I missed "without the bridge"

 

minus bridge ( both E-ports removed from bridge, bridge survived) with dhcp from router NUC NIC gets a 192.168.xx ip number which can be pinged from both NUC and PC  but it is still invisible on the android phone

[kyrill]

Dank je Sandyk

 

That is very good advice

[kyrill]

My signal wires are DIY from finest OCC copper insulated by cotton sleeves which you can buy ( steep price) on the internet.

However they dont attenuate RFI and EMI.

 

SO I do another sleeve around it. You might ask "will that not negatively interfere with the low DK of cotton?" IT will be inaudible because of the betterment in sq with shielding The RFI/EMI shielding is  very very thin (conductive textile) so its DK is by itself because of that attenuated. Also because the cotton keeps it away from the wire. The more airy the cotton is the lower its DK  This is the shield I use. In audiophile land it is close to cheap. there are many different widths https://hollandshielding.com/Cable-shielding-materials-shielded-entries-materials. I never use the heavy metallic

 

I never use the heavy metallic flexible tube of audiophile death. .  although a transparent death. ( EMI and RFI attenuate strongly transparency) Its DK is a nightmare between 6 and infinity. A high metal DK gives a sharpness ( tiredness after longer listening) and typical HiFI" sound the opposite of  listening to acoustic signals around you. Non transparent systems how ever with "damped highs" may benefit from higher DK. It suggests focused and outlined "identities" of the different "players" in the sound stream  Replace it with the fabric, and when living under heavy RFI bombardement of mobile hotspots or towers around you, use two layers.

But a mobile phone  wrapped in a thin layer fabric  of less than a half mm cannot receive phone calls anymore.

 

[kyrill]

14 hours ago, ElviaCaprice said:

My experiment with JS 360 for signal cables is over.  This may be signal cable dependent.  Meaning, it all depends on where you are in the stream.  Data only, I don't think suffers from the same issues as a stream in motion.  

 

I found that the PCie ribbon from my NUC mobo to my PCIe 1X lane for my tXUSBexp card to make me suffer greatly in fatigue with the JS 360.  But it was the best transparency I've heard, which is what kept me at it, trying, to make it work.  But to no avail, it was like the timing was out of wack, slow motion leading to fatigue in speed??  

 

So I resulted in going to a triple cotton wrap for now.  I don't have an answer to this issue.  How can I enclose this ribbon from all RF/EMI noise without killing the sound leading to fatigue?   Big open spaced metal enclosure?  Which of course is not practical.   

 

Does this apply to USB or SATA cables?  I notice on taking my PPA SATA chord apart that it is double wrapped with metal braid.  This doesn't seem to lead to the great fatigue as in the PCIe ribbon.  But this only sends data and OS.  Music data is streamed from memory with JRiver.  I'll still need to do some testing with this.  

 

Quote

Honour your love for transparency, it is key next to musicality. To choose between the two is a disaster. Still regular metallic RFI/EMI attenuation adds transparency but at the same time because of their high DK ( dielectric constant) attenuate by adding a kind of subliminal harshness, musicality as well. I have your solution of a ribbon ( between NUC M.2 slot and Pcie card)  too. Happy to find one with low DK ribbon  insulation. Still It needs shielding and any good shielding in audiophile land is based on compromises.

Electrons ( both wave and particle properties) moving through a conductor need zero DK in order to please your ears. Zero DK is outside earth in a vacuum. Next best is air around the wire. But it starts at the wire. The more low DK is direct around the wire the lower the negative effect of a next layer of higher DK component will be. That is why a lot of wires have teflon ( low DK) around the wire but the the most outer layer is   still often made of PVC. A second  rule is, the thickness of the materials. A 0.06 mm layer with relative high DK sounds much better than when it is 1mm thick. The worst sounding ones are the5 0 years + old traditional industrial metal flexible braided tubes often >1mm thickness  you see in more than 50% of audiophile components...

So my advice is to use a shield around your ribbon made of conductive fabric* specially designed for ribbons BUT as even a thin layer of ( less than 1 mm) conductive fabric has a relative high DK compared to 1,5- 2,5 of cotton / teflon, you can lower it by using a bigger sleeve and use the space between sleeve and ribbon by carefully wrap the ribbon first with a thin layer of cotton wadding evenly spread over the ribbon. Try it out and you can even use a double shield if it sounds better. Instead of a ready made sleeve,.

 

a more economical approach is to use

flexible conductive fabric foil.  I would not wrap it so thickly as shown in this photo and still would wrap it first ( wire) in a cotton sleeve to keep away the the higher DK material from the actual conductor. I told more in this thread . A lot of something steely ( subliminal) or just heard around sibilants or instruments will be attenuated.

* carbon foil has conductive properties with cotton-teflon like DK

 

Quote

http://jplay.eu/forum/index.php?/topic/1223-the-battle-against-harshness/page-2?hl=%2Bkyrill+%2Bharshness#entry21637

 

 

 

[kyrill]

23 hours ago, ElviaCaprice said:

 

Thanks Kyrill, for your input.  I have read your findings a few times over on that forum.  I've ordered the following copper fabric off of ebay and plan to make my own sleeve over some cotton wrap.  

https://www.ebay.com/itm/EMF-RFID-RF-Shielding-Copper-Fabric-Roll-42-x-1-Conductive/173229145787?ssPageName=STRK%3AMEBIDX%3AIT&_trksid=p2060353.m2749.l2649

 

https://www.ebay.com/itm/Graham-Cellu-Cotton-Beauty-Coil-100-Cotton-40ft-Sanex/131346555189?ssPageName=STRK%3AMEBIDX%3AIT&_trksid=p2060353.m2749.l2649

 

Wish you success, it is not that difficult. One remark for the  copper fabric though. It will do it job very well, but in time copper will oxidize into CuO copper oxide and that is a bad conduct. The opposite happens with silver. Silver fabric is hard to get and expensive. Maybe the copper foil is treated with a chemical in order to stop the.oxidation or the blended in nickel has reducing properties. The oxidation will also happen at both sides of the thin fabric and in time the conductive function will get thinner and loses its anti RFI/EMI properties. The metallic looking foil or fabric does not have that property. You can also before using the foil, loosely paint it with a diluted  l(violin) varnish. That will keep the oxygen out.

[kyrill]

On 10/4/2018 at 2:30 PM, Mark62 said:

Thanks for the lead on the 3rd dummy wifi in the bridge.  I'll try that first.  And, thanks for the great summary post.  As I've tried to get into this topic, it has been extremely helpful.  I just recently got an ultraRendu and have been so impressed with its sound.  If it can get even better, I'll be quite happy!

hi Mark

you have two complaints

one NAS based and two Roon stuttering based

I can only explain complex things for myself and so to others by deconstructing the complex to the most basic simplicity ( which can still be complex as we shall see

 

What i have in home working after severe stuttering problems is this

one NUC ( OS on  Supermicro SATA DOM 32 Gig to avoid ssd) with two Ethernet ports ( NICs)   bridged.  One NIC to a WiFi router with a SUPRA 8 ethernet cable ( 6 meters) for internet and wifi (to smart phone to control what to play), and the rest of the private home network, the other NIC is connected to a UltraRendu with 1 mtr supra 8 ethernet cable.

The NUC is physically settled in the stereo system (DEQX + Jadis amps etc). I am a purist and found music stored on a shielded USB stick sounds noticeable more musical and transparent than any other solution I tried So no  ssd or nas for me. So main music is in my PC on terra hdds. so my music source

 

suggestion make the bridge work in its most simple form ( you need temporarily mouse and  monitor and keyboard for the NUC)

1) Make the bridge as simple as possible , make it work without stuttering and from then on expand the bridge

A bridge is within 1 computer  ok?

So one NUC has two NICs which are bridged.

You need internet for Tidal or other stations. So one bridge is linked to a router/switch with WiFi with FTP cable

the other one to the UltraRendu with Ethernet FTP cable

That  is all for the moment only to test

2) bridge the  NICs with fixed ip numbers incl gateway and 2 DNS servers numbers

Dont activate Roon. Download a free DNLA player for your UltraRendu from google store on your phone

3) set some music ( also 96 khz or 192 khz on a usb stick and plug it into the NUC. Add usb stick to library by Rclick mouse and choose add to library

Set UltraRendu to DLNA

From the android phone with for instance HiFi Cast you should see ultraRendu, the library and the music nuc as server and all should play

The player has a shoutcast  radio entree, try if you can get music from the internet. should work but mp3 sounds afwul

 

Tell me does it work?

 

Forget what is here below that is phase two when above does not work

 

I had   severe stuttering too.

what helped is  with trial and error change the properties of the bridge under

(A) control panel/network and internet/network connections

Rklick  properties   I can only tell you what I did . My bridge is between UltraRendu and Router/switch . They come together in that NUC having two NIC ports bridged. the router provided an internet connection as well connecting to other PCs 0r maybe in you case NAS?

What i canceled in  properties was  ( W10 PRO)

client for Ms networks and all the others except Ipv4 e.g. file and printer sharing  QoS and so on. when done click OK at bottom of page.

Dont start with Tidal. I put some songs in the music folder under "users" in the boot disk. i only want to test the bridge

 

If all play well to UltraRendu go to the next point (B)

if not find out which particular property you have to activate again. Roon may like QoS. I dont use Roon yet and can do without.

(B) Go To on same "page" to configure  of the ethernet card

you see a lot of tabs

1) Power management unclick all of them

2)Advanced        This is the tricky one . lot of choices. But this one resolved my stuttering. the principle is disable as much as possible. every activated property infuses latency and or extra jitter

Then we

 

[kyrill]

doc yes it can be a bug but on what level? So better for Mike to sort things out from the basics up.

 

Mike you can find the the dns if you, on any computer in the house or laptop that has  internet by DHCP ( so no manual configuring) goto network and sharing center/ change adapters settings /  OR

go the window where you NIC or Ethernet connection can be seen' DOnt click properties but Status / Details and then you see all the numbers. By the way a NUC  powered by an audiophile DC supply sounds better as a server than any PC unless  a dedicated audiophile  itx or microatx heavily modded and in an aluminum casing and fastly it becomes very expensive and even then how to avoid OS in ssd?

 

Keep us posted

[kyrill]

good thinking Elan

But being the perfectionist you are you can also replace the outer pvc layer ( ugly influence because of high dielectric constant=DK) with cotton sleeve ( be careful in removing it as the copper shield underneath it is not woven) and replace the heat-shrink tube ( also not an audiophile invention  straight from the electronic industry)  with cotton self adhesive tape.

I have done it with all my Gotham dc cables as soon as I heard with my 1st trial  ( second cable) that replacing it shortens the playing in of the cable by days and makes it sounds more soft like the Lush cable sounds more soft than the Curious usb cable. This softness does not cost transparency or definition, it enhances it and transcends it more into that difficult to define concept musicality. However more relatively muted sounding systems need an extra "bite' from higher DK to articulate the multiple separate sound identities

 

A worthfull alternative especially inside noisy internal RFI compartments with digital clocks inside is to use Lessloss new Litz DIY hookup wire.  The hundreds of more intertwinings of their symmetrical configured Litz wire compared to quad  cables  does not need 360 multiple  metallic shields with their VERY high DK over 10+

 

https://www.lessloss.com/cmarc-bulk-wire-and-cable-p-225.html

I have not tried it yet but their reviews  and theory behind it are promising

[kyrill]

THX M

 

Nice to see how intelligently you found your way through the audiophile dense forest with as many guides as there are religions

THe Rcores + LT 3045 is one of my newest discoveries

I have 2 questions

The Nimak Pulsar Mk 2 how do you use it as a USB cable? You begin your paragraph with " In terms of USB I have found a cable ..."  Have you compared it with the Curious and or "Lush" usb cable?

2) The modified Ghent 360 JSSG DC cable have you compared that one against PoE?

 

And again thx for the detailed last post

[kyrill]

Doc  there too many things that can be heard what contemporary methodology cannot measure yet. Only when subjective hearing persist in hearing a phenomenon clever technician find a  new measurement technique which sheds light on the subject.

The beginning years of the  digital CD sounded awful while all measurements swore that bit perfect should sound ehh perfect. Nobody knew at that time the core effect  jitter had on listening. Then it took decades to understand that digital is a thought category and not an electrical one. Digital is just analogue with a special kind of waveform. carried by old analogue physics:  current and voltages . Music signals are pleasant and consist of mostly sinus waves. Digital is ugly music made of square waves. In that sense they are both analogue only teh carried information is different. Now it is normal to know that all what plagued analogue musical streaming ( EMI/RFI, Necessity of damping, dirty current out of the wall, sound of cables and so on) plus jitter problems analogue does not have  also plague digital with the  exception of lower noise floor and "perfect" copies. Also terrible sounding is data correction. perfect for storing not so in real time streaming.

In general you cannot measure factor X if you dont know beforehand  in what category X falls.

[kyrill]

Does nuc properties accept 15V?

 

[kyrill]

Is it possible under Audiolinux to bridge two Ethernet ports being the only OS in the NUC?

 

[kyrill]

Greenleo thx .

 

question about the last picture. I understand "do not click start"

The option then is "close" I think, but you suggest write finished? or is it it "close"

after the writing is finished and that is the case when the start button is not grey again?

[kyrill]

you can bridge with a NUC by adding a better ( JCAT ETHERNET card)  via its M2.socket with an audiophile M2. pcie adapter. Then you have 3 NICs with two clocked by a femto clock. That is what I did. The NUC as server for the UltraRendu.

[kyrill]

Ah that is the question this topic should have answered. is network bridge possible in w2016 core? I have w2016 GUI and that is easy

I have not found what demands can be done in core for that outcome.

 

With audiolinux headless bridging apparently is possible and for audiolinux apparently the consensus is it sounds better than w2016 core.

 I am now building and taken care of my dc power and its cables incl inner dc cabling  ( sotm usb ultra for instance) and inner signal cabling in my tube amps. EMI/RFI is a real gremlin. After all that audioliniux hopefully has a manual for dummies

 

by the way,  Peter, could it be possible w2016 gui +bridge may sound better than core without one?

[kyrill]

I anticipated that see http://jplay.eu/forum/index.php?/topic/2473-intel-nuc-alikes-which-accept-jcat-femto-ethernet-card/page-2

while most cards are around 10 dollars or even cheaper this one is 32 dolars, i bought three different ones, two mainstream and this one. the quality difference in materials was immediately  obvious

 

under the brand name you can see where to order

[kyrill]

yes, good bet. but linux  headless is awfully user unfriendly when you rely on "intuition" ,  still made by programmers for programmers especially when leaving the known paths of routine processes to explore boldly the unknown, where no man has been before. Precisely what this thread is about.

[kyrill]

8 hours ago, elan120 said:

Big thanks to both @austinpop and @lmitche for this Gotham JSSG360 cable idea!

I am slow in adapting this idea, but finally made the same cable, and think this is a great upgrade.  Now I need to find time making a 24pin ATX cable as well.

 

Here are few detail pictures I took during cable making:

 

Pins crimped:

 

 

Pins inserted to connector prior to heat-shrink cable shields:

 

Both cables completed:  8 Pin cable installed:  4 Pin cable installed:

The devil is in the details Elan. The last .75  inch is naked ( no shield at all)

the .6 inch copper shield before that has wires underneath stripped of their individual shield 

So Real 360 shield ends where the blue pvc outer jacket stops. So there is actually a big "leak" in the shield where RFI and EMI slips in. There is metalized fabric in the market  a kind of cloth that you can wrap around it before applying the heat  shrink. The difference is easy to hear

 

 

 

[kyrill]

There is another BIG disadvantage of the fan (s) at least as bad as adding noise: vibration.

Any moving metal in a magnetic field will induce electricity. And vice versa. moving electrons in metal

will induce a magnetic field.

[kyrill]

100% is not necessary. That purity is in no sense all sense input has "noise" or impurities  Our brain filters it out in "constructing" our experience out of it. A good example is our vision. even people with 20/20 vision what actually falls on our retina is a far cry from  a good camera image.

[kyrill]

pff  Tidal 96khz+ is on my not to miss list, like Audiolinux is...

[kyrill]

" But Jussi's image wins on performance for me - when playing new tracks there are pops just at the beginning of tracks with Audiolinux (and nearly every Linux distro I've had running NAA before). "bold added

 

This rather alarming for me. Do more audiolinux users have these pops or clicks with teh beginning of a new track?

[kyrill]

Hi Larry

I cant help you.

I don't do upsampling.

I don't do NAS

 

My endpoint is a modest modded UltraRendu. My NUC is the server.   7th generation I5

I want to keep it  as simple and pure as possible. My DAC is a DEQX  HDP 4 192 khz

at the moment W2016*  GUI + AO 2.2 on a 32 Gig Supermicro SATA DOM (own Hynes 5v psu ) (is not an SSD),     bridged Ethernet ports  with JCAT FEMTO netcard and state of the art DC power for each component + DC silver quad Litz wire with extra shields

* because of this forum I will try Audiolinux + bridged NICS + Tidal+ own files on a separated powered USB stick

because of Audiolinux I was worried about the pops remark

Forgot to mention  tX-USBultra,  and Uptone isoregen between UltraRendu and DAC (DEQX)