Streaming Sound Quality - Want the very best sound quality? Network Acoustics ENO...

Just added another small box.  I have a DJM Electronics GigaFOILv4-INLINE Ethernet Filter on the way.  I will probably position it immediately after my Bonn N8 switch and just before the Network Acoustics ENO, which is right before my server.  The GigaFOILv4-INLINE uses Fiber Optic Isolation Link (“FOIL”) technology to prevent EMI from passing through the filter.  I suspect my use of fiber in place of Ethernet cable is already effective at blocking EMI but I like the idea of having the GigaFOIL just in front of the ENO so I will give it a try and assess the results.  I may also try going back to Ethernet in place of the fiber to assess whether the fiber is redundant with the GigaFOIL in place.

Thank you, finally you are realizing it 👍

Ignorance is bliss, so they say.

@cakyol so, you are never wrong and everyone else is a "phool". nice.

@soma70 i know what i know. Most auidophiles (or phools sometimes) suffer from the "i paid more so i must be hearing better" syndrome, especially about wires , fuses and networking gear.

Listen to what @invalid has commented. That is almost always the issue.

99% of the so called audiophiles, and humans for that matter, once having reached about 40 years of age, will not hear anything above 16khz, let alone all the sound intricacies they claim they hear.

I invite everyone on this forum to be "real" with themselves and everyone else and comment accordingly.

AS LONG AS THEY ARE WITHIN 802.11 specs will work PERFECTLY.

That's the problem a lot of them do not meet specs even though they are advertised as such.

Good thread, thanks to OP and all.

It inspired me to try upgrading my ethernet cables between a Bonn N8 network switch and a Roon Nucleus.

I opted for Audioquest Cinnamon (about £100 each / solid copper core with silver) and have had very good results.

@cakyol is it possible that you don't know what you don't know? 

I’m a huge fan of @richtruss and Rob Osborne’s ENO Streaming System, which I incorporated into my listening chain this past fall. Both Rich and Rob were extremely helpful with answering all of my questions and concerns about their products and how I could best use them. Placing the ENO Streaming System between my D-Link network switch and my Roon Nucleus music server was an absolute revelation. Everything sounded much smoother and more analog through my headphones. What I’m hoping is for Network Acoustics to come out with their own network switch so they could offer a true synergistic setup from a listeners router to their music server. Rich and Rob make a great product that I wish more audiophiles would try in their own music streaming systems. 

There is no clocking in ethernet transmission. IT IS ASYNCHRONOUS. Do not spend your money on frivolous stuff like switches and cables. Most basic cables, AS LONG AS THEY ARE WITHIN 802.11 specs will work PERFECTLY.

Spend your money on the DAC instead. That is the most important component in a digital system.

@sns , The Ediscreation FIBER BOX II is an optical isolation solution not unlike using fiber and two FMCs, but somewhat less clunky as it is contained in a single box.  I suspect it would work well right in front of the ENO.  They also make the SILENT SWITCH OCXO, a switch using an OCXO clock with internal LPS.   Couldn't find a review but here are some nice pictures.

@audiom3 Prior to inserting Opticalrendu I had diasy chained SOTM SMS200Neo to Uptone usb regen, I had gone back and forth any number of times leaving Uptone out or replacing, always preferred the daisy chain. I then inserted Opticalrendu forgetting that I still had Uptone in chain. Listened in daisy chain configuration for couple sessions when suddenly realizing still had Uptone (by the way powered by lps) connecting to dac. Pulled it out and shazam, so far better, even lower noise floor, and most importantly newfound sense of ease and relaxation. My digital now may match or nearly match my vinyl setup in this area, along with much higher resolving powers (need more updates on vinyl side). The Uptone was obviously putting a slight veil and masking the ease of OR. I've now removed from my chain one switch, two ethernet cables and two lps, so less complex with far better sq.

@ddafoe My ethernet solution cable was all Audioquest Vodka, my optical solutions, even with generic fmc and one ethernet conversion bested ethernet solution

I was also recently made aware of a another formerly unknown to me optical solution with the following, https://ediscreation.com/product/fiber-box/ This may be solution for me upstream of server.

For the ethernet fans here, I'm not stating with absolute certainty optical superior to ethernet, but so far my optical setups have been superior, others seem to concur with this view. I certainly hear far more deeply into recordings, and I've not heard any sound signature or flavor added with optical. Thus far my optical explorations have gone further than ethernet, the Eno may still be in my future if I wish to compare ethernet under more favorable conditions.

I seriously question the additional conversion for runs under 6-8m. A well shielded cat8 ethernet cable is the better solution in that case, even before discussing the quality of PSUs needed for the convertera. The Etherregen in combination with a well clocked switch is a cleaner and better solution.

This didn't hold true in my system using my ears at least...

Replacing ~10 ft of ethernet from a noisy equipment closet to my audio rack with fiber made a BIG difference.   I heard a darker background and more clarity/detail, with no added harshness.   I terminated the fiber with each of the following: EtherREGEN, OpticalRendu, and $100 StarTech converter (+ after market power supply) and the results were all about the same minus differences between USB and SPDIF (my preference) which I assume to be related to my DAC.  All of these were much better than straight Ethernet to the M5 from my perspective.

Yes, my 10 ft run was not cat8 or of high quality, but the difference in sound quality vs copper was so significant I don't think buying a better 10 ft copper cable would have made up all of the differences.

Baseline:

• switch with sfp port -> 10 feet cheap ethernet cable -> EtherREGEN -> short cat8 -> Bricasti M5 -> SPDIF -> DAC

Fiber 1-SPDIF:

• switch -> 10 feet of fiber -> EtherREGEN -> short cat8 -> Bricasti M5 -> SPDIF -> DAC

Fiber 1-USB:

• switch -> 10 feet of fiber -> EtherREGEN -> short cat8 -> Bricasti M5 -> USB -> DAC

Fiber 2-USB:

• switch -> 10 feet of fiber -> OpticalRendu -> USB -> DAC

Fiber 3-SPDIF:

• switch -> 10 feet of fiber -> $100 Fiber Converter -> short cat8 -> Bricasti M5 -> SPDIF -> DAC

I'm currently running the latter as I didn't detect a drop after replacing the more expensive converters.   I've been tempted to buy another EtherREGEN to drop back in to see if I would now hear a difference vs the much cheaper converter...

I seriously question the additional conversion for runs under 6-8m. A well shielded cat8 ethernet cable is the better solution in that case, even before discussing the quality of PSUs needed for the convertera. The Etherregen in combination with a well clocked switch is a cleaner and better solution.

@sns said: As debjit_g mentioned above, conversion of optical back to ethernet is somewhat detrimental, place for noise to re-enter network. And the boards in generic FMC are not optimum. Understanding this is what motivated me to upgrade to Opticalrendu.

This is what prompted me to use a Sonore oM from my NUC and an Uptone EtherRegen right before my DAC.  The only two copper/fiber conversions in the entire path.  The differences in adding both (I added each in at separate times) were subtle but definitely noticeable.

anybody (possibly @grannyring ?) been a beta tester of the Muon USB cable and shed some comparison with the Eno ?

As debjit_g mentioned above, conversion of optical back to ethernet is somewhat detrimental, place for noise to re-enter network. And the boards in generic FMC are not optimum. Understanding this is what motivated me to upgrade to Opticalrendu.

Still, I did find two generic FMC preferable to my less than optimal ethernet chain. It would be good to continue to hear from folks who've compared optimized optical vs optimized ethernet chain. Prior to purchasing opticalrendu I planned on adding Eno after second FMC, this based on speaking with Rich at NA in regard to best placement in my particular setup. I still may purchase as I could remove opticalrendu, go to Eno connecting to SOTM SMS200Neo or use upstream of server with optical remaining between server and streamer.

This is correct! 

“I am looking at the NA website but there is no mention of Eno Ag anymore”

@debjit_g 

They may have reduced the line to one version of ENO since they are getting to release upscaled version of ENO called Muon. 

My experience too.  I had switched back to 45 feet of Ethernet over the weekend and just switched back to fiber.   Even with the two additional converter boxes (both with LPS) the slight effect was a very slight smoothing or taking an edge off but more to the positive than the negative and without losing detail or sparkle.  Both sound good but I like the fiber just a touch better.

Folks, I think my use case might be completely different than yours. I am NOT talking about fiber through FMC vs direct copper. I am talking about the differences between the Fiber NIC vs JCAT Net NIC. This itself carries a lot of weight and sets the right context.

I am looking at the NA website but there is no mention of Eno Ag anymore as has been mentioned several times in this thread. Did they remove the Ag and copper version and now has only one version to sell ? Is it Ag or copper ?

In my experimentation/complete overhauling of my copper to fiber LAN, there was never a marring out of details.  

My experience too.  I had switched back to 45 feet of Ethernet over the weekend and just switched back to fiber.   Even with the two additional converter boxes (both with LPS) the slight effect was a very slight smoothing or taking an edge off but more to the positive than the negative and without losing detail or sparkle.  Both sound good but I like the fiber just a touch better.

Interesting findings.  In my experimentation/complete overhauling of my copper to fiber LAN, there was never a marring out of details.  It was quite the opposite actually.  Much better extension, space and micro details when I removed all of the added distortion/noise from the infrastructure.  Fiber optics > copper+filters every day of the week.

It's vastly system dependent and also depends on who has the worst noise copper or fiber in a particular deployment. Just having fiber doesn't not mean better sound - that's a misconception. A lot of folks uses FMC to convert copper to fiber and then another FMC to convert is back to copper. Remember the Xilinx fpga used in most FMC creates a ton of noise and jitter by themselves and to top it over you have added power supply noise. All adds up but the bottom line is if you find fiber sounds better, that's what matters :-) 

Having said that, I still use fiber between the Buffalo and the EdgeRouter.

@debjit_g wrote: "of all these yrs that I have been experimenting, introducing fiber not necessarily translates to better sound - it’s just sounds different. In most cases (in some of my friends setup and mine) I have seen a softening effect to a point where it mars out the details slightly."

Interesting findings.  In my experimentation/complete overhauling of my copper to fiber LAN, there was never a marring out of details.  It was quite the opposite actually.  Much better extension, space and micro details when I removed all of the added distortion/noise from the infrastructure.  Fiber optics > copper+filters every day of the week.

Also thining to try Euphony Stylus endpoint vs Roon endpoint.

I used to run Euphony in my build sometime back but moved on since then. My optimized Windows sound better in my system. Also, the current version of Euphony runs quiet a older Linux kernel which doesn’t have the needed support for the latest Intel/AMD cpu/chipset. So be careful with the h/w and s/w combination. Their is a talk of 4.0 version but don’t know when it will be released.

Now back to NA :-)

Yes, future experiments will absolutely compare optimized optical and usb comparisons, also, one computer vs two computer setups. I'd also like to try return to Audirvana, compare with Roon and Roon/ HQPlayer. When I first changed over from previous iteration of Audirvana (prior to Studio) to Roon was quite sure I preferred Audirvana. Also thining to try Euphony Stylus endpoint vs Roon endpoint.

@debjit_g , this is the board used in Mojo,

http://www.jlsounds.com/i2soverusb.html

I suspect the modified Sonore Opticalmodule within Small Green Computer SonicTransporter Optical I9 would provide better isolation from motherboard.

it’s the same isolation that you would get if you had the opticalmodule outside and connected via copper to the motherboard. To run fiber end-to-end, one needs to design the motherboard from scratch and don’t think anyone is upto the task (well, Sotm did but they don’t have fiber)

of all these yrs that I have been experimenting, introducing fiber not necessarily translates to better sound - it’s just sounds different. In most cases (in some of my friends setup and mine) I have seen a softening effect to a point where it mars out the details slightly.

you can’t go wrong with either the PF or the JCAT. Both are equally good. It’s just a matter to taste which one you like.

@debjit_g  Forgot to mention, I have second ethernet port on my server, so I can do FMC without having to go through an added switch in network.

@debjit_g That seems like logical explanation, I suspect the modified Sonore Opticalmodule within Small Green Computer SonicTransporter Optical I9 would provide better isolation from motherboard. Unfortunately, this doesn't provide for optimal usb out. I want the choice of both optimized fiber and usb with my next server, diy may be only solution for me.  I'd have difficult choice between Pink Faun and JCAT.

Also, I presume having to go back through Buffalo switch with fiber vs. direct to dac with usb would have bearing on your preference.

@hchilcoat the JL sound board is fabulous. Uses xmos chip. Recently Lampizator started using their board across all their models, low and top tier, as well.

@debjit_g 

Based on what I see showing up in Roon settings, looks like a JL Sounds USB board. 

Anybody know what USB controller is used in Mojo DAC ? It’s it Xmos based or Amanero or something else ?

I think I haven’t found a DAC that doesn’t benefit from a good USB cable.

Well the EE 8Switch was a definite improvement over a standard $50 switch in my system.  When coupled with a Network Acoustics cable and filter the results are just great in my system.  When the new NA switch comes out I will definitely try it.  
 

I owned the Innuos Phoenix and Ideon $4000 usb filter/reclockers and ended up selling them.  I found my Mojo Evo dac did not benefit from them. Just a great usb cable is what the Evo seems to like the most. My Innuos Zenith really partners well with the ENO filter.  I really think much of this is gear dependent. 

@sns  I think the issue is not with fiber connection as such but the noise a fiber nic creates back into the motherboard. Unless you have filters downstream PCIe, there is a high chance that noise is going through seep through. Unlike a generic fiber nic, the JCAT is purposefully designed to deal with the least noise as possible and its also have external 5v power input (which I use as well). I think this is a major contributing factor, at least in my case. 

Yes, I am taking usb from the JCAT XE USB straight to the DAC. When I used the fiber nic it was connected to a Buffalo switch.

The end of production for DejaVu was bummer for me, JCAT usb major calling card for me.

Debjit_J, interesting you found JCAT superior to Fiber NIC, wonder if fiber prior to server had bearing on this? It may also be what comes after usb or fiber out of your server has something to do with it? Assume your taking usb out of server straight to dac. What does fiber out go to?

@mitch2 thanks for the detail. Your setup is very very similar to mine. My router connects to EdgeRouteX SFP switch and then I run a 30ft long fiber into a Buffalo BS-2016 switch (same used in Melco S100) near the music server. The music server is DIY and uses quality choke based linear power and its also equipped with JCAT XE USB and Femto Net card. I am thinking of upgrading to JCAT XE Net and mod the Buffalo with a PF OCXO clock. I tried playing with Fiber NIC in the server but went back to the JCAT. Hence I am trying to see what ways I can further optimize my setup but I don't stream and all my playback in from local files (CD rips and high bitrate downloads). Therefore, I am trying to access the return of investment. The Eno seems like a good addition for folks streaming, I just will have to experiment if it makes further improvements and without trying one is just not possible to say theoretically. Good that NA has a 30 day return policy.

@debjit_g 

This recent exploration of my streaming set-up all started because I purchased the ENO Ag and a Network Acoustics network cable.  A member here alerted me to this unit being for sale and I was fortunate to jump on it quickly.

I have a Mojo Audio server/streamer (DejaVu) and their DAC (EVO B4B).  I have a 45 foot long network cable from my router into the server and then USB into the DAC.  I added the ENO ahead of the server, by first bringing the network cable into a Bonn N8 switch, then the Network Acoustics cable from the switch to the ENO, then the short network cable attached to the ENO into the server.  The server runs Roon and serves as both the core and player.  I stream both Tidal and Qobuz but the addition of Qobuz is recent.  

Even before trying the different set-ups, I thought the system sounded very nice, organic, full-bodied tone, and powerful, with good extension and staging.  My initial opinion was that the ENO added something that made the music sound just a little more real somehow, but I haven't yet been able to point to one particular thing that causes that.  My plan is to stick with it for some time and then take it out and decide whether I am missing anything.  Leaving new stuff in for a while and then taking it out to assess differences has been a good method over the years for me to evaluate what I like the sound of.  Right now I would say the ENO filter was a positive addition but I really don't know why yet until I listen more.

The Mojo Audio stuff is built great, with outstanding power supplies, and both pieces use high quality I/O boards by JCAT such as this XE USB board.  I suspect the already high quality pathway of my server/streamer and DAC have a lower need for corrections brought on by small add-on boxes.  Even the ENO and their cables are certainly less than a revelation, but may be one of the few add-ons to make a positive difference.

In working through my network connections, I have recently done some reading on switches.  The more I read, the more skeptical I become about what is being marketed and sold as an "audiophile" switch, and the high prices of these switches.

The thread on Audiophile Style titled "Uptone Ether Regen - English Electric 8 Switch Comparison" is interesting.  I am left wondering whether any of these have value beyond that of a basic Netgear or TP-Link switch.

@mitch2 are you streaming (perhaps Tidal or Qobuz or from NAS) through the DejaVu server using Roon ? How much do you think the Eno added to the SQ ?

Been listening and comparing the sound through Roon with:

• DejaVu server connected directly to the EVO B4B DAC using USB (DejaVu is core and player), and
• DejaVu connected by Ethernet to the Metrum Ambre and then by AES/EBU to the EVO B4B DAC (DejaVu is core and Ambre is player)

In both cases, the ENO was connected after a Bonn N8 switch and in front of the DejaVu server.

My initial impressions remain unchanged in that both connections sound very good with the direct DejaVu>EVO connection resulting in a slightly fuller, more rounded presentation and the DejaVu>Ambre>EVO connection being slightly more incisive and dynamic.  Some may describe it as musicality vs. clarity.  The differences are slight but seem repeatable.  However, there are other variables not least of which being the USB vs. AES/EBU connection at the DAC.  If I had to select one option now, I would probably choose the direct DejaVu>EVO connection, without using the Ambre, as it seems the direct connection is slightly more musical and enjoyable on a wide range of material - but I could live with either.  

@lalitk I see, you've learned all there is needed to be learned. And yes, I will carry on.

@sns

Once again, you missed my point. As you pointed out in one of your earlier post, ‘I still think much to be learned’. Carry on!

@lalitk There are two different audiences here, yes, basics important to those just venturing into streaming, for the more mature streamers this is explorative information. I see no harm in trying and reporting on every single streaming device available. This only adds to knowledge base, most mature streamers can read between the lines and determine for themselves the veracity of those reporting on these devices and the devices themselves.

Trying to limit discussion to only devices and setups that meet some subjective base line is only censorship. I for one like to hear about many setups, so keep the experiments and reviews coming.

@antigrunge2

I’m well aware of your somewhat idiosyncratic system approach. My point is simple, it is very easy to get carried away in our pursuit to improve performance to nth degree. There is this breaking point where adding more ‘tweaks’ will only result in marginal or subtle improvements if your baseline is weak. It’s like putting a ‘lipstick on a pig’ :-)

Streaming is no longer in its infancy, it has matured. I see some of the posters here regularly peddling their convoluted biases like setting a good streaming system is equivalent of setting up a science lab. The way I see, it’s not very helpful to readers that are trying to venture into streaming. 

I am very interested in trying out the Eno filter in my system. However, I don’t stream and all playback is from locally sourced files residing in the music server itself. The server is connected to the home network only to control Roon. Curious if anyone found improvement with the Eno filter in such scenario ? 

@antigrunge2 Exactly, my experience as well. Still, my search is on for the most efficient means to extract maximum sound quality. My present thoughts along this line is server with optical out capability into streamer or streaming dac with optical in capability.

I'd also add complexity vs maximizing simplicity is a trade off. More complexity means more cables, connectors, devices, power supplies, all can generate, be conduits or allow leakage in of EMI/RFI. Greater separation of devices at the same time may theoretically decrease this same noise. I have Trifield Meter and the greatest generation of RFI comes from server/computer motherboards, transformers of all kinds and storage devices such as ssd and hard drives. Put these all under one roof, not good.

Paradoxically, the best solution should be something in between the two extremes. While I'd like to see power supplies,  and perhaps storage in separate enclosures, I wouldn't mind seeing optimized usb within servers which may negate need for all kinds of usb thingies, or separate streamers for that fact. They could also offer clean ethernet inputs within servers, devices like Uptone Etherregen wouldn't have to be separate and external. Storage could also be well shielded within server, eliminating the need for NAS or USB external storage. Having the ethernet and usb optimized within a single server also serves to eliminate possible clocking incompatibility, per what you mentioned above.

I do think the Taiko Extreme and Wadax servers address all my requests, minimal devices and maximum sound quality all under one roof. Unfortunately, this comes at prohibitive price for most of us!

Unlike analogue, where any additional device will affect the sound through changing capacitance and impedance effects on the transmitted signal, the signal path of digital requires only avoidance of noise incursion and maintenance of clock accuracy. It seems therefore to follow that the number of devices and cable links -as long as they satisfy the above conditions- is immaterial. While I agree with the simplicity mantra on analogue, my experience on digital points in a different direction.

I'd say add to our streaming knowledge and try, I still think much to be learned.

I hope to have most or all of my questions about streaming answered by end of year. I know someone who is in process of diy build, aka self assembly of nearly state of art server/streamer, based on final spec this will be equivalent to off the shelf server in price range of $10k-$20k. It is based on full size atx motherboard, will run windows or linux os, have enough processing engine to run HQPlayer with highest level DSP, folded into Roon, which means it can be run sans HQPlayer or with no DSP engaged. Industrial spec RAM and SSD. Ports will include optical, at least two ethernet, USB off motherboard, hopefully, USB off PInk Faun or JCAT, and high end lps.

With this server will be able to compare HQPlayer processing at any level vs no processing. With the optical port it can be used with Sonore Opticalrendu or some other optical capable streamer or streamer/dac. With direct ethernet out it can be used with ethernet streamer/usb renderer/Roon endpoint or streamer/dac. With two choices of USB, motherboard vs onboard usb rendering directly into dac can be compared. By the way, I consider the direct off motherboard usb the greatest flaw with the vast majority of off the shelf servers, this will directly compare that to optimized USB, these boards are $1200-$2500 each and are reportedly worth all the money, here's link to discussion, https://www.whatsbestforum.com/threads/jcat-usb-xe-or-pink-faun-ultra-ocxo.32332/

I've been doing self builds of general service computers for years, music server are simply general service computers optimized as music players only. The vast majority use off the shelf motherboards,processors, ssd, RAM, etc., some have custom built lps and some upgraded ports. The only bespoke servers that exist are something like Taiko Extreme and Wadax Reference, may be others I'm unaware of. These are truly one of a kind, fully engineered from scratch servers and likely worth every penny for those that can afford. Mike Levine here, I believe has posted on both. I only wish this build could be on that level. The vast majority of sub $10k servers I've looked at don't have the necessary processing engine  to run HQPlayer at highest processing levels, don't use industrial grade RAM or SSD, have less than optimal port availability and/or those ports not fully optimized, and not all that impressive lps. Going the self build route allows one to build a much better server at a reasonable price, not saying the off the shelf servers not worth the money, only comparing business models here. Self build uses free R&D, free labor, no marketing costs, no profit margin. For many opportunity cost estimates or lack of computer building experience may make self build prohibitive. This is not meant to dis on off the shelf, these servers provide value to those purchasing them.

Anyway, we will listen and compare various configurations of this server, hopefully it will be finished by early spring. I will then build my own server based on my particular needs. Listening tests will hopefully help me determine the exact components I will use in my build. This endeavor will test my theories that separate server, streamer and optical isolation solutions are optimal. It will also have allowance for multi optimized ports which will make this server maximally versatile, many different streamers and or streamer/dacs will be able to optimally connect to server.

While the above will test server to dac leg of network, what comes in front of server remains to be tested in like manner.