Innuos Zenith Mk3 or OpticalRendu + separate server?

[matthias]

31 minutes ago, barrows said:

 I would suggest (but realize it will fall on many deaf ears) that one would be much better off implementing optical isolation and using a relatively generic upstream computer as the file server (or just a NAS if one has no need for DSP/oversampling in software) and then applying the thousands of dollars saved on better loudspeakers, where big improvements really can be realized. 

 

Ivor Tiefenbrun of Linn has proven that this is a bad idea. The quality of the source determines the enjoyment and quality you get from music and not the quality of the speakers. Source first rules.

 

Matt

[BigAlMc]

33 minutes ago, barrows said:

I challenge you to propose a  technical theory by which an upstream device will matter in an optical Ethernet based set up.  

 

You're embarrassing yourself now.

 

You've preached the theoretical benefits of the MicroRendu for two years now but many peoples EXPERIENCE has led them to move onto better solutions.

 

I don't need to explain or propose the scientific theory. I'm telling you that my Innuos MKII sounded noticeably better than my MicroRendo.

 

This optical is the holy grail argument is frankly beneath you. It might well be a great advance. I truly hope it is. Sincerely. But I just can't see it being the endgame you're purporting.

 

Cheers,

Alan

 

 

[BigAlMc]

@barrows,

 

Upon reflection I apologise for that previous post. I was annoyed at your argument but I was overly personal. My bad.

 

I know your work and generally respect you. But guess we need to agree to disagree here.

 

My apologies for the embarrassing yourself comment.

 

Cheers,

Alan

[barrows]

30 minutes ago, matthias said:

You are walking on very thin ice. Software upstream matters, but hardware not?

Sorry, it sounds like I was not entirely clear, I'll put it another way and perhaps be able to get my point across:

 

The software upstream only matters because of how the Renderer has to operate to accommodate the network protocol in use.  What the Renderer is doing is what makes the sound quality differences.  That is: if the server (upstream) is using ROON, then the Renderer also has to use its RAAT protocols, and using these protocols in the Renderer results in different sound quality than if the Renderer is using DLNA (for say, miniMserver, or Squeezelite, etc.).  Again, I stress, it is what is going on in the Renderer which determines the sound quality.

 

Another detail: The upstream server hardware does not matter, because it makes no difference anymore to the Renderer: In a direct connected set up (Server-USB-DAC) it is the noise signature of the hardware (server box and its power supplies) which comes over to the DAC via USB which makes for the sound quality differences.  With the optical Ethernet connection from server to Renderer/DAC, that noise signature is no longer present.

[matthias]

3 minutes ago, barrows said:

The software upstream only matters because of how the Renderer has to operate to accommodate the network protocol in use.  What the Renderer is doing is what makes the sound quality differences.  That is: if the server (upstream) is using ROON, then the Renderer also has to use its RAAT protocols, and using these protocols in the Renderer results in different sound quality than if the Renderer is using DLNA (for say, miniMserver, or Squeezelite, etc.).  Again, I stress, it is what is going on in the Renderer which determines the sound quality.

 

Before Roon you used Audirvana. 

What happens when you change the izotope or sox settings in Audirvana? Nothing to do with your Sonore renderer, it happens long before the signal reaches the opticalRendu. Not audible? 

 

Matt

[matthias]

9 minutes ago, barrows said:

Another detail: The upstream server hardware does not matter, because it makes no difference anymore to the Renderer: In a direct connected set up (Server-USB-DAC) it is the noise signature of the hardware (server box and its power supplies) which comes over to the DAC via USB which makes for the sound quality differences.  With the optical Ethernet connection from server to DAC, that noise signature is no longer present.

 

So per your definition the source is the device which connects directly to the DAC, either the server or the renderer and this device determines the SQ.

So we can try what sounds better, a DAC connected to a Sonore or a Innuos for example. When it sounds better, it is better.

 

Matt

[barrows]

6 minutes ago, matthias said:

 

Before Roon you used Audirvana. 

What happens when you change the izotope or sox settings in Audirvana? Nothing to do with your Sonore renderer, it happens long before the signal reaches the opticalRendu. Not audible? 

 

Matt

As I mentioned before, of course if you use DSP in a server (whether that be oversampling, room correction, or whatever) of course the sound will change.  In fact, if you like to use lots of DSP, then an optical Ethernet connected approach is ideal, as one needs a powerful (and hence noisy) server in order to use lots of DSP functionality.

I have used Audirbana +, miniMserver, Squeezlite, and ROON here, so plenty of experience with the different Network protocols, except for HQPlayer and NAA (soon).

[FredM]

39 minutes ago, barrows said:

Another detail: The upstream server hardware does not matter, because it makes no difference anymore to the Renderer: In a direct connected set up (Server-USB-DAC) it is the noise signature of the hardware (server box and its power supplies) which comes over to the DAC via USB which makes for the sound quality differences.  With the optical Ethernet connection from server to Renderer/DAC, that noise signature is no longer present.

 

With no relevant knowledge on this subject, I did remember a viewpoint in the SOtM switch review: “There is no single product which removes noise completely, even though the product are good at reducing noise so it improves sound quality, it doesn’t mean that the products remove noise completely”. I guess only listening in practice will give the answer (in both cases).

[barrows]

10 minutes ago, FredM said:

With no relevant knowledge on this subject, I did remember a viewpoint in the SOtM switch review: “There is no single product which removes noise completely, even though the product are good at reducing noise so it improves sound quality, it doesn’t mean that the products remove noise completely”. I guess only listening in practice will give the answer (in both cases).

The optical connection stops all upstream noise from getting to the downstream components, period.  From that downstream point (Renderer) all the noise and signal integrity (USB) issues that exist matter for sound quality, so use the best Renderer you can.

 

I can site sources from the Internet which will support any view one can imagine: in fact, on this very site I can find posts which say that optical Ethernet sounds bad, that Wifi Ethernet is best, that wired Ethernet is best, that optical Ethernet is best, heck, I can find sources who say that a direct connected laptop running iTunes is as good as it can get!

 

Do not believe everything you read.

 

I also often see the opinion that "we do not understand everything there is to know about digital audio"; while of course that is true, this notion does not mean that we should disbelieve the technical matters which we actually do completely understand.

 

 

[Dutch]

35 minutes ago, barrows said:

The optical connection stops all upstream noise from getting to the downstream components, period.

 

The designer of the equipment you advocate seems to disagree.

 

 

[barrows]

8 minutes ago, Dutch said:

The designer of the equipment you advocate seems to disagree.

I am referring to electrical noise, not esoteric phase noise from the upstream clock(s).

 

At this point I would like to note we are getting into a very esoteric area.  There is presently no accepted science which explains if, and how, the phase noise of an upstream clock can affect the performance of a downstream Renderer.  First consider that the Renderer operates asynchronously, that is it ignores the upstream clock, and re-clocks all data to its own internal clock.  So the performance of the internal clock in the Renderer is paramount.  Once we accept this fact, we must first make sure our choice in renderers use a good internal clock(s).

OK, to discuss upstream clock phase noise, again, there is no accepted science of if/how upstream phase noise might effect the performance of a downstream Renderer...  Let that sink in before you go on believing this is a real problem.  After that has sunk in, let's consider: there is some anecdotal evidence to suggest the somehow upstream clocks are effecting downstream, even asynchronous, Renderers.  I believe there is enough anecdotal evidence to suggest further investigation into this.  John is currently making those investigations, and he is hoping to be able to measure these effects (if there are any).  I look forward to seeing what results he comes up with.  Let us also note that there are some really sketchily "engineered" products out there (like modified switches) that claim to reduce clock phase noise on Ethernet connections, and that some of the anecdotal evidence reported is based on these devices.

In the mean time I do look forward to testing for myself if a better upstream clock makes any audible difference (via an audiophile FMC with good engineering), and I will do that as soon as possible.  If it does, it is an easy "problem" to address, and does not change the fact that no electrical noise is carried on optical Ethernet connections. 

[BigAlMc]

20 minutes ago, barrows said:

I am referring to electrical noise, not esoteric phase noise from the upstream clock(s).

 

At this point I would like to note we are getting into a very esoteric area.  There is presently no accepted science which explains if, and how, the phase noise of an upstream clock can affect the performance of a downstream Renderer.  First consider that the Renderer operates asynchronously, that is it ignores the upstream clock, and re-clocks all data to its own internal clock.  So the performance of the internal clock in the Renderer is paramount.  Once we accept this fact, we must first make sure our choice in renderers use a good internal clock(s).

OK, to discuss upstream clock phase noise, again, there is no accepted science of if/how upstream phase noise might effect the performance of a downstream Renderer...  Let that sink in before you go on believing this is a real problem.  After that has sunk in, let's consider: there is some anecdotal evidence to suggest the somehow upstream clocks are effecting downstream, even asynchronous, Renderers.  I believe there is enough anecdotal evidence to suggest further investigation into this.  John is currently making those investigations, and he is hoping to be able to measure these effects (if there are any).  I look forward to seeing what results he comes up with.  Let us also note that there are some really sketchily "engineered" products out there (like modified switches) that claim to reduce clock phase noise on Ethernet connections, and that some of the anecdotal evidence reported is based on these devices.

In the mean time I do look forward to testing for myself if a better upstream clock makes any audible difference (via an audiophile FMC with good engineering), and I will do that as soon as possible.  If it does, it is an easy "problem" to address, and does not change the fact that no electrical noise is carried on optical Ethernet connections. 

 

Dude you are digging a hole for yourself.

 

You want to lecture science over our experience but then state things like:

 

22 minutes ago, barrows said:

There is presently no accepted science which explains if, and how, the phase noise of an upstream clock can affect the performance of a downstream Renderer. 

 

 

Let's just accept there are differing views here.

 

Beyond that if you want to maintain this "our isolation makes the server irrelevant" argument its gonna be put up or shut up time.

 

I have a Zenith SE as well as an Audiolinux NUC server on hand and am happy to host a few CA/AS people for a bakeoff  if you want to send a Optical Rendu prototype or whatever here. Likewise I'm sure there will be US based members willing to do the same.

 

But this stance that your science trumps everyone else's experience and that experience is "anecdotal" isn't winning this argument for you. Its also pissing people off to be blunt. So either accept there are many unknowns here or prove your claims.

 

Cheers,

Alan

 

[barrows]

7 minutes ago, BigAlMc said:

Dude you are digging a hole for yourself.

Umm, no.  

 

If the upstream clock phase noise does turn out the an actual issue in terms of an optical Ethernet based architecture, there is an easy and (relatively) inexpensive way to address that without resorting to the (what, $7K to $XXK) expense of an audiophile sever: just add an UpTone Etherregen, problem solved, with a well engineered solution at a reasonable price.  There will soon be other well engineered, reasonably priced solutions for such as well.

 

But I will be very surprised if the upstream clock performance even really matters, at this point I am remaining open minded on this possibility until i test it myself here.  But there is considerable evidence that this clock phase noise is a non-issue: such as the fact that the Internet even works at all, just consider all those thousands of clocks involved in getting a Tidal stream to your house, all adding their own heinous phase noise to your system.  I would suspect that Tidal would be unlistenable if clock phase noise (excepting the last in the chain) really was cumulative.

 

But by no means do I expect audiophiles to just blindly believe what I say, go ahead and test for yourself, like I suggested, just do not blindly follow some Internet hobbiest based on their experiences, test for yourself, in your own system.

[BigAlMc]

7 minutes ago, barrows said:

But by no means do I expect audiophiles to just blindly believe what I say, go ahead and test for yourself, like I suggested, just do not blindly follow some Internet hobbiest based on their experiences, test for yourself, in your own system.

 

I'm at a loss for words in general but to be clear my own journey included:

 

ÑAS

MicroRendu

FMC

Singxer SU-1

Zenith MKII

Zenith SE

NUC Audiolinux

 

So I would argue with strong conviction that I have tested.

 

Regardless, bring it on. Send me an Optical Rendu and I'm happy to test and invite a few others to observe whether it renders the SE, NUC and my Lenovo laptop equal as you claim.

 

I'll pay full price as a holding deposit and keep it if it wins the bakeoff. But I'd expect to return it at no cost to myself if it doesn't live up to expectations.

 

Put up or shut up my friend, because beyond that this is getting tired and we're doing the Op a disservice.

 

Cheers,

Alan

[barrows]

19 minutes ago, BigAlMc said:

Put up or shut up my friend, because beyond that this is getting tired and we're doing the Op a disservice.

 

I find your attitude distasteful, and am unsure as to why I would be obligated to prove anything to you?

 

In any case, you are free to try whatever you would like, and you are free to post anything you would like, just as I am.  I have nothing to prove to you, or to anyone else for that matter, I am only here to offer the benefit of my experience and to present some of the actual technical facts surrounding different approaches to playback.  It is up to readers here to come to their own conclusions as to which approaches they might want to try in their own systems.

I feel I have made my position clear here, and presented the facts as well as possible, so now I will leave this thread so that others have some space to weigh in with their experiences.

[Blackmorec]

The problems with discussions like these and with digital audio in general is that you get the theory and you get the practice and in my recent experience they don’t necessarily match up perfectly. 

The problem with digital audio is that unlike the past, where hi-fi could be easily isolated from other household appliances, streaming has to co-exist with a number of other clients, so designing a truly optimal network is a lot more complex given that the same network needs to serve UHD Video streaming, gaming maybe and a host of fixed and roaming clients. 

When I designed my original system I initially used a lot of the theory in my decision making....for example clock accuracy (and therefore jitter) does’t matter for asynchronous systems;  high frequency noise is not an issue in optical or fully isolated systems etc. 

What I found in practice is that these things may not matter until you measure them in terms of sound quality, then they do.  An example is asynchronous USB. In theory an asynchronous DAC discards the embedded clock signal and uses its own internal clock so the quality of the clock signal arriving at the DAC is in theory irrelevant. However, I and a significant number of audiophiles have found that reclocking the signal before the DAC has a very positive impact on SQ.  A product like the Innuos Statement has 2 USB outputs...one on the mother board and the second with its own OCXO 3 ppb clock and dedicated LPS. My DAC is asynchronous so the 2 USBs  should in theory sound the same yet the OCXO USB is utterly superior in terms of SQ. 

 

Theoretically an optical system will sound better than wired ethernet because of the full isolation of electrical noise. In practice, the optical system needs to convert, transmit, receive and reconvert the data stream and while noise may be reduced, jitter probably isn’t, so SQ gains made in reducing noise could well be lost through increased jitter.  What this swings and roundabout situation means is that there really isn’t a one-size-fits-all formula and theory doesn’t necessarily align with practice when SQ is the arbiter. 

As a result, audiophile consumers should not rely only on theoretical benefits without confirming that they indeed translate into actual sound quality benefits. Many of the improvements I made to my network flew directly in the face of what theoretical logic would dictate....for example vibration control of the router,  improved DC quality to modem and router etc. 

In light of the above its also important to treat manufacturer’s claims with a reasonable degree of scepticism. I implemented an audiophile switch where the manufacturer clearly stated that their SMPS was fully optimised and could not be improved upon.  This turned out to be be incorrect and while their SMPS sounded perfectly adequate, the aftermarket LPS I implemented brought jaw dropping improvements in practically every sonic area valued by audiophiles....improvements so profound it led me to implement LPSs elsewhere in my network.

 

[BigAlMc]

13 hours ago, barrows said:

I find your attitude distasteful, and am unsure as to why I would be obligated to prove anything to you?

 

Right back atcha. I find your attitude distasteful as your insinuating that I'm foolish for spending money on a good server because all I needed was to wait for Sonores completely unproven optical decrapifier. But in fairness to the Op I guess we park it there and agree to disagree.

[vortecjr]

There are people with all in one servers solutions and people with a server and endpoint solution. Clearly Sonore's approach is the latter. In our solution we are trying to keep things as simple as possible so you can use a regular computer or NAS as a server. In the all in one the emphasis is in improving off the shelf hardware. In our solution the emphasis is on a well designed endpoint that doesn't need improvement. There are going to be people on both sides of this argument...to each his own!  

[Blackmorec]

4 hours ago, vortecjr said:

There are people with all in one servers solutions and people with a server and endpoint solution. Clearly Sonore's approach is the latter. In our solution we are trying to keep things as simple as possible so you can use a regular computer or NAS as a server. In the all in one the emphasis is in improving off the shelf hardware. In our solution the emphasis is on a well designed endpoint that doesn't need improvement. There are going to be people on both sides of this argument...to each his own!  

I have to say that I disagree with your summary. The ‘one box’ servers I’ve had experience with, namely from the Innuos range have been extremely well designed and optimised,  offer SOTA sound, and are extremely easy to use, requiring neither regular computer nor NAS.  Practically any device these days uses off-the-shelf parts, as no audio companies manufacture SSDs, CD drives, memory chips, oscillators etc. 

[vortecjr]

25 minutes ago, Blackmorec said:

I have to say that I disagree with your summary. The ‘one box’ servers I’ve had experience with, namely from the Innuos range have been extremely well designed and optimised,  offer SOTA sound, and are extremely easy to use, requiring neither regular computer nor NAS.  Practically any device these days uses off-the-shelf parts, as no audio companies manufacture SSDs, CD drives, memory chips, oscillators etc. 

I'm sure your happy with it..no argument there. To me they are no different than the line products from Small Green Computer which offer the same functionality at a much lower price point. I have no issue with people using them and just prefer a different approach. The main board on the Rendu series is custom built for the task with no extra unused components not related to audio. People already have a computer or a NAS in their homes and the point of the Rendu series is to utilize them as servers. Personally, I'm either streaming Tidal or Qobuz without the need for a computer or using my Synology NAS as a server. Either way the electrically noisy and acoustically noisy drives are elsewhere away from my audio room.       

[BigAlMc]

1 minute ago, vortecjr said:

. To me they are no different than the line products from Small Green Computer which offer the same functionality at a much lower price point.

 

Sorry Jesus but this is either a naive comment or an uninformed one.

 

I went from the Audiostore Prestige 2 (Martin Smiths version of Andrew Gillis  Small Green Computers server) to an Innuos Zenith MKII and the difference was night and day. The Zenith is far superior. The Zenith SE is even better still.

 

Cheers,

Alan

[vortecjr]

Just now, BigAlMc said:

 

Sorry Jesus but this is either a naive comment or an uninformed one.

 

I went from the Audiostore Prestige 2 (Martin Smiths version of Andrew Gillis  Small Green Computers server) to an Innuos Zenith MKII and the difference was night and day. The Zenith is far superior. The Zenith SE is even better still.

 

Cheers,

Alan

I never questioned your perception. As discussed I don't use SGC servers for local playback only as streamers. I also don't make those products or support them directly in the manufacturing of them. Martin is indepently making those units and licensing Sonic Orbiter from SGC. I do contribute to the Sonic Orbiter project because it's a vital part of the Rendu system. BTW SGC also licenced software to Innuos in the past so don't think I'm against them. They have their solution and we have ours. Anyway, all off the shelf hardware aside and the use of similar Linux Operating systems the only difference IMHO is the power supply and the price. I don't sell servers so take my input as an outsider looking in.    

[vortecjr]

Also, Barrows is a sub-contractor to Sonore so his opinions are his own. I value his opinion though and he has samples of the opticalRendu boards so he has heard them and compared them to the entire Rendu series. As for putting the money on the table I think we have put a lot of money on the tableWe are happy to do this because we are passionate and proud about our solutions. I also don't see the Rendu series as "decrapifiers". They are just purpose built endpoints or renderers designed with audiophile inspirations. Anyway, I don't go around calling our gear the "endgame" solution like I read so many people here professing they have discovered. Instead the opticalRendu will be reviewed by our customers and they will tell us what they think in a public forum no less. Chris also reviewed the opticalRendu and said, "it was the best Rendu Sonore has ever made." Emphasis on the word "made" because we didn't by parts on NewEgg and just assemble these. 

[BigAlMc]

2 minutes ago, vortecjr said:

. I also don't see the Rendu series as "decrapifiers". They are just purpose built endpoints or renderers 

 

I had the MicroRendu and later upgraded it to the 1.4 board - both courtesy of Martin Smith as I'm in the UK.

 

I liked both and admire your products.

 

I used the phrase decrapifier in response to Barrows continual argument that what's upstream of the Optical Rendu doesn't matter at all. So it's a decrapifier in terms of Barrows argument that it doesn't matter if you use a crap server because total isolation will save the day.

 

This is not my experience and I'm firmly in the 'everything matters' camp. If you don't introduce the crap you don't need to remove it.

 

I got a little strident with Barrows as his one sided argument was getting on my nerves. But I probably overreacted. My bad!

 

Look at my first post on this thread. I was very balanced despite currently owning an Innuos.

 

I hope the Optical Rendu does very well and I have no doubt it'll be a great product and great value. I just take issue with this 'your server doesn't matter because my science says so', mantra.

 

Cheers,

Alan

[vortecjr]

1 minute ago, BigAlMc said:

"I just take issue with this 'your server doesn't matter because my science says so', mantra."

I never said that. However, I don't think they matter as used with our Rendu series. The server is far away from the audio room and isolated from the Rendu via the network. Some times over cable ethernet, some times over wireless Ethernet or some times over optical Ethernet. From memory...Chris reviewed the microRendu over cable ethernet and stated then he could hear a difference with our suggested configuration using DLNA. Your point of reference is with a server directly attached to the USB device where I do think I matters a lot and hence the reason to not connect things that way:) Again just my opinion and your are free to enjoy whatever you think sounds best in your system. I also appreciate your support even if it's through Martin who is a great industry member and a top Sonore dealer and support center.