Digital audio has evolved to be facilitated/controlled by computers, which must communicate with one another, some across the world. The easiest and most common way for them to communicate is wired ethernet, which unfortunately is probably the worst for sound quality (compared to fibre and wifi). So in simple terms, the audio reproduction industry has chosen to use the hardest thing to solve (wired ethernet - probbaly based on the network experts view the ethernet is perfect), and we audiophiles are trying to solve it one cable and device at a time. No wonder this pursuit will seem endless.
Consider taking a different tack.
The proverbial penny dropped for me when I did the following:
substituted the Devialet Pro integrated streamer (Core Infinifty) using a super wired connection with an Auralic Aries G1 using wifi
substituted the wired SPDIF connection from the G1 to Devialet Pro with a very good toslink
Both these were ‘ah-ha’ moments. Lessons:
the wires collect noise
keep the noise out of each device and do it well.
dont rely on the integrated galvanic isolation.
But, we must use wire sometimes, there is no alternative, so choose wisely.
It is quite possible the PheonixNet was simply illuminating problems elsewhere in the network that you did not like.
Yes, I am fortunate that in Australia we have nbn which in my case delivers internet to my premises by fibre. It is mandatory to use the nbn termination device (which above I called the fibre internet service device for simplicity).
Thanks for pointing me there. Much appreciated. You do the community a great service by taking the time and making such a great effort to share what you know and are learning as you go so freely and comprehensively. No ego. No hidden agenda. No hubris. No BS. Not sure you get thanked enough so i will say it again. Thank you.
Thanks for reporting on your thoughts and experiences. It all helps us to work towards better sound.
It is my understanding that wifi can be good but all wifi is not equal and to do it properly is not easy and is expensive for the device manufacturer. I’m not quite sure how we could tell which devices have done it properly and which haven’t.
Likewise optical is not a universal panacea to eliminate noise. At least one well respected streamer manufacturer has rejected optical for its top of the range devices on the grounds that the supplementary circuits/boards required for optical generate more noise than a well implemented wired connection such as usb or spdif. My own experience with a K50 and Oladra is that optical from either of these is inferior to bnc spdif.
It is a freakin mystery which products do it well or not. Both the Aries G1 and Devialet Pro use plug in type wifi modules. Probably off the shelf from a manufacture, but carefully chosen. I suspect the key is how and where the noise is shunted (presumably to ground) and perhaps fine tuned firmware.
Firstly, the lesson I learned was that toslink was less prone to conveying noise. I translated that to focussing more on fibre ethernet. But I agree toslink could be done poorly in some cases, even though some manufacturers rely on it or something similar (I recall ST fibre is fundamental in Bel Canto Black, and there is MSB ProISL Interface).
The biggest lesson I have learned it to not dismiss any format due to its reputation. I had dismissed wifi on that basis and was frankly shocked by its performance when implemented well. Now I have a dedicated wifi access point to deliver data from my EX (server) to Devialet (streamer). To take it another optimising step , it is very low power/gain, and powered by battery which has Akiko USB stick sapping more noise!
So I try to listen for the potential/best qualities of something and then tinker to see if the not so good qualities can be improved, and sometimes the latter turn out to be somewhere else.
With toslink, for example, I tried different cables, and even heard an improvement cleaning the end of the fibre.
Thanks for your recent contributions. Just jumping in here, maybe it was mentioned elsewhere but since all is in the details;
does your ER have the SR ethernet cable on the A or B side plugged in?
Synergistic Research cable with(out) bullet and/but with ground ?
does ER have a better LPS and ?
a better DC power cable?
a clock?
you confirmed ER is grounded
Or did you possibly compare a few of above options?
Bonus question, if you don’t mind: The SR cable is grounded to a SR ground block and further connected to an unfiltered/unconditioned exclusive AC circuit?
Fibre to the premises > fibre termination device > SR Atmosphere X Ref > ER Side B - Side A > SM fibre > CRS 305 (router) > SR Atmosphere X Ref > Antipodes EX > Shunyata Sigma > Mikrotik mAP lite — wifi — > Devialet Pro.
Both SR cables have the bullets in, and use HD ground cables.
The ER is powered by Antipodes LPS and its attached DC cable (from an Edge) with stock power cord. No external clock, but grounded with a basic SR ground cable. The ER has 2 SR ECTs inside and unused RJ45s are capped with Audioquest and AQVox caps.
The ground block is plugged into the ground input of an ifi Audio AC iPurifier which is plugged into a wall outlet of a general household power circuit (not dedicated, isolated etc, actually I think the fridge is on the circuit). That is the only AC treatment for the ER and fibre termination device, very basic.
The SR cable was preceded by trying quite a few cables, from cheap Cat 5e, commercial grade Cat 6 and 6a UTP and S/FTP, Elecom Cat 8, Wireworld and JCat cables.
It is a messy experiemental prototype pretty much thrown together on the fly, but the prototype is motivating me to redesign that bit of the network properly (better power sources, cable layouts, etc).
Next I am inclined to interrogate the fibre termination device (change its SMPS, put ECTs in it, etc) which is risky because it is owned by the national network provider.
“God is in the details”, good SQ is in the details. Thank you, it definitely helps me to reflect on my network. Apparently we (did) use some similar parts. I do understand it is evolving, great! Do you only stream or play local files (NAS or SSD) as well?
The second SR ethernet cable and Shunyata Ethernet cable can not be called details though . I wasn’t aware, but I thought there was maybe more to it to obtain your great results. Glad I asked.
Being the last downstream connections the Mikrotik CRS 305 and mAP lite are on a separate PSU? or respectively included smps? Hopefully at least as good UPS as a S60? and connected to the audio AC circuit? (Not the ER S60 psu or the kitchen AC circuit?)
FWIW some things to consider for ER: Do I understand correctly the DC cable for ER is the standard 50cm cable that came with the S60? One can improve on canare 4s8 4 x 16awg Star Quad, for much less $ (than for example one of the SR cables) but probably very audible. I would think it is even more beneficial to pay attention to DC and cables for both Mikrotiks.
I am sure you compared the usual Finisar and Planet SFP modules?
I do not understand why you isolate twice. Once with the ER moat. Second with the microtik 305. You insist on fibre I guess and the
SM fibre > CRS 305 (router) > SR Atmosphere X Ref
must make a substantial difference? and you need a router there ? Not only a switch.
I just started using the mAP lite (thanks Emile ) as well but to access the audio network w/o internet connection. I am going a different way, now eliminating wires and cables as much as possible and giving up on fibre isolation.
Audio-router only (Ubiquity + good LPS) + mp lite
l
Switch (swapped ER for phoenixNET)
l
Server/Streamer
I only stream. I removed the SSD and its wires from the EX to eliminate that as a possible noise source so I could properly assess the impact of ethernet noise.
Yes, the CRS 305 is powered via W4S PS 1, SR Atmosphere Level 3 digital power cable and my principal power conditioner (SR Powercell 12 SE) and mAP lite powered by a USB power bank battery.
The LPS is ODAPS which has a DC cable attached. Yes DC cables are on the to do list. I plan to reorganise the network to swap ER and CRS 305, or put another ER after CRS 305, and power them each with seperate batteries for total isolation from AC and each other.
Not yet, I use Cisco SFP with Afterdark noise reduction. There can be issues with compatibility. Part of reorganising will be possibly moving the national network device to the hifi room and using SFP+ Active Optical Cables.
I isolate 3 times … MOAT, fibre and wifi because each one provides improvement. I conclude/suspect each ethernet device is affected by noise that has an affect on SQ. I also isolate the rest of my household from the CRS via fibre and then wifi.
The CRS 305 is my only router, because the 1G router I had was a weakness and the 10G sounds so much better.
Agreed about eliminating unnecessary wires and cables. I only use them when there is no better/lower noise option, and then aim to use only very lower noise wires and cables.
It is really great to see my similar thoughts and advancements regarding an audio network for computer audio reflected in your approach. I also use(d) a SR Powercell + level 3 AC cord btw. Besides enjoying the SR Atmosphere X Ethernet cable and SR grounding principles and the ER and a multitude of network and AC cables and plugs and what else more besides two dedicated AC lines.
Did you isolate your audio network completely from the home network? Hard- and/or software isolation? Does he modem or ‘‘fibre termination device’’ split ethernet in dedicated home and audio ethernet streams?
That first device definitely can benefit from a better psu, if it is on a smps now. I have a Farad Super S shared between modem and a switch. The latter physically splits in 1) home (WiFi AP) and 2) audio stream for streaming and updates and connection the other UPnP devices in the home. But I also have the option to disconnect internet from the audio stream and use only the Ubiquity X audio router + mAP lite + switch + server /streamer.
In my quest to eliminate the latter and not be able anymore to distinguish streaming (Qobuz & Tidal) from local music, my current ‘problem’ is the replacement of ER by phoenixNET and now local SSD sounds absolutely better, I will give it more time though.
So today I ask myself did my elimination of (network) cables not pay off? In the next months I probably will re-introduce ER, fiber, SFP modules and so on but at the same time hopefully not.
I have to admit I am a firm believer in local SSD + good power being superior. No batteries. To raise the bar even more I should install the (in this Antipodes forum ) lauded Samsung PM893 SSD in my server. But I wish it was not so and I could live happily ever after with only Qobuz streaming, but I know that will take a superior (audio) network.
Again, dbastin , I really appreciated your replies. Thanks very much.
I apologize for any confusion. The objective is to completely isolate the audio from your home network and its interference. One approach is to use two separate modems, each dedicated to audio and home network usage. Alternatively, one can utilize a 4G or 5G ‘‘audio only’’ modem with a SIM card, which has been successfully implemented in Europe as far as I know, although it is not commonly practiced.
Another possibility is to employ software-based solutions, such as a VPN, to achieve full separation. Taiko addresses this issue through their Extreme router’s software.
As for Antipodes/Oladra it is uncertain whether they provide 100% isolation of network inputs from the home network. However, achieving complete isolation within the streamer itself may be challenging, and it is likely that the solution requires addressing the issue further upstream in the network setup.
I believe you mentioned that you also experimented with a non-shielded Cat5 or Cat6 Ethernet cable, but ultimately found that the SR Ethernet cable provided the highest SQ in this upstream section.
I can have only 1 modem, the national fibre termination. A 2nd service is available via 4G and 5G but I dont have good reception and it is an extra expense I cant justify.
I will review router settings to ensure the audio gear gets only relevant traffic, but as I understand that is already the case.
A VPN still comes through the same hardware so I’m not sure what the advantage is.
Honestly I am not a pro, not pretending to either. But you want isolation ‘‘like a VPN’’. I think in software it is possible to virtually separate in two networks that can not see each other but still share the same router (The Taiko solution as far as I understand it).
I do not know if a physical separation for example connecting a Wi-Fi access point to that same router for exclusive home network use results in the same quality isolation comparable to a software solution. If the goal is full separation, again, it is all in the details, maybe you need both?
I understand your point. I admire European audiophiles who enjoy the luxury of paying very little (less than 10 Euros per month) for a data SIM connection. Unfortunately, here in Canada, not only is it sometimes impossible to have a connection depending on your location, but we also have the highest communication rates globally.
On the bright side, if a 4g modem can be that good and I can cut down on buying three of your best Ethernet cables, I could potentially cover the cost of five years’ worth of data connection + the purchase of a 4g modem . I apologize if my previous statement seemed impolite; just want to highlight that perspectives can vary.
I would say isolation from noise sources is paramount and considerably more beneficial than isolating traffic, and without addressing the former it is probably impossible to hear benefits of the latter… maybe unless your network has tonnes of traffic. But Taiko are taking it to the extreme and addressing both, like the new Synergistic Research Router.
. I wasn’t aware, but I thought there was maybe more to it to obtain your great results. Glad I asked.